Reporting framework for value based care

StickyValue-Based Care

As the cost of healthcare continues to rise, both payers and providers continue the “triple aim” of delivering effective quality care while managing their costs. Value based care (VBC), or value based reimbursements, has emerged as an effective mechanism for achieving this goal. In fact, based on estimates developed by National Business Group on Health, almost 40% of employers are incorporating some form of value based incentives for their employee health plan [1]. Moreover, there is increasing motivation and willingness amongst payers to move to value based care and away from the traditional fee-for-service payment structure.

It is important that both the payers and providers are aligned in terms of the methodologies and metrics used for evaluating performance and how “value” is defined and measured. In order to instill confidence for all stakeholders, it is imperative that the VBC evaluation framework be designed such that it is simple and transparent. To perform this type of evaluation, organizations must set up robust data and reporting frameworks that can provide insightful analysis to compare performance by peer provider groups. In this article, we have outlined an approach we used with one of our regional payer clients to design a quantitative methodology that represents the value of providers and provider groups. Please note that there is not a one size fits all solution for all providers and the appropriate methodology should be re-considered for a given arrangement. However, we believe this example is instructive.

A provider’s relative performance, or “value”, was evaluated through three key components: quality, efficiency, and cost. Together, these three components can be combined in the following equation in Figure 1.

Figure 1. Conceptual equation for calculating the value of a healthcare provider.

  • Quality was evaluated through a variety of measures, including evidence based measures (EBMs) published by multiple organizations such as the National Committee for Quality Assurance (NCQA [2]), Agency for Healthcare Research and Quality (AHRQ [3]) and Choosing Wisely® [4]. In our analysis of the providers, quality was defined by several specific EBM and MedInsight Health Waste Calculator (HWC) measures, tailored to each type of provider. For Primary Care Physicians (PCP’s), we identified 14 EBMs specific to primary care, and also assessed the provider’s overall Patient Harm Index (% of members with evidence of an unnecessary harmful service) and Wasteful Service Index (% of services measured that were identified as wasteful) as defined by the HWC.
  • Efficiency was evaluated based on the utilization and number of healthcare services performed by the provider. For members attributed to one PCP, we also measured the total services performed based on the member’s claims in the following care settings: inpatient, outpatient, emergency department, professional, pharmacy, and ancillary. For each setting, resource use was measured using the MedInsight Global Relative Value Units (RVUs) product, which assigns an equal RVU value to similar types of services performed in similar regions in order to conduct fair comparison of resource utilization.
  • Cost was evaluated as the as the overall cost of care provided and also in terms of unit price per service. We defined the overall cost as total allowed dollars and the latter as the allowed dollars per RVU.

One additional critical factor to manage in this evaluation process was to account for the case-mix and severity of members attributed to different providers. The resource use and unit cost for the providers was therefore risk adjusted using Milliman’s Advanced Risk Adjuster (MARA). Risk adjustment can be useful tool, but selecting the appropriate one for a given risk arrangement is outside the scope of this paper.

The practical evaluation of the three components of value was driven by the type of providers as shown in Figure 2.

Figure 2. Four major provider types that drive design of VBC evaluation

There were variations in terms of the metrics used for calculating cost, efficiency and quality for different types of providers. For example, ‘length-of-stay’ was an important factor for evaluating hospitals whereas PCP evaluations included preventative services rendered. These variations were especially nuanced when defining quality measures. A final composite provider score was developed by assigning flexible weight factors to each of these three components. This methodology for composite scoring was based on the MedInsight Provider Composite solution, a new product currently in development.

For this client, we calculated the provider composite scores along with other cost and utilization data, and leveraged the new capabilities within the MedInsight portal to create several dashboards. These dashboards focused on summarizing client provider systems, comparing their cost and efficiency, and identifying contract improvement opportunities associated with ‘lower value’ provider systems. Metrics such as attributed member months, risk adjusted relative cost of care, risk adjusted relative resource use, relative unit price, and the provider composite score were combined to categorize health systems into value tiers. Based on this methodology, clients have the ability to incorporate multiple data components into their value based contracting analyses and develop appropriate strategies for calculating reimbursement based on a provider or provider group’s value.

References:

[1] Bruce Japsen, 2017. Forbes. Employers Accelerate Moves To Value Base-Based Care In 2018. Published: August 17, 2017. Available at: https://www.forbes.com/sites/brucejapsen/2017/08/17/employers-accelerate-move-to-value-based-care-in-2018/#5d009ebf1ec6

[2] NCQA. HEDIS and Performance Measurement. Available at: https://www.ncqa.org/hedis/

[3] Agency for Healthcare Research and Quality. CAHPS: Assessing Healthcare Quality From the Patient’s Perspective. Available at: https://www.ahrq.gov/cahps/index.html

[4] Choosing Wisely. Available at: https://www.choosingwisely.org/

Chronic condition analysis: Selecting a Diagnostic Grouper

StickyValue-Based Care

With over 77,000 codes, ICD10 diagnosis codes provide a much needed level of granularity for clinicians to use in caring for patients, but they are not particularly useful for large scale analytics. As a result, analysts have developed many different diagnostic groupers to group ICD10 codes into meaningful categories for analysis. Several are available within MedInsight; the MARA risk scores, the Chronic Condition Hierarchical Groups (CCHGs), several available episodic groupers, and our Evidence Based Measures quality metrics.

With multiple groupers to choose from, how do you know which one to use? When choosing a grouper, it is important to keep in mind that each grouper is designed to serve a specific primary purpose, so the application or use case for which you need the grouper will be the determining factor in deciding which one will best meet your requirements. 

Some of the most common use cases for groupers are chronic condition cost and prevalence analysis, chronic condition management, estimating patient and population risk, disease pathways and cost of treatment, performance management and quality measures, and care management case finding.

We have provided guidance below on the best grouper to use for each indication (called primary) and have identified other groupers that may also be useful for each use case (called secondary).

Chronic Conditions Cost and Prevalence Analysis:

One common use of a diagnostic grouper is tracking chronic condition prevalence within a population over time.

This is often expressed through common questions such as: Do I have more patients with diabetes and hypertension than in the past? Is the increase in treatment costs for patients with cancer increasing faster than the overall population? Is that impacting my bottom line?

These questions can easily be answered by MedInsight’s Chronic Conditions Hierarchical Groups (CCHGs). Because CCHG covers 100% of a patient population and spend, it is designed for population trending over time. The focus on chronic conditions within the CCHGs, and the manageable number of conditions tracked also makes CCHGs ideal for this application.

In the absence of CCHGs, the individual diagnostic categories within the Milliman Advanced Risk Adjusters (MARA) grouper, or an Episodic Grouper could be used similarly as a secondary option.

Chronic Condition Management:

Chronic condition management poses a similar question to the one above on chronic condition cost and prevalence: How well am I managing my pool of patients with a chronic condition?

In this case, CCHGs is also the primary grouper, particularly for measuring the impact of condition management programs and efforts, or fact-checking the impact of an outside vendor program. CCHGs’ embedded hierarchy will filter out other complex conditions, and give you the ability to measure a clinically similar group of members.

Evidence Based Measures can also be useful here to see if compliance rates for recommended care (e.g. HEDIS measures) for patients with chronic conditions are improving.

Episodic groupers can help monitor performance managing procedures and other episodes of care for members/patients with chronic conditions.

Estimating Patient and Population Risk:

For estimating overall population risk, risk-scoring for a physician panel, or any sub-population of patients, MARA is the best grouper to use. MARA’s patient-level concurrent risk score can be aggregated to create a historical risk score to normalize costs across time and between patient populations. The prospective risk score can also be used to predict future risk of a patient or population. 

In the absence of MARA, CCHGs can be used to trend population prevalence. The costs to treat patients with chronic conditions over time can provide a good view of population risk over time or point in time risk/chronic condition disease burden for a single patient.

Disease Pathways and Cost of Treatment:

Disease pathways and cost of treatment helps to measure the efficiency of care for a specific condition.  Episodic Groupers are the best option when answering questions such as: What is a standard pathway for cancer treatment? Which providers are delivering diabetic care most efficiently?

Evidence Based Measures can also be considered as an additive measure to reflect quality of care in addition to cost efficiency. In the absence of an episodic grouper, CCHGs can add some insight about efficiency by measuring the average cost of a clinically-similar group of patients within each condition group.

Performance Management Compliance and Quality:

Evidence Based Measures are the right place to start when thinking about performance management and quality measures, due to the ability to look at specific compliance with HEDIS or other quality measures. Again, CCHGs can give a good starting point to identify patients with a diabetes and see if they have received an HbA1c, or if your patients with CAD have medication refills if the Evidence Based Measures are unavailable. 

Case Finding and Patient/Member Identification:

Finally, to identify patients for chronic condition management programs, the most effective method is to layer groupers together to prioritize patients for evaluation and possible enrollment. CCHGs is used here to zero in on patients with specific, chronic conditions, such as diabetes or CAD. Adding MARA brings patients to the surface who are most likely to be impactable in better clinical outcomes, lower costs or both. This is a very powerful way to prioritize your scarce and valuable care manger time.

In conclusion, the suite of groupers available in the MedInsight Ecosystem are each designed to serve a specific primary purpose and complement each other for patient or population level analysis. However, some of these tools can provide meaningful insight even outside of their primary application.  For each project or initiative you should choose the grouper within your toolkit that is the best match for the desired application.

Data Confidence for ACOs

StickyACO, Data Confidence, Healthcare Analytics, Value-Based Care,

All healthcare data has data quality challenges. However, as Accountable Care Organizations (ACOs) have taken on more risk and are working on improving care processes data quality has become a more important issue. Below are some common data quality issues that ACOs face and some of the solutions ACOs can use to build confidence in their data.

Incomplete Provider Data

Provider level analysis is extremely important for ACOs. ACOs need to know which providers both within and outside the ACO network are providing services to patients attributed to the ACO. This is important not only to work towards bringing a larger percentage of services in-network (leakage management), but also for quality and efficiency improvement. There are multiple issues with provider data from payer data sources that can make it difficult to correctly identify in and out of network providers. These include:

  • Claims missing complete provider information. Medical claims need to have both the billing and servicing/rendering providers listed, and pharmacy claims need to have both the prescriber and pharmacy listed. It is critical that ACOs work with their data suppliers to ensure that these multiple provider fields are complete on claims.
  • Custom provider identifiers. Some data sources use custom provider identifiers, instead of National Provider Identifiers (NPIs). To perform analysis across data sources from different suppliers, any custom identifiers need to be cross-walked and mapped to a consistent standard such as NPI. For facilities or large practices, which generally have multiple NPIs or may use alternative identifiers, it is important to roll up the identifiers present in the data for analytic purposes.

Incomplete Financial Fields

Data suppliers often remove or mask financial data to ensure that provider reimbursement terms for providers outside of the ACO network remain confidential. Financial values are useful in ensuring that the data is complete, and are necessary to determine the magnitude of differences in resource cost between different services. A variety of tools, including MedInsight Global RVUs, . The conversion factor can be derived using benchmarks or by dividing the total cost of the contract by total RVUs. While this does not replicate actual unit cost it can provide reasonable approximations ACOs can use to make decisions.

Incomplete Diagnosis Coding

Many ACO contracts include financial parameters that are risk adjusted and it is important to have all diagnosis codes available for analysis, as these diagnoses drive risk scores. To test for the quality of the diagnosis coding in a given data source, users can audit both the number of codes per claims and the ACOs can use benchmarks to ensure that their claims have reasonable population of diagnosis codes and use other tools to review the consistency of diagnosis coding over time.

Completeness of Electronic Medical Record (EMR) Data

More analysis is being done using EMR data and combined EMR/claims data. In order to appropriately incorporate EMR data, the ACO needs to ascertain how much of a patient’s clinical care was delivered by providers using that EMR system, as clinical data from providers using alternative EMR systems would not be included in the data. It’s also important to gauge the relative quality of the EMR fields. An example is to measure the completeness of fields in the encounter file.

These are a few examples of the importance of ACO data quality and how ACOs can use analytics tools to improve data quality. As healthcare analytics continue to play an increasingly important role in decision-making, utilization and cost, ACOs will need to work closely with their data suppliers to continue to improve data quality.

Rising to the challenge: Adapting to meet patient care needs during the COVID-19 pandemic

StickyACO, Value-Based Care

As accountable care organizations (ACOs) consider options for advancing their population health initiatives in the midst of the coronavirus pandemic, some organizations are choosing to opt-out of risk based programs for 2020, some are staying in and putting proactive activities on hold but some are carrying the mantle of population health forward, to serve patient needs in this time. For organizations that are truly invested in a long-term population health strategy, there is a need to continue to proactively manage patients, even as traditional medical services are disrupted. As noted in the Professional Case Management Journal, this pandemic has provided opportunities for innovation and creativity including use of digital and telecommunication technology in new ways to ensure the continued delivery of health and to those who need them regardless of location. This is consistent with what we have heard from several client organizations, and we wanted to share their ideas and strategies with you.  

The Care Coordination Institute (CCI) in Greenville, SC (affiliated with Prisma Health), is maximizing patients’ ability to interface with providers to care for their chronic conditions at home. In response to the pandemic, CCI has implemented new strategies for monitoring chronic condition health, for example, asking patients to enter daily blood glucose results electronically for care manager review.  In order to making sure that chronic condition patients still receive the care they need when unable to attend an in-person visit or worried about presenting in person, CCI implemented a Virtual Visit for Chronic Condition patient strategy, focusing first on the highest risk patients. Additionally, they have creatively adapted the in-vehicle COVID-19 testing model and created a new drive-through blood pressure monitoring and lab draw clinic. Pushing the boundaries of preventive care that patients can receive at home, CCI is helping physician practices send patients Cologuard® tests to complete at home. Finally, CCI is partnering with chronic condition patients to develop solutions to their medication adherence challenges. While these programs were developed to mitigate the virtual transmission risks of in-person healthcare services  while facilitating continued access to necessary care during the coronavirus pandemic, CCI is evaluating whether some should last for the longer term; for example, both patients and providers have high rates of satisfaction with virtual visits.

The coronavirus pandemic hit New York City-based Mt. Sinai Health System hard, but they were determined to continue their focus on proactive outreach to their attributed patients. The population health team at Mt. Sinai recognized that patients were reluctant to engage in- person with the healthcare system because of fears of COVID-19, and that the population health team was uniquely positioned to conduct a campaign of proactive outreach to ensure patients’ needs were being met. First, they conducted an outreach campaign to all patients currently enrolled in care management to discuss their needs. From that initial outreach, they identified five key themes:

  • Chronic condition management
  • Pharmaceutical access
  • Food insecurity
  • Behavioral health – exacerbated by financial distress, social isolation, etc.
  • And finally, COVID-19 related symptoms

Mt. Sinai then expanded their outreach to include a broader set of at-risk patients. They created a list of patients at high-risk during the pandemic, including patients that are immunosuppressed or have heart disease, COPD, or other select conditions. They paired this patient list with the development of a resource guide that provides a description of resources to help address each of the key themes. Then, Mt. Sinai’s care coordinators, or where possible, their primary care physician’s staff, contacted these high-risk patients and leveraged the resource guide to help resolve their issues. Outreach staff reported higher than typical level of engagement in these calls, as patients were largely home, in need, and had questions about their health.

Both CCI and Mt. Sinai have adapted quickly to adopt telehealth. Mt. Sinai is currently completing more telehealth visits each day as they completed across the whole of 2019. Additionally, the team at CCI tells us that both patients and providers have grown to support the telehealth model, and they are hopeful that enhanced reimbursement for telehealth will remain after the crisis is over. CCI is considering continuing some of their remote initiatives anyway, recognizing the benefit of enhanced population healthcare. It is possible that some of the innovative initiatives these organizations are deploying today will have longer-term benefits in their overall population health strategy.

To learn more, click here to watch our webinar, Telehealth Expansion and Other Creative Strategies for Caring for Patients during COVID-19 – Stories from Two Clients.

Chronic condition analysis: Selecting a Diagnostic Grouper

StickyValue-Based Care

With over 77,000 codes, ICD10 diagnosis codes provide a much needed level of granularity for clinicians to use in caring for patients, but they are not particularly useful for large scale analytics. As a result, analysts have developed many different diagnostic groupers to group ICD10 codes into meaningful categories for analysis. Several are available within MedInsight; the MARA risk scores, the Chronic Condition Hierarchical Groups (CCHGs), several available episodic groupers, and our Evidence Based Measures quality metrics.

With multiple groupers to choose from, how do you know which one to use? When choosing a grouper, it is important to keep in mind that each grouper is designed to serve a specific primary purpose, so the application or use case for which you need the grouper will be the determining factor in deciding which one will best meet your requirements. 

Some of the most common use cases for groupers are chronic condition cost and prevalence analysis, chronic condition management, estimating patient and population risk, disease pathways and cost of treatment, performance management and quality measures, and care management case finding.

We have provided guidance below on the best grouper to use for each indication (called primary) and have identified other groupers that may also be useful for each use case (called secondary).

Chronic Conditions Cost and Prevalence Analysis:

One common use of a diagnostic grouper is tracking chronic condition prevalence within a population over time.

This is often expressed through common questions such as: Do I have more patients with diabetes and hypertension than in the past? Is the increase in treatment costs for patients with cancer increasing faster than the overall population? Is that impacting my bottom line?

These questions can easily be answered by MedInsight’s Chronic Conditions Hierarchical Groups (CCHGs). Because CCHG covers 100% of a patient population and spend, it is designed for population trending over time. The focus on chronic conditions within the CCHGs, and the manageable number of conditions tracked also makes CCHGs ideal for this application.

In the absence of CCHGs, the individual diagnostic categories within the Milliman Advanced Risk Adjusters (MARA) grouper, or an Episodic Grouper could be used similarly as a secondary option.

Chronic Condition Management:

Chronic condition management poses a similar question to the one above on chronic condition cost and prevalence: How well am I managing my pool of patients with a chronic condition?

In this case, CCHGs is also the primary grouper, particularly for measuring the impact of condition management programs and efforts, or fact-checking the impact of an outside vendor program. CCHGs’ embedded hierarchy will filter out other complex conditions, and give you the ability to measure a clinically similar group of members.

Evidence Based Measures can also be useful here to see if compliance rates for recommended care (e.g. HEDIS measures) for patients with chronic conditions are improving.

Episodic groupers can help monitor performance managing procedures and other episodes of care for members/patients with chronic conditions.

Estimating Patient and Population Risk:

For estimating overall population risk, risk-scoring for a physician panel, or any sub-population of patients, MARA is the best grouper to use. MARA’s patient-level concurrent risk score can be aggregated to create a historical risk score to normalize costs across time and between patient populations. The prospective risk score can also be used to predict future risk of a patient or population. 

In the absence of MARA, CCHGs can be used to trend population prevalence. The costs to treat patients with chronic conditions over time can provide a good view of population risk over time or point in time risk/chronic condition disease burden for a single patient.

Disease Pathways and Cost of Treatment:

Disease pathways and cost of treatment helps to measure the efficiency of care for a specific condition.  Episodic Groupers are the best option when answering questions such as: What is a standard pathway for cancer treatment? Which providers are delivering diabetic care most efficiently?

Evidence Based Measures can also be considered as an additive measure to reflect quality of care in addition to cost efficiency. In the absence of an episodic grouper, CCHGs can add some insight about efficiency by measuring the average cost of a clinically-similar group of patients within each condition group.

Performance Management Compliance and Quality:

Evidence Based Measures are the right place to start when thinking about performance management and quality measures, due to the ability to look at specific compliance with HEDIS or other quality measures. Again, CCHGs can give a good starting point to identify patients with a diabetes and see if they have received an HbA1c, or if your patients with CAD have medication refills if the Evidence Based Measures are unavailable. 

Case Finding and Patient/Member Identification:

Finally, to identify patients for chronic condition management programs, the most effective method is to layer groupers together to prioritize patients for evaluation and possible enrollment. CCHGs is used here to zero in on patients with specific, chronic conditions, such as diabetes or CAD. Adding MARA brings patients to the surface who are most likely to be impactable in better clinical outcomes, lower costs or both. This is a very powerful way to prioritize your scarce and valuable care manger time.

In conclusion, the suite of groupers available in the MedInsight Ecosystem are each designed to serve a specific primary purpose and complement each other for patient or population level analysis. However, some of these tools can provide meaningful insight even outside of their primary application.  For each project or initiative you should choose the grouper within your toolkit that is the best match for the desired application.

Doing predictive analytics with Python – Podcast series

StickyAnalytics, Healthcare Analytics

Python is a very popular coding language for doing predictive modeling and data science. We have been discussing python as part of our ongoing Predictive Analytics podcast series for the Society of Actuaries. Four episodes are available for listening now: 

Episode 1:

This is the first of a few podcasts focused on Python, a popular tool for predictive modeling and machine learning. In this episode, join Anders Larson, FSA, MAAA and Shea Parkes, FSA, MAAA as they cover some basics and talk about when Python may be a good idea to try.

Listen to episode 1

Episode 2:

Join hosts Anders Larson, FSA, MAAA, and Shea Parkes, FSA, MAAA, for the second in a series of podcasts focused on Python. After giving an introduction to the popular programming language in our previous episode, they discuss some key concepts in Python, such as its object-oriented framework, the idea of namespaces, the ability to create package sets and a few other topics.

Listen to episode 2

Episode 3:

Join hosts Anders Larson, FSA, MAAA, and Shea Parkes, FSA, MAAA, for the third in a series of podcasts focused on Python. Moving on from the foundational concepts and background from the prior two episodes, this episode moves into more practical advice for actuaries looking to get started with Python. The discussion includes pros and cons of various editing software and user interfaces. To provide some useful context, Shea also discusses the key considerations that his own team made as they implemented Python into their operations.

Listen to episode 3

Episode 4:

Join hosts Anders Larson, FSA, MAAA, and Shea Parkes, FSA, MAAA, for the fourth in a series of podcasts focused on Python. The previous episode covered how to get started with Python. This episode covers useful packages for data analysis. Python is a general purpose language at heart, so you will likely need to use a variety of packages to perform most data science tasks. Luckily, the Python ecosystem is full of feature-rich data science packages. Listen to this episode to learn about some of the most important ones. 

Listen to episode 4

The full backlog of podcasts in this Predictive Analytics series is available here (and on Stitcher etc).

The increasing relevance of risk scores on MSSP ACOs

StickyACO, Value-Based Care

Medicare Advantage plans have long recognized the importance of coding to ensure accurate risk scores and clinical documentation.  By contrast, Accountable Care Organizations (ACOs) participating in the Medicare Shared Savings Program (MSSP) often focus less on the implications of risk adjustment or assume they have little ability to affect their risk scores.  To understand why, there are two key aspects of the MSSP to consider:

  1. Regional adjustment: In the early years of the MSSP, benchmarks were purely based on the ACO’s historical experience. These benchmarks were not adjusted based on the absolute risk scores for the ACO beneficiaries, but rather the change in risk scores over time.  Therefore, these ACOs were not penalized if their risk scores were understated by a similar amount in each year. This began to change in 2017 with the introduction of the regional benchmark adjustment, in which the benchmark is adjusted for the absolute difference between the ACO and the region average risk scores – making the ACOs’ absolute risk scores an important factor.  However, before the Pathways to Success rule, this regional benchmark adjustment only applied to ACOs that started a second agreement period in 2017 or later.
  2. Demographic adjustment: Prior to the Pathways to Success rule, the MSSP limited the risk score changes for “continuously assigned” beneficiaries to be no more than a demographic adjustment.  Because of the way continuously assigned beneficiaries were identified[1], this rule effectively made it impossible for ACO efforts to improve coding accuracy to significantly increase the ACO’s historical benchmark.

Given the considerations above, it is not surprising that ACOs have historically deemphasized coding accuracy. However, the landscape is now changing in ways that make risk scores even more important for two reasons: 

  1. Under the Pathways to Success rule, the limitation on risk score changes for continuously assigned beneficiaries has been removed.  Instead, there is now a limit of 3% in the total increase in risk score for the ACO from the historical benchmark to the performance year.  This limit prevents ACOs from dramatically increasing their benchmark strictly due to risk score increases, but it should not deter ACOs from trying to ensure their risk scores are accurate and complete. 
  2. Further, the year to year risk score adjustment can be greater than the 3% cap. For example, if the ACO’s underlying risk score decreases from baseline year 3 (BY3) to performance year 1 (PY1), then the risk score change from PY1 to PY2 could be greater than 3%.  

The figure below illustrates this possibility using a hypothetical ACO that improves coding accuracy from PY1 to PY2.

  • In the “No BY3 to PY1 Risk Score Change” scenario, the ACO’s historical benchmark is limited to a 3.0% risk score change even though coding accuracy efforts led to an 8.3% risk score increase.
  • In the “Negative BY3 to PY1 Risk Score Change” scenario, the ACO’s risk score would have decreased by 2.5% relative to BY3 if risk scores remained flat from PY1 to PY2, but the coding accuracy efforts actually led to a 5.8% risk score increase relative to BY3.  Even though the change from BY3 to any PY is capped at 3.0%, the value of the coding accuracy efforts is 5.5% (3.0% minus -2.5%).

Effect of Risk Scores on Historical MSSP Results

Even before the implementation of the Pathways to Success rule, coding accuracy efforts could limit risk score decreases for the continuously assigned population.  Based on a review of the 2014 through 2018 MSSP Public Use Files (PUFs), we found that approximately 49% of ACOs had an aggregate risk score change from Benchmark Year 3 (BY3) to Performance Year (PY) of less than 1.000.  Although the PUFs do not split out the continuously assigned and newly assigned risk scores, it is likely that many of these ACOs had a decrease in their continuously assigned risk scores, which means these ACOs could have benefitted from coding accuracy efforts to mitigate this decrease.  If these ACOs had all been able to maintain an aggregate risk score change of 1.000, the average gross savings across all ACOs from 2014-2018 would have increased from approximately 1.1% to 1.9%[2].  The graph below shows the distribution of gross savings before and after applying a 1.000 floor to the risk score change.

The impact on shared savings depends on the specific situation of each ACO, such as their minimum savings rate, Track selection, and quality performance.  However, for example, if we applied the sharing parameters of the ENHANCED Track and a 2.0% minimum savings rate/minimum loss rate (MSR/MLR) to each ACO, this would have resulted in an average increase in average net shared savings/(losses) of $51 per beneficiary per year (PBPY).  For a 20,000-life ACO, this would equate to approximately $1 million.

Regional Benchmark Adjustment

The benchmark for most ACOs in their second or later agreement period, and ACOs starting under the Pathways to Success rule, is adjusted based on the ACO’s risk-adjusted expenditures relative to all Medicare FFS beneficiaries in the ACO’s region.  The regional benchmark adjustment is based on each ACO’s risk scores for their last benchmark year. Therefore, the risk scores in an ACO’s last benchmark year are particularly important. For an ACO starting an agreement period under Pathways to Success in CY 2020, the risk-adjusted regional expenditures for their next agreement period will be based on risk scores in CY 2024, which use diagnosis codes from CY 2023.  While not immediately relevant for active ACOs, this will have a significant impact on the benchmark for the next agreement period.

Conclusion

No matter their specific situation, ACOs should ensure their assigned beneficiaries have accurately coded risk scores that reflect their full morbidity level.  As a result of changes to MSSP rules over the past few years, particularly the Pathways to Success rule, risk adjustment is becoming increasingly relevant to financial results for ACOs participating in the MSSP, and neglecting the importance of accurate risk score coding accuracy is a risk that is not worth taking.

References:

Shared Savings Program Accountable Care Organizations (ACO) Public Use Files. Centers for Medicare & Medicaid Services. Available at: https://www.cms.gov/Research-Statistics-Data-and-Systems/Downloadable-Public-Use-Files/SSPACO/index

[1] Any beneficiary who was either assigned to the ACO or received a primary care service from any of the ACO participants in the previous year was considered “continuously assigned.”

[2] Due to the limitations of the PUFs, this analysis does not precisely capture the impact of risk score decreases on each ACO’s benchmark, but approximates the overall impact across all ACOs.  Some ACOs with a risk score decrease may still have been subject to the demographic adjustment on the continuously assigned population, meaning they may not have benefited from a higher performance year risk score.  However, it is also possible that some ACOs with a risk score increase may have had a decrease in risk score for the continuously assigned population, and therefore they could have benefited from a higher performance year risk score.

Thank you to Grant Churchill for his contributions to the analyses in this post.

Telehealth: Adoption and value

Sticky, Value-Based Care

Telehealth is the use of electronic information and telecommunication technologies to support remote clinical healthcare, patient and professional health-related education, public health, and health administration. Currently millions of Americans are residing in areas with a shortage of primary healthcare providers and often experience delays to see a provider.[i] Telehealth is believed to improve access to healthcare for patients living in both rural and urban areas and ensure that patients receive the right care at a place and time most accessible to them. In addition, telehealth is also believed to reduce healthcare costs by:

  • Optimizing staff distribution and healthcare resources across healthcare facility or system
  • Reducing unnecessary office and emergency room visits and hospital admissions
  • Reducing financial impact on providers in case of no-shows by patients.[ii]

Currently, 31 states and the District of Columbia have parity laws that mandate commercial insurers pay for telehealth services.[ii] Unfortunately, there are barriers to wide adoption of telehealth. For example, Medicare generally still limits coverage and payment for many telehealth services, lagging behind other payers. [iii]

We analyzed one of MedInsight’s client’s data for 2010-2017 with over 3.9 million Medicaid, Medicare, and Commercial plan members, to explore the use of telehealth services. We only explored those services recognized by federal and commercial health plans [iv] as billable for telehealth services.

Trends in telehealth use

We analyzed the data to study the utilization of telehealth services over 8 years and its distribution across different age groups and gender (Figure 1, Figure 2 and Figure 3). We observed that:

  • Although the proportion of telehealth visits increased over recent years, it still remains well below one percent
  • Over the years the average cost difference between telehealth and non-telehealth services has increased from almost nil to about $40 per visit with telehealth being cheaper than non-telehealth services

Figure 1: Trends in telehealth use

Figure 2: Average cost trend

  • Females were found to have used telehealth more frequently as compared to males
  • Members below 19 or above 64 years old were less likely to use telehealth services as compared to members 19 to 64 years old. This could be due to parental preference for their children, Medicare coverage limitations, or a matter of trust on the conventional methods amongst the older age groups.

Figure 3: Telehealth use across age groups and gender

Specialty based use of telehealth

Apart from analyzing the telehealth use by patients, we also analyzed the use across provider specialties (Figure 4) for the year 2017.

  • The top ten specialties with highest number of telehealth visits constituted about 87% of the total telehealth visits
  • Nurse practitioners were the most likely to use telehealth amongst all provider specialties followed by clinical psychologist and psychiatrist

Figure 4: Telehealth use by specialty

Clinical Classifications Software category based use of telehealth

Instead of analyzing data at the individual diagnosis level, we compared the telehealth use for the year 2017 at the Clinical Classifications Software (CCS) category[v] level, (Figure 5) which provides a method for classifying diagnoses into clinically meaningful categories.

  • The top ten CCS categories with highest number of telehealth visits constituted about 82% of the total telehealth visits
  • Nine out of the top ten CCS categories were related to mental health and behavioral disorders. This is in line with the top ten specialties with maximum telehealth use
  • Overall, telehealth constituted below 2% of all included services, the rest being accounted by other type of visits

Figure 5: Telehealth use by CCS diagnosis category

Although the analysis was based on limited administrative data, it illustrated that telehealth is a far less expensive option in comparison to the conventional face-to-face visits. Even a one percent shift from face-to-face visits to telehealth can save millions of dollars. However, members, payers, and providers need to start embracing telehealth as effective as any other form of healthcare service delivery method.

Resources:

[i] Telehealth Programs, Health Resources and Services Administration. Available at: https://www.hrsa.gov/rural-health/telehealth/index.html

[ii] Jamal H. Mahar, MD et al. 2018. Telemedicine: Past, present, and future. Available at: https://www.mdedge.com/ccjm/article/189759/practice-management/Telehealth-past-present-and-future

[iii] American Hospital Association. Fact sheet: Telehealth. Available at:  https://www.aha.org/system/files/2019-02/fact-sheet-telehealth-2-4-19.pdf

[iv] United Healthcare, telehealth and telemedicine policy. Policy Number: ADMINISTRATIVE 114.32 T0. April 2019. Available at: https://www.uhcprovider.com/content/dam/provider/docs/public/policies/index/oxford/telemedicine-ohp.pdf

[v] Clinical Classification Software, Tools and Software, Healthcare Cost Utilization Project. Available at: https://www.hcup-us.ahrq.gov/tools_software.jsp

Reporting Framework for Value Based Care

StickyValue-Based Care

As the cost of healthcare continues to rise, both payers and providers continue the “triple aim” of delivering effective quality care while managing their costs. Value based care (VBC), or value based reimbursements, has emerged as an effective mechanism for achieving this goal. In fact, based on estimates developed by National Business Group on Health, almost 40% of employers are incorporating some form of value based incentives for their employee health plan [1]. Moreover, there is increasing motivation and willingness amongst payers to move to value based care and away from the traditional fee-for-service payment structure.

It is important that both the payers and providers are aligned in terms of the methodologies and metrics used for evaluating performance and how “value” is defined and measured. In order to instill confidence for all stakeholders, it is imperative that the VBC evaluation framework be designed such that it is simple and transparent. To perform this type of evaluation, organizations must set up robust data and reporting frameworks that can provide insightful analysis to compare performance by peer provider groups. In this article, we have outlined an approach we used with one of our regional payer clients to design a quantitative methodology that represents the value of providers and provider groups. Please note that there is not a one size fits all solution for all providers and the appropriate methodology should be re-considered for a given arrangement. However, we believe this example is instructive.

A provider’s relative performance, or “value”, was evaluated through three key components: quality, efficiency, and cost. Together, these three components can be combined in the following equation in Figure 1.

Figure 1. Conceptual equation for calculating the value of a healthcare provider.

  • Quality was evaluated through a variety of measures, including evidence based measures (EBMs) published by multiple organizations such as the National Committee for Quality Assurance (NCQA [2]), Agency for Healthcare Research and Quality (AHRQ [3]) and Choosing Wisely® [4]. In our analysis of the providers, quality was defined by several specific EBM and MedInsight Health Waste Calculator (HWC) measures, tailored to each type of provider. For Primary Care Physicians (PCP’s), we identified 14 EBMs specific to primary care, and also assessed the provider’s overall Patient Harm Index (% of members with evidence of an unnecessary harmful service) and Wasteful Service Index (% of services measured that were identified as wasteful) as defined by the HWC.
  • Efficiency was evaluated based on the utilization and number of healthcare services performed by the provider. For members attributed to one PCP, we also measured the total services performed based on the member’s claims in the following care settings: inpatient, outpatient, emergency department, professional, pharmacy, and ancillary. For each setting, resource use was measured using the MedInsight Global Relative Value Units (RVUs) product, which assigns an equal RVU value to similar types of services performed in similar regions in order to conduct fair comparison of resource utilization.
  • Cost was evaluated as the as the overall cost of care provided and also in terms of unit price per service. We defined the overall cost as total allowed dollars and the latter as the allowed dollars per RVU.

One additional critical factor to manage in this evaluation process was to account for the case-mix and severity of members attributed to different providers. The resource use and unit cost for the providers was therefore risk adjusted using Milliman’s Advanced Risk Adjuster (MARA). Risk adjustment can be useful tool, but selecting the appropriate one for a given risk arrangement is outside the scope of this paper.

The practical evaluation of the three components of value was driven by the type of providers as shown in Figure 2.

Figure 2. Four major provider types that drive design of VBC evaluation

There were variations in terms of the metrics used for calculating cost, efficiency and quality for different types of providers. For example, ‘length-of-stay’ was an important factor for evaluating hospitals whereas PCP evaluations included preventative services rendered. These variations were especially nuanced when defining quality measures. A final composite provider score was developed by assigning flexible weight factors to each of these three components. This methodology for composite scoring was based on the MedInsight Provider Composite solution, a new product currently in development. Additional information on the MedInsight Provider Composite Solution will be discussed in an upcoming MedInsight blog article. The subsequent article will feature additional use cases and wider applications of the product.

For this client, we calculated the provider composite scores along with other cost and utilization data, and leveraged the new capabilities within the MedInsight portal to create several dashboards. These dashboards focused on summarizing client provider systems, comparing their cost and efficiency, and identifying contract improvement opportunities associated with ‘lower value’ provider systems. Metrics such as attributed member months, risk adjusted relative cost of care, risk adjusted relative resource use, relative unit price, and the provider composite score were combined to categorize health systems into value tiers. Based on this methodology, clients have the ability to incorporate multiple data components into their value based contracting analyses and develop appropriate strategies for calculating reimbursement based on a provider or provider group’s value.

References:

[1] Bruce Japsen, 2017. Forbes. Employers Accelerate Moves To Value Base-Based Care In 2018. Published: August 17, 2017. Available at: https://www.forbes.com/sites/brucejapsen/2017/08/17/employers-accelerate-move-to-value-based-care-in-2018/#5d009ebf1ec6

[2] NCQA. HEDIS and Performance Measurement. Available at: https://www.ncqa.org/hedis/

[3] Agency for Healthcare Research and Quality. CAHPS: Assessing Healthcare Quality From the Patient’s Perspective. Available at: https://www.ahrq.gov/cahps/index.html

[4] Choosing Wisely. Available at: https://www.choosingwisely.org/

More Than Half of CT Scans of the Head for Dizziness in the Emergency Department are Considered Wasteful

Sticky, Evidence Based Decision Making, Health Waste Calculator, Healthcare Data Collection, Value-Based Care

Dizziness accounts for 3.3% of all Emergency Department (ED) visits annually. This corresponds to about 2.6 million ED visits per year in the United States.¹ A Computed Tomography (CT) scan of the head is one of the most frequently used diagnostic procedures for evaluation of dizziness in the ED, despite various studies reporting its low sensitivity, low predictive value, and low diagnostic yield for evaluation of dizziness.²

The American College of Emergency Physicians (ACEP) recommends against the use of CT scans of the head in the ED for asymptomatic adult patients with a normal neurological evaluation, and without significant trauma.³ The ACEP considers use of CT scans of the head for dizziness as inappropriate and wasteful, as it is not conclusive. The test is also expensive and potentially hazardous due to unnecessary exposure to radiation.

To identify the prevalence of CT scans of the head in the ED for patients reporting dizziness, we conducted an analysis on a large commercial dataset with claims from major health plans (approx. 13 million members). We used MedInsight’s Health Waste Calculator (HWC) application to identify service count, waste index (percentage of services measured that were wasteful), and associated costs of CT scans of the head for dizziness in ED. The HWC identifies CT scans of the head as not wasteful for members with associated head injury, benign or malignant tumors of head and neck or neurological deficit. Table 1 summarizes the utilization of CT scans of the head for dizziness in ED for a commercial population in the above dataset for 2017. There were 6,794 CT scans of the head for dizziness in 2017, out of which 4,066 (60%) services were wasteful. This was almost 0.2% of all the wasteful services for 2017, and contributed to 0.8% of all the wasteful health expenditure for 2017.

Table 1: Utilization of CT Scans of the Head for Dizziness in ED – 2017

Analysis of Utilization services by age showed higher utilization of CT scans of the head in members with advancing age. Figure 1 shows the distribution of population and services of CT scans of the head for dizziness in ED by age. The proportion of services was higher in members above the age of 50. This was in contrast to the pattern seen in members less than 50 years old.

Figure 1: Distribution of Population & Services of CT Scans of the Head for Dizziness in ED by Age – 2017

Further Analysis of utilization of CT scans of the head services by gender (summarized in Table 2) showed higher utilization, as well as higher waste index in females, as compared to males. A total of 3,377 CT scans of the head were performed in females, as compared to 3,017 in men. The waste index in females was 62%, as compared to 57% in males. The prevalence of dizziness was also seen to be higher in females as compared to males. This was in line with reports of dizziness in females in various publications.⁴

Table 2: Services of CT Scans of the Head for Dizziness in ED by Gender – 2017

Interestingly, although there was high utilization of services in adults above the age of 50, the prevalence of wasteful services was found to be slightly lower in this age group for both genders. Figure 2 shows the waste index by age bands for both genders. The waste index (WI) in young adults (less than 50 years old) was above 60% and decreased gradually to around 50% for members older than 70 for both genders.

Figure 2: Waste Index of Head CT Scans for Dizziness in the Emergency Department by Age Band – 2017

To review the year over year trend of services and cost associated with these services, we analyzed head CT scans for dizziness in the ED from two years prior for the same population. Table 3 displays the utilization and cost analysis for this service.

Table 3: Trend in Utilization and Cost of Head CT Scans for Dizziness in the Emergency Department: 2015-2017

Despite the recommendations against its usage from various medical societies, such as American Society of Emergency Physicians and associations such as Choosing Wisely, the waste index was consistently above 50% for all three years. Figure 3 summarizes the trend of utilization of CT scans of the head in the ED for dizziness for 2015-2017. An overall increase in the cost of these services was noticed, although there was a slight decrease in provision of the services over the years. Further, the contribution of this series to the overall waste increased year over year, both in terms of wasteful services and wasteful dollars.

Figure 3: Service Count and Cost Trend for Head CT Scans for Dizziness in the Emergency Department: 2015-2017

It is important to note that claims data alone allows only an approximate identification of wasteful CT scans of the head in the ED for dizziness. The HWC application identifies wasteful services, which can help identify opportunities to address avoidable costs and achieve appropriate clinical care. This analysis confirmed the high prevalence of wasteful CT scans of the head for a commonly reported problem in the ED in line with various publications. Studies recommend that dizziness requiring only symptomatic management should be carefully differentiated from cases requiring further diagnostic work-up.⁵

References:

  1. Mitsunaga M, Yoon H. Head CT Scans in the Emergency Department for Syncope and Dizziness. American Journal of Roentgenology. 2015;204(1):24-28.
  2. Lawhn-Heath C, Buckle C, Christoforidis G, Straus C. Utility of head CT in the evaluation of vertigo/dizziness in the emergency department. Emergency Radiology. 2012;20(1):45-49.
  3. ACEP – Avoid head CT for asymptomatic adults with syncope | Choosing Wisely [Internet]. Choosingwisely.org. 2019 [cited 8 January2019]. Available from: http://www.choosingwisely.org/clinician-lists/acep-avoid-head-ct-for-asymptomatic-adults-with-syncope/ (http://www.choosingwisely.org/clinician-lists/acep-avoid-head-ct-for-asymptomatic-adults-with-syncope/)
  4. Kerber K, Callaghan B, Telian S, Meurer W, Skolarus L, Carender W et al. Dizziness Symptom Type Prevalence and Overlap: A US Nationally Representative Survey. The American Journal of Medicine. 2017;130(12):1465.e1-1465.e9.
  5. Ahsan S, Syamal M, Yaremchuk K, Peterson E, Seidman M. The costs and utility of imaging in evaluating dizzy patients in the emergency room. The Laryngoscope. 2013;123(9):2250-2253.

Top 7 Critical Point Podcasts

StickyOpioids, Value-Based Care, Waste

Critical Point is a podcast by Milliman, hosted by subject matter experts including actuaries, clinicians, technology specialists, and economists. These seven episodes take a deep dive into subject matters that affect the health and well-being of people across the world. Critical Point examines the inherent challenges and innovative solutions that Milliman consultants encounter throughout their daily work around healthcare, benefits, risk management, technology, investment consulting, insurance, and financial services.

Episode 1 – Healthcare Waste:

In this episode of Critical Point, Milliman MedInsight’s Dr. David Mirkin, Marcos Dachary, and Jackie Sehr discuss the history of healthcare waste, why identifying healthcare waste is so important, and how to identify low-value care.

Listen to Healthcare Waste

Episode 2 – Alternative Payment Models 101:

This episode of Critical Point includes a discussion from Milliman Senior Healthcare Consultant, Pamela Pelizzari on alternative payment methods, bundled payment, ACOs, and MACRA.

Listen to Alternative Payment Models 101

Episode 4 – A Primer on Telehealth in the U.S.:

In this episode of Critical Point, Milliman’s Senior Healthcare Management Consultant, Susan Philip discusses the history of telehealth, and how it encompasses access to quality care and convenience.

Listen to A Primer on Telehealth in the U.S.

Episode 7 – Diagnosing and Predicting Opioid Use Disorder:

In this episode of Critical Point, Milliman employees, Soddard Davenport and Joseph Boschert discuss their latest research on opioid use in the United States, including underdiagnoses and how advanced analytics can help predict whether a patient may develop opioid use disorder.

Listen to Diagnosing and Predicting Opiod Use Disorder

Episodes 3, 9, and 11 – Shea and Anders Talk Blockchain (Three-Part Series)

In this three-part series, Milliman team members, Shea Parkes and Anders Larson discuss blockchain and its insurance applications. Throughout this series, the two start by giving an overview of blockchain basics, and end with the insurance applications of blockchain.

Listen to Shea and Anders Talk Blockchain (Part 1)

Listen to Shea and Anders Talk Blockchain (Part 2)

Listen to Shea and Anders Talk Blockchain (Part 3)

Diagnosis Related Group-Based Readmission Profiling

Sticky, Value-Based Care

Introduction

High rates of unplanned readmissions indicate poor quality of care or gaps in care coordination, and bring an increased financial burden on the healthcare system.[i] The major reasons for unplanned readmissions include premature discharge from initial hospitalization, gaps in post-operative care, unnecessary inpatient admissions from the Emergency Department, no integrated system for patient records leading to incomplete information sharing, and misunderstandings of patient discharge instructions.[ii]

A study published by the Healthcare Cost and Utilization Project (HCUP) in 2014, found that 3.3 million 30-day readmissions occurred in the United States in 2011 and cost $41 billion.[iii] Healthcare policy makers, payers, and providers are making efforts to reduce hospital readmissions. The Centers for Medicaid and Medicare Services (CMS) in 2012 initiated the Hospital Readmissions Reduction Program (HRRP) to bring attention to unnecessary readmissions. Since then, readmissions within 30 days of an acute inpatient discharge have become a key indicator for healthcare quality.[iv]

Penalties imposed by Medicare for high readmission rates increased by $108 million from 2015 to 2016, putting the total withheld reimbursements for 2016 at $528 million.1 However, between 2007 and 2015, the overall readmission rate for the targeted diseases of HRRP decreased from 21.5% to 17.8% and rates for non-targeted conditions decreased from 15.3% to 13.1%.[v]

MedInsight provides a number of evidence-based measures to assess the quality of inpatient care such as the Plan All-Cause Readmissions (PCR) Healthcare Effectiveness Data and Information Set (HEDIS®[1])measure which provides a comparison between the expected and observed readmission rate. We analyzed a MedInsight client’s 2016 data containing over 3.9 million Medicaid, Medicare Advantage, and Commercial plan members, to explore the most common conditions resulting in readmissions within 7, 15, and 30 days of an acute inpatient discharge.

Diagnosis Related Group-Based Summarization

Out of all reported admissions in the year 2016, 3.2%, 6.3% and 10.3% of cases resulted in unplanned readmissions within 7, 15, and 30 days, respectively. Table 1 displays the percentage of index admissions that had an unplanned readmission within 7, 15, and 30 days, for ten Medicare Severity-Diagnosis Related Groups (MS-DRGs) families with high admission counts and significant readmission rates. Instead of individual DRGs we used DRG families created by Milliman’s New York health practice which groups together clinically related DRGs. We found that not all DRG families with high index admission counts have high readmission counts. For example, the DRG family ‘Lower Extremity Arthroplasty’ had the highest admission count but a low readmission count.

A 2017 statistical brief by the HCUP showed the top 20 index admission primary diagnoses for 7- and 30-day all-cause readmission rates in 2014. The top 10 primary diagnoses for 30-day readmission rates were congestive heart failure, schizophrenia, respiratory failure, alcohol related disorders, anemia, hypertension, diabetes, renal failure, chronic obstructive pulmonary disease, and implant/graft complications.[vi] The findings of our analysis are shown in Table 1.

Table 1: DRG Families with High Index Admission Counts and Significant Readmission Rates

DRG FamilyDRGsAdmit Count30-Day Readmission Rate15-Day Readmission Rate7-Day Readmission Rate
Heart Failure291, 292, 2934,86715.92%9.10%4.40%
Mental Disorders876, 880, 881, 882, 883, 884, 885, 886, 8872,84612.12%7.66%3.69%
Gastrointestinal Disease - Medical368, 369, 370, 371, 372, 373, 391, 392, 393, 394, 3953,95511.88%7.03%3.77%
Sepsis870, 871, 8728,94511.66%6.55%3.40%
Cardiac Arrhythmias308, 309, 3102,65710.80%7.00%3.73%
Pneumonia193, 194, 1953,4049.78%5.58%2.47%
Bowel, Rectal, Adhesion Surgery329, 330, 331, 332, 333, 334, 335, 336, 337, 344, 345, 346, 347, 348, 3492,6908.55%5.76%3.46%
Cerebrovascular Disease- Medical061, 062, 063, 064, 065, 066, 067, 068, 069, 070, 071, 0723,8418.23%5.08%2.71%
Spinal Procedures028, 029, 030, 453, 454, 455, 456, 457, 458, 459, 460, 471, 472, 473, 490, 491, 518, 519, 5203,5474.51%3.21%1.58%
Lower Extremity Arthroplasty466, 467, 468, 469, 47010,6862.88%1.71%1.02%

Same DRG Family Readmissions for the Top 10 DRG Families by Index Admission Count

Approximately 22% of the 7-, 15-, and 30-day readmissions had the same DRG family as the index admission. For the top 10 DRG families with the highest number of inpatient index admissions in 2016, the bar plot below (Figure 1) shows the percentage of unplanned readmissions within 7, 15, and 30 days with the same DRG family. The values range from very low (below 2%) for lower extremity arthroplasty to very high (80% and above) for mental disorders.

Figure 1: Same DRG Family Readmission Rates for the Top 10 DRG Families by Index Admission Count

Same MDC Readmissions for the Top 10 MDCs by Index Admission Count

We further examined the major diagnosis category (MDC) of the index admissions and readmissions. Figure 2 shows the percentage of readmissions within 7, 15, and 30 days for the top 10 MDCs with the highest number of index admissions in 2016.

  • For each of the top 10 MDCs, more than 25% of unplanned readmissions had the same MDC as the index admission.
  • About 80% of all unplanned mental health disorder readmissions had the same MDC as the index admission.
  • Musculoskeletal disorders index admissions had the lowest percentage of readmissions with the same MDC.

Figure 2: Same MDC Readmission Rates for the Top 10 MDCs by Index Admission Count

A high percentage of unplanned readmissions that we reviewed in our analysis had the same DRG family or MDC as their corresponding index admission. This observation might be attributed to possible gaps in the quality of care or the care delivery process. Plans and providers can choose from several inpatient quality measures provided in MedInsight to get an overview of their respective standing in comparison to national/regional benchmarks and design interventions to improve the quality of inpatient care.

References:

[1] HEDIS® is a registered trademark of the National Committee for Quality Assurance (NCQA).

[i] National Quality Forum, All-Cause Admissions and Readmissions Technical report, September, 2017. Available at: http://www.qualityforum.org/Publications/2017/09/All-Cause_Admissions_and_Readmissions_2017_Technical_Report.aspx.

[ii] T. Edelman, 2016. Center for Medicare Advocacy, Reducing Hospital Readmissions by Addressing the Causes, April 18, 2016. Available at: http://www.medicareadvocacy.org/reducing-hospital-readmissions-by-addressing-the-causes/.

[iii] Anika L. Hines et al, 2014. Conditions with the Largest Number of Adult Hospital Readmissions by Payer, 2011. Published: April, 2014. Available at: https://www.hcup-us.ahrq.gov/reports/statbriefs/sb172-Conditions-Readmissions-Payer.pdf.

[iv] Center for Medicare and Medicaid Services- Readmissions Reduction Program (HRRP). Last Modified: April, 2018. Available at: https://www.cms.gov/medicare/medicare-fee-for-service-payment/acuteinpatientpps/readmissions-reduction-program.html.

[v] Rachael Zuckerman et al, 2016. The New England Journal of Medicine, Readmissions, Observation, and the Hospital Readmissions Reduction Program. Published: April 21, 2016. Available at: https://www.nejm.org/doi/full/10.1056/NEJMsa1513024.

[vi] Kathryn R. Fingar et al, 2017. A Comparison of All-Cause 7-Day and 30-Day Readmissions, 2014. Published: October 2017. Available at: https://www.hcup-us.ahrq.gov/reports/statbriefs/sb230-7-Day-Versus-30-Day-Readmissions.jsp?utm_source=ahrq&utm_medium=en1&utm_term=&utm_content=1&utm_campaign=ahrq_en11_7_2017.

Drug Cost Trends 2015 to 2017 and the Top 20 Drugs

Sticky, Drug Utilization

The Centers for Medicare and Medicaid Services (CMS) identifies prescription drugs as one of the major components of U.S. national health expenditures, after hospital care and physician services.1 The US Assistant Secretary for Planning and Evaluation (ASPE ; principal advisor to the Secretary of the U.S. Department of Health and Human Services on policy development) estimates that retail prescription drug spending in the United States was about $328 billion in 2015, or 12 percent of overall personal health care services compared to 11 percent in 2013. This excludes non-retail prescription drug spending (spending by medical providers for drugs they provide directly to patients). The ASPE further observed that expenditures on prescription drugs are rising and would continue to rise in the coming years. 2

To identify the trend and pattern of drug spending, we conducted an analysis in one of MedInsight’s client databases with commercial population and included all the retail prescriptions. We analyzed the number of prescriptions (“scripts”) and the total cost for prescription drugs for three consecutive years (2015 to 2017) with respect to total healthcare spending and member enrollment data for the year. We identified an upward trend in the percentage of prescription spending over the years 2015 through 2017, as seen in Table 1. Despite a decrease in the average number of lives covered, prescription drug spending increased from approximately $1,123 million in 2015 to $1,229 million in 2017, a 9% increase.  Total healthcare spending increased 2% over the same time period. Prescription drug spending as a percentage of total healthcare spending also increased from 20% in 2015 to 21% in 2017. The per member per month (PMPM) prescription drug spending increased by 15% from $76 in 2015 to $87 in 2017.

Table 1: Retail Prescription drug spending, 2015-2017

 201520162017
Healthcare spending$5,596 million$5,388 million$5,732 million
Retail prescription drug spending$1,123 million$1,082 million$1,229 million
Prescription drugs as % of healthcare spending20.10%20.10%21.40%
Total member months14,751,51414,009,27114,070,879
PMPM prescription spending$76.14$77.24$87.32

The top 20 drugs by spending accounted for 31% of prescription drug costs in 2017, as shown in Figure 1. Table 2 lists the top twenty drugs in order of prescription cost and shows their frequency and unit cost.

Figure 1: Contribution of Top 20 Drugs to Total Pharmacy claims

Table 2: Top 20 high cost drugs- 2017

DrugPrescription Drug Spending (in millions)Relative to Total Drug SpendingNumber of scriptsRelative to Total Script CountAverage Cost per Script
Humira Pen$887%17,3000.15%$5,084
Enbrel Sureclick$323%6,9000.06%$4,674
Lantus Solostar$222%40,4000.34%$539
Novolog Flexpen$192%23,8000.20%$810
Victoza$192%22,5000.19%$829
Tecfidera$182%2,7000.02%$6,826
Novolog$171%20,6000.17%$827
Methylphenidate Hydrochloride$171%92,5000.78%$182
Humira$161%3,3000.03%$4,974
Copaxone$151%2,5000.02%$6,046
Gilenya$141%1,9000.02%$7,208
Enbrel$141%3,1000.03%$4,433
Vyvanse$131%47,9000.40%$280
Orkambi$131%6410.01%$20,439
Januvia$121%23,2000.20%$536
Invokana$121%20,9000.18%$552
Revlimid$111%8990.01%$12,478
Advair Diskus$111%26,5000.22%$417
Lyrica$111%20,7000.17%$532
Norditropin Flexpro$101%3,0000.03%$3,471
Total (Top 20 drugs)$38631%381,4003.21%$1,012
All other drugs$84369%11,515,50096.79%$73

These top 20 drugs include some high cost drugs in the following therapeutic areas: antidiabetics, analgesics – anti-Inflammatory, miscellaneous psychotherapeutic and neurological agents and drugs for ADHD, Narcolepsy/Obesity, as shown in Figure 2. The other categories in the list are miscellaneous respiratory agents, antiasthmatic and bronchodilator agents, anticonvulsants, assorted class of drugs and miscellaneous endocrine and metabolic agents.

Figure 2: Contributors to high drug cost

We further reviewed the data for drugs ranking high on total cost, average prescription cost (average cost per script) and utilization (number of scripts) for all three years, 2015 to 2017. We found that:

  • Humira Pen, an analgesic anti-inflammatory immunosuppressive drug used to treat rheumatoid arthritis, psoriatic arthritis, ankylosing arthritis, Crohn’s disease and plaque psoriasis had the highest total cost for all the three years. Our analysis shows an increase in both – the number of scripts as well as the average cost per script in a period of 2 years. The trend of prescription drug spend for Humira Pen is shown in Table 3.

Table 3: Pattern of cost and prescription for Humira Pen, 2015 to 2017

 20152017% increase (2015-2017)
Total Cost$43 million$88 million105%
Number of prescription11,60017,40050%
Average prescription Cost$3,714 $5,083 37%

  • The drug Orkambi, a respiratory agent used to treat Cystic fibrosis, had the highest average cost per script among top 20 high cost drugs in Table 2. The drug was approved by the US FDA for use in late 2015. A significant increase in its prescription and cost was observed in 2017 as shown in Table 4.

Table 4: Pattern of cost and prescription for Orkambi, 2015 to 2017

 20152017% increase (2015-2017)
Total Cost$2 million$13 million550%
Number of prescription117*641448%
Average prescription Cost$19,849 $20,439 3%

*The drug was approved in August 2015, hence the utilization pattern represents only partial year and cannot be comparable for year on year analysis.

  • Methylphenidate Hydrochloride, a drug used for ADHD, narcolepsy and obesity, had the highest number of prescription count over the years (highest in 2016 and 2017, second highest in 2015), as seen in Table 2. The pattern highlights an increase in the average prescription cost over a two-year period despite the decrease in the number of prescription count, as shown in Table 5.

Table 5: Pattern of cost and prescription for Methylphenidate Hydrochloride, 2015 to 2017 

 20152017% increase (2015-2017)
Total Cost$16 million$17 million6%
Number of prescriptions99,65692,475-7%
Average prescription Cost$160$18214%

Our analysis confirms findings from multiple published studies that drug spending continues to rise year after year. This continuous increase in total spending for prescription drugs results from an increase in the number of scripts, an increase in the unit cost per script, and the emergence of new drugs in the market. It is important to note that this analysis is based only on claims data from one large MedInsight client with commercially-insured lives, which may not be representative of all populations and may not identify all drug utilization and cost. Also, the days’ supply or the quantity of drug dispensed is not taken into consideration in this analysis.

The findings suggest that payers and policy makers need to continually monitor prescription drugs costs and utilization. Various publications 3 suggest some realistic short-term strategies to address high prices which include enforcing more stringent requirements for awarding and extending exclusivity rights; enhancing competition by ensuring timely generic drug availability; providing greater opportunities for meaningful price negotiation by governmental payers and employers; generating more evidence about comparative cost-effectiveness of therapeutic alternatives; and more effectively educating patients, prescribers, payers, and policy makers about these choices. Milliman does not endorse any specific policy proposal.

1 Centers for Medicare and Medicaid Services. National Health Expenditures Projections 2016-2025. Accessed at: https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/NationalHealthExpendData/Downloads/proj2016.pdf

2 Observations on Trends in Prescription Drug Spending. Department of Health and Human Services Office of the Assistant Secretary for Planning and Evaluation. March 8, 2016.

3 Kesselheim AS, Avorn J, Sarpatwari A. The high cost of prescription drugs in the United States. The Journal of American Medical Association. 2016;316(8):858-871






CMS Pathways to Success Final Rule

StickyMSSP

On December 31st, 2018, the Centers for Medicare and Medicaid Services (CMS) issued a final rule change to the current Medicare Shared Savings Program (MSSP) structure called “Pathways to Success”.  While this rule looks to accomplish additional goals, the main focus is to transition Accountable Care Organizations (ACOs) into risk-bearing contracts sooner.  The corresponding proposed rule was released on August 8, 2018, and more in-depth information from it can be found here.  This post will highlight the key changes in the final rule from the proposed rule.

Amongst the various changes that occurred between the proposed rule and the final rule is the shared savings rates for some of the levels within the BASIC track.  Levels A and B will have up to 40% shared savings rate instead of the originally proposed 25%, and Levels C and D will have up to a 50% shared savings rate instead of the proposed 30% and 40%, respectively.

Track progression for some scenarios also changed.  Low revenue ACOs are now defined as those historically having ACO participants’ Parts A and B FFS expenditures under 35% of total FFS expenditures for the ACO’s assigned beneficiaries instead of 25%. New low revenue ACOs will have the option to stay in Level B for an extra year if they commit to then be in Level E for the rest of their agreement period.  Additionally, high revenue ACOs who transitioned to Track 1+ in their current agreement period are eligible to remain in the BASIC track under Level E for an additional agreement period.

Another change that occurred with the final rule involves the benchmarking methodology.  The final rule presents a 15% regional adjustment to ACOs with historic expenditures above the regional benchmark in order to help these ACOs slowly phase into the use of regional adjustment factors.  This is a decreased factor from the proposed 25%.  Additionally, the proposed rule presented benchmark changes during agreement periods due to a shift in risk scores of up to 3% in either direction.  Under the final rule, the benchmark increase remains capped at 3%, whereas the benchmark decrease cap was removed.

In aggregate, CMS made numerous changes to the Pathways to Success between the proposed rule and final rule, many of which aid ACOs in the transition to more risk and new methodologies.  More information regarding the final rule and the implications it could have on ACOs can be found here.






The Value in Retrospective Data and Reporting in Provider Analytics

StickyData Confidence, Provider Efficiency

Technology continues to advance almost daily and data sources and real-time data access are rapidly expanding in healthcare.  These advances, combined with the shift from fee-for-service to value based reimbursement, mean healthcare analytics and reporting are changing too, and access to timely data is critical.  But do all these advances and changes call into question the utility and value of monthly retrospective data for provider analytics?

The short answer is no.

There are advantages to real-time (or near real-time) data and there are advantages to retrospective data. The two can, and should, peacefully co-exist and complement each other.

There are certain critical analytics, including financial analytics, which require more static retrospective data to support strategic and long term operational decisions.  Retrospective data also affords the opportunity to analyze and learn from how patients were managed in the past, which will benefit both providers and patients.

Examples include but are not limited to:

  • Utilization Efficiency Analysis and Cost Savings Opportunity Identification – This can include statistical analysis comparing against internal and industry accepted benchmarks and established guidelines.
  • Provider Efficiency – Using historical data to identify providers who are less or more efficient than their peers. This combined with the ability to drill into the claims detail will provide meaningful insight to both cost and utilization efficiencies while also considering the risk burden of the population.
  • Trend Reporting – Analyzing historical trends across time periods and targets. This could include cost, utilization, population changes, including member demographics and chronic populations.
  • Care Management Program Efficiencies – Understanding the effectiveness of programs from a cost and resource utilization perspective on the outcomes of patient health.
  • Population Health Management – Broad analytics that can identify the most impactablesegments of the population, identify avoidable waste, target and prioritize member populations for care management interventions, and target providers for education opportunities, as well as many other critical use cases.

In contrast, real-time (or near real-time) operational reporting or analytics is better suited to the following needs:

  • Ability to access the most up-to-date clinical information. – This means clinicians can assess patient-specific eligibility, gaps in care, lab data and other EMR related data.
  • Allowing users to be alerted to important data changes or where thresholds (KPIs) are exceeded in real-time.
  • Situations where users need to quickly view, analyze and drill down on real time displays, or dashboards in real-time.

Being able to support a broad range of use cases with both retrospective and real-time data allows an organization to be more competitive, adaptive and gain more strategic and operational insights to their performance efficiencies and improvement opportunities.  Clinicians benefit by having the ability to make better-informed decisions at the point of care, a key factor in providing the most appropriate care for their patients.






Antidepressant Monotherapy in Bipolar Disorder: Prescription Patterns

Sticky, Drug Utilization, Evidence Based Decision Making

Mental disorders are considered among the leading causes of ill health and disability worldwide. The United States spends an estimated $201 billion on mental disorders, making it the most costly medical condition in the country.1 Bipolar disorder is a common mental health condition, and the US has the highest lifetime rate of bipolar disorder at 4.4%, according to the National Institute of Mental Health.2 Bipolar disorder causes significant social, economic and emotional problems for affected individuals, their caregivers, and society.

Treatment and management for bipolar disorders are fairly well established. The American Psychiatric Association recommendation for treatment of bipolar disorder includes medications such as mood stabilizers, anticonvulsants or antidepressants depending on the specific symptoms.3 Guidance for the appropriate combination of more than one medication is well documented. For example, The International Society for Bipolar Disorders and the National Institute of Mental Health recommend that antidepressant monotherapy should not be used in bipolar disorder patients because of the potential risk for excessive mood elevation or mood destabilization. Antidepressants should be used sparingly, and if antidepressants are used at all, they should be combined with a mood stabilizer.2, 4 Taking an antidepressant without a mood stabilizer is likely to cause unfavorable (i.e., more challenging and costly) outcomes, including rapid and frequent mood swings and higher tendency for suicide.

To determine the extent of compliance with established treatment guidelines, we conducted a study of prescription patterns for antidepressant medications in patients with bipolar disorder using claims data for a US health plan with approximately 2 million members. To measure inappropriate and potentially wasteful antidepressant monotherapy, MedInsight’s Health Waste Calculator application was used on claims data for the time period between January 2015 and December 2015. A key focus of analysis was to identify any antidepressant monotherapy prescription which is considered inappropriate and potentially harmful.

The MedInsight Health Waste Calculator application identified 2,139 members with an antidepressant prescription and a corresponding diagnosis of a bipolar disorder. Of the total members, 1,711 members had a concurrent prescription of a mood stabilizer hence were considered as not receiving antidepressant monotherapy. A total of 438 members or 20% of the total members (with 16% of the total services) were identified who had antidepressant monotherapy only. The utilization/cost of antidepressant monotherapy for these members was identified as clinically inappropriate and potentially harmful which are considered wasteful by the MedInsight Health Waste Calculator application.

Figure 1: Profile of Antidepressant Prescription Services in Bipolar Disorder

Amongst the wasteful population, 60% of the population were prescribed Selective Serotonin Reuptake Inhibitors (SSRIs) or Serotonin-Norepinephrine Reuptake Inhibitors (SNRIs), which have been linked to heightened risk of subsequent mania (Figure 1). 5

Further analysis of the follow-up impact of antidepressant monotherapy highlighted that members in the Wasteful population had 50% more frequent psychotherapy visits than the Not Wasteful population.

Although claims data are limited in the ability to quantify downstream effects of inappropriate antidepressant monotherapy (such as mania), the results of this study confirm the prevalence of ineffective prescription patterns that are contrary to established treatment guidelines and result in wasteful spending and potential harm to patients.

1 Charles Roehrig. Mental Disorders Top The List Of The Most Costly Conditions In The United States. Available at https://www.healthaffairs.org/doi/full/10.1377/hlthaff.2015.1659

2 Bipolar Disease. National Institute of Mental Health. Available at https://www.nimh.nih.gov/health/statistics/bipolar-disorder.shtml

3 American Psychiatric Association. Bipolar Disorder. Available at https://www.psychiatry.org/patients-families/bipolar-disorders/what-are-bipolar-disorders

4 Pacchiarotti I, Thase M., et al, The International Society for Bipolar Disorders (ISBD) Task Force Report on Antidepressant Use in Bipolar Disorders. Am J Psychiatry. November 2013, 170:1249–1262.

5 Patel R, Reiss P, Shetty H, Broadbent M, Stewart R, McGuire P, Taylor M. Do antidepressants increase the risk of mania and bipolar disorder in people with depression? A retrospective electronic case register cohort study. Online First. Available at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4679886/

 






Communicating with Patients with Hearing Loss

StickyProvider Efficiency, Uncategorized

Poor compliance with healthcare provider advice for health and wellness actions is a growing issue. For older patients, hearing loss may contribute to lack of adherence with provider guidance.

BACKGROUND

The U.S. Census Bureau projects that by 2035, there will be 78 million people who are age 65 and older[i]. Research[ii] shows that nearly half of those 65 years of age and over have hearing loss, with age as the greatest predictor of hearing loss for adults between 20 and 69 years of age[iii]. Of these, about a third have significant hearing loss. However, most people over age 60 do not receive testing for hearing loss[iv].

Hearing loss negatively impacts multiple aspects of patients’ lives by creating communication barriers. A University of Manchester study[v] showed that people with hearing loss understand about 21 percent of speech. Use of either a hearing aid or speechreading[1] results in comprehension increasing to 64 percent. If both of these aids are used, comprehension increases to 90 percent; however, less than 30 percent of those 70 and older who could benefit from a hearing aid use them.[vi] Given the significant increase in understanding by use of hearing aids, why do people avoid the use of this readily accessible tool?

Aging is often stigmatized in U.S. culture, which can contribute to avoidance of addressing common physical deficits of aging, including hearing loss.  Hearing aids can be perceived as an outward, visible sign of aging and impairment. Johns Hopkins research shows the average adult waits 15 years to address hearing loss[vii]. Additionally, assistive devices are expensive, not covered by most insurances, and can be seen as a low priority expense for those on a fixed income. These factors contribute to patients not using hearing aids and thereby not understanding what their provider is communicating in the office visit.  Other negative impacts to health include a link between hearing loss and increased falls, physical and cognitive decline, brain shrinkage, and increased risk of dementia[viii].

APPROACHES

Once impaired hearing has been identified and where providers are part of a larger integrated health system or practice, a patient Electronic Medical Record or database and tools may be available to provide reporting that identifies patients with hearing loss or risk factors for impaired hearing. Desktop applications or manual processes are alternative options, although less efficient. Examples of data that can be informative include patient age, history of environmental exposure to loud noise, or existing sensory deficits. Once identified, this information can generate alerts for regular hearing screening.

Alerting practice staff with a chart or system flag for patients with a hearing deficit or risk factors can provide the opportunity for advanced planning for office visits. For example, preparing written materials for those with hearing loss with a larger font can be useful for those with hearing loss and vision deficits as well. Content might include topics such as:

  • May I see your insurance information and picture identification?
  • Do you need a sign language interpreter?
  • Please have a seat and we will let you know when your provider is ready to see you.

Training staff on effective communication techniques for those with sensory deficits can be very helpful for both patient and staff. Approaches include directly facing the person, speaking clearly and a bit more slowly than usual, minimizing background noise, and providing written materials, such as a summary of the treatment plan written at a literacy level appropriate for broad comprehension. Small whiteboards can be helpful to augment communication for both the provider and the hard-of-hearing patient.

By using patient data for proactive identification, the patient visit can result in comprehension of the recommended plan of care and actions by use of planned techniques to address hearing communication barriers.

[1] Speechreading is defined as the act or process of determining the intended meaning of a speaker by utilizing all visual clues accompanying speech attempts, as lip movements, facial expressions, and bodily gestures, used especially by people with impaired hearing. Source: https:// dictionary.com. Accessed 10/17/2018.

[i]United States Census Bureau. (March 13, 2018). Older People Projected to Outnumber Children for the First Time in U.S. History. Retrieved on 5/28/2018 from https://www.census.gov/newsroom/press-releases/2018/cb18-41-population-projections.html.

[ii]StratisHealth. (September 2011). Deaf and Hard-of-Hearing Minnesotans. Retrieved on 5/9/2018 from http://culturecareconnection.org/documents/InformationSheet_Deaf.pdf.

[iii]National Institutes of Health, National Institute on Deafness and other Communication Disorders. (December 15, 2016). Quick Statistics About Hearing. Retrieved on 5/25/2018 from https://www.nidcd.nih.gov/health/statistics/quick-statistics-hearing.

[iv]StratisHealth. Ibid.

[v]StratisHealth. Ibid.

[vi]National Institutes of Health, National Institute on Deafness and other Communication Disorders. Ibid.

[vii] Minnesota Commission of Deaf, Deafblind and Hard of Hearing Minnesotans. (December 14, 2017). First-of-its-Kind Pilot Program Aims to Increase Awareness of Hearing Loss Among Older Minnesotans. Retrieved on 5/9/2018 from https://mn.gov/deaf-commission/news/?id=1063-320251.

[viii]Johns Hopkins Medicine. (January 22, 2014). Hearing Loss Linked to Accelerated Brain Tissue Loss. Retrieved on 5/25/2018 from https://www.hopkinsmedicine.org/news/media/releases/hearing_loss_linked_to_accelerated_brain_tissue_loss_.

 






Data Confidence for ACOs

StickyACO, Data Confidence

All healthcare data has data quality challenges. However, as Accountable Care Organizations (ACOs) have taken on more risk and are working on improving care processes data quality has become a more important issue. Below are some common data quality issues that ACOs face and some of the solutions ACOs can use to build confidence in their data.

Incomplete Provider Data

Provider level analysis is extremely important for ACOs. ACOs need to know which providers both within and outside the ACO network are providing services to patients attributed to the ACO. This is important not only to work towards bringing a larger percentage of services in-network (leakage management), but also for quality and efficiency improvement. There are multiple issues with provider data from payer data sources that can make it difficult to correctly identify in and out of network providers. These include:

  • Claims missing complete provider information. Medical claims need to have both the billing and servicing/rendering providers listed, and pharmacy claims need to have both the prescriber and pharmacy listed. It is critical that ACOs work with their data suppliers to ensure that these multiple provider fields are complete on claims.
  • Custom provider identifiers. Some data sources use custom provider identifiers, instead of National Provider Identifiers (NPIs). To perform analysis across data sources from different suppliers, any custom identifiers need to be cross-walked and mapped to a consistent standard such as NPI. For facilities or large practices, which generally have multiple NPIs or may use alternative identifiers, it is important to roll up the identifiers present in the data for analytic purposes.

Incomplete Financial Fields

Data suppliers often remove or mask financial data to ensure that provider reimbursement terms for providers outside of the ACO network remain confidential. Financial values are useful in ensuring that the data is complete, and are necessary to determine the magnitude of differences in resource cost between different services. A variety of tools, including MedInsight Global RVUs, . The conversion factor can be derived using benchmarks or by dividing the total cost of the contract by total RVUs. While this does not replicate actual unit cost it can provide reasonable approximations ACOs can use to make decisions.

Incomplete Diagnosis Coding

Many ACO contracts include financial parameters that are risk adjusted and it is important to have all diagnosis codes available for analysis, as these diagnoses drive risk scores. To test for the quality of the diagnosis coding in a given data source, users can audit both the number of codes per claims and the ACOs can use benchmarks to ensure that their claims have reasonable population of diagnosis codes and use other tools to review the consistency of diagnosis coding over time.

Completeness of Electronic Medical Record (EMR) Data

More analysis is being done using EMR data and combined EMR/claims data. In order to appropriately incorporate EMR data, the ACO needs to ascertain how much of a patient’s clinical care was delivered by providers using that EMR system, as clinical data from providers using alternative EMR systems would not be included in the data. It’s also important to gauge the relative quality of the EMR fields. An example is to measure the completeness of fields in the encounter file.

These are a few examples of the importance of ACO data quality and how ACOs can use analytics tools to improve data quality. As healthcare analytics continue to play an increasingly important role in decision-making, utilization and cost, ACOs will need to work closely with their data suppliers to continue to improve data quality.






ACO Rule Change: Impact of Proposed Rule on Track 1 and New MSSP ACOs/The BASIC Track

Sticky, , ACO, MSSP

The new rule proposed by the Centers for Medicare & Medicaid Services (CMS) on August 9, 2018 would result in significant changes to the way the Medicare Shared Savings Program (MSSP) works for new Accountable Care Organizations (ACOs) and those currently in a one-sided risk model. Under the current program, ACOs can enter the MSSP under an upside-only risk arrangement referred to as a Track 1 ACO for up to two three-year agreement periods. After that time the ACO is required to transition to a two-sided risk agreement or exit the program. However, the proposed rule would end the Track 1 program and replace it with a new five-year agreement called the BASIC track. This new track would transition from one-sided to two-sided risk throughout the agreement period. This change has the potential to significantly impact ACOs that were planning on entering the MSSP or renewing in Track 1 for their second agreement period by exposing them to two-sided risk earlier than expected.

The proposed BASIC track would consist of five levels that gradually increase the amount of risk and reward taken by the ACO. The first two levels, called “Level A” and “Level B” are identical one-sided models designed to allow new ACOs to establish themselves and gain familiarity with the program. Levels A and B are structured similarly to the current Track 1 program. After that there are three two-sided risk levels with progressively increasing amounts of risk and savings potential, called “Level C”, “Level D”, and “Level E”. Level E is similar to the current MSSP Track 1+ program, while Levels C and D have lower shared risk and savings potential. Each year, the ACO is automatically moved up one level in the progression until reaching Level E, where it would remain until the end of the five-year agreement period. The ACO can also choose to move into a higher level in the progression in any given year; however, once an ACO has moved up the automatic progression continues from its newly-selected level– the ACO cannot reverse or stop the progression. The only level an ACO can remain in for more than one year is Level E at the end of the progression, where the ACO would remain until the expiration of the five-year agreement.

Whether an ACO can enter into the BASIC track, and where they are eligible to enter, depends on its makeup and history in the MSSP. New ACOs are eligible to enter the BASIC track at Level A. New ACOs are defined as those where fewer than 50 percent of its participating physicians have recent experience as part of a Track 1 ACO. Current Track 1 ACOs and re-entering ACOs, those where greater than 50 percent of the participants have recent experience in a Track 1 ACO, are eligible to enter the BASIC track; however, they are required to enter at Level B. This restricts more experienced ACOs to a single year of one-sided risk as opposed to the two years that new ACOs can choose. The ability to renew into the BASIC track for a second agreement period depends on if the ACO is determined to be high or low revenue.  More information about what qualifies an ACO as high or low revenue can be found here (link to other blog).  Low-revenue ACOs can enter a second BASIC agreement period limited to Level E only. High revenue ACOs are required to move up to the ENHANCED track or exit the program at the conclusion of their first agreement period.  Low revenue ACOs also have a lower loss sharing limit.

The proposed changes to the MSSP program have the potential to significantly impact both existing ACOs and those considering entering the program. For more detailed information on this proposed rule and its potential impact please see please click here.






ACO Rule Change: Pathway to Success for Current Performance-Based Risk ACOs

Sticky, , ACO, MSSP

The Centers for Medicare and Medicaid Services (CMS) proposed a new rule in August of 2018 to change the structure of the Medicare Shared Savings Program (MSSP).  This will affect current and future ACOs whether they participate in a track with one-sided or two-sided risk.  Currently, three MSSP tracks incorporate performance-based, or two-sided, risk: Track 1+, Track 2, and Track 3.  Performance-based risk means these ACOs share in a larger portion of any savings below the benchmark, but they also share in losses if their expenditures surpass the benchmark.  In the proposed new rule, ACOs will choose between two tracks:  BASIC and ENHANCED.  ACOs that have participated in performance-based risk tracks in the past are limited to participation in the ENHANCED track or, if they are determined to be low-revenue, ACOs may enter the highest level of risk/reward in the BASIC track.

In order for an ACO that has previously participated in performance-based risk to enter the BASIC track, it must be considered low revenue.  An ACO is determined to be high or low revenue based on the percentage of the ACO’s assigned beneficiaries’ total fee-for-service expenditures that are paid to physicians participating in the ACO. If the ACO participants’ total FFS revenue (including revenues for beneficiaries not assigned to the ACO) is more than 25% of the ACO’s assigned beneficiaries’ total FFS expenditures, the ACO is classified as high revenue, otherwise it is low revenue. In general, high-revenue ACOs will be those that include facilities, resulting in a higher degree of control over total expenditures for its assigned beneficiaries.  If an ACO is determined to be low revenue, it can choose to participate in the highest risk/reward level of the BASIC track:  Level E.  Level E of the BASIC track contains the same level of risk and reward as the current Track 1+.  The other option available to ACOs that previously participated in performance-based risk is the ENHANCED track.  High-revenue ACOs must enter this track, and low-revenue ACOs will be given the option to enter this track.   ACOs that choose to participate in the ENHANCED track will receive the same level of risk and rewards as current Track 3 ACOs.

According to CMS, ACOs in performance-based risk models are improving quality of care and perform better over time than ACOs in one-sided risk models.   CMS believes the new tracks will encourage ACOs to participate in performance-based risk models, while continuing to reward ACOs that take on the two-sided risk with new tools and a greater reward for beating the benchmark.  For more detailed information on this proposed rule and its potential impact please click here.






ACO Rule Change: Non-Track Changes

Sticky, , ACO, MSSP

On August 9th, 2018, the Centers for Medicare and Medicaid Services (CMS) announced a proposed rule change to the current Medicare Shared Savings Program (MSSP) structure.  Along with updates to the tracks that are available to Accountable Care Organizations (ACOs) through the program, a number of new features and methodologies would be introduced under the new rule.  These new aspects are intended to help ACOs innovate and be successful in care coordination, promote beneficiary engagement, encourage program integrity, and improve care.

The proposed rule would introduce a more rigorous benchmarking methodology for ACOs.  Currently, the benchmarking methodology only incorporates regional Fee for Service (FFS) expenditure trend factors when an ACO is on their second or subsequent 3-year agreement period.  Under the new rule, the regional factors will be applied to all of the agreement periods, which will become 5 years in length.  Additionally, benchmarks can be adjusted in either direction up to 3% per year during an agreement period in order to reflect changes in the population’s health status.

ACOs will be given the option for each performance year whether they want prospective member assignment or preliminary prospective member assignment with a retrospective reconciliation.  Currently, the methodology is determined based on the track that an ACO is enrolled in.  There will also be other changes to the beneficiary assignment methodology.  The definition of primary care service will be expanded, and beneficiaries may be given the option to opt in to an ACO.  Additionally, voluntary alignment would continue to be allowed and allowing the member to designate any physician regardless of specialty.

Beginning January 1, 2020 ACOs would also potentially be eligible to have participating physicians receive payments for using telehealth services.  Physicians of ACOs that are in a 2-sided agreement with prospective assignment would be eligible for this.  Additionally, use of the SNF 3-day waiver would be expanded to apply to any ACO in a 2-sided arrangement and to critical access hospitals and small rural hospitals part of these ACOs, and 2-sided ACOs could provide incentive payments up to $20 to beneficiaries for each related primary care service they receive.

Under the proposed rule, CMS would help provide additional resources and features to help ACOs succeed.  Beyond the track structure change that would occur, various aspects will be updated, including new benchmark calculations and support of ACO programs.  For more information about the proposed rule and the various aspects involved, see the article found here.






New ACO Rule Change

Sticky, , ACO, MSSP

On August 8, 2018, the Centers for Medicare and Medicaid Services (CMS) released a sweeping proposed regulation that, if enacted, will significantly change the Medicare Shared Savings Program (MSSP). The proposed regulation, titled “Pathways to Success,” accelerates the path for accountable care organizations (ACOs) to participate in shared risk arrangements while simultaneously softening key provisions, allowing lower revenue ACOs to participate with reduced total financial risk.

Currently the MSSP is composed of four different tracks, all with different shared saving and shared loss structures, that ACOs can enter into for a three-year agreement period.  Under the proposed rule, there would be two different tracks:  BASIC and ENHANCED each with a five-year agreement period.  Under the BASIC track, there are five different levels with various risk and reward levels that ACOs transition through during their five-year agreement period.  These various levels range from a one-sided risk model to a model that closely aligns with the current Track 1+.  The new ENHANCED track is largely the same as the existing Track 3.

In addition to these new tracks, the proposed rule affects other key program provisions.  There will also be a variety of new features available to ACOs through the proposed rule.  ACOs will be able to choose every year whether to receive prospective assignment or preliminary prospective assignment with retrospective reconciliation.  Additionally, beneficiary incentive programs, expanded telehealth services, and three-day SNF rule waivers will be supported for ACOs who are in two-sided risk arrangements under the new tracks.

If approved, the rule proposed by CMS would go into effect on July 1, 2019 for a one-time six-month performance year that would aid in the transition.  For more information about the proposed rule, reference the articles found here.






Beta-Blocker Therapy: Why Adherence Matters

Sticky, Avoidable Admissions, Drug Utilization, Evidence Based Decision Making, Value-Based Care

Heart failure (HF) is a global pandemic affecting at least 26 million people worldwide that is increasing in prevalence. Currently about 6  million people in the United States (US) have heart failure, and prevalence is projected to increase to more than 8 million people by 2030 [i],[ii] Direct medical costs for HF in the US are expected to increase from $20.9 billion in 2012 to $53.1 billion in 2030.[iii]

Beta-blocker therapy is generally accepted as a key therapy for patients diagnosed with chronic systolic heart failure.[iv] The American College of Cardiology Foundation (ACCF) and the American Heart Association (AHA) Guideline (2013)  for the Management of Heart Failure recommends long-term beta-blocker treatment on the basis that it can lessen the symptoms of HF, improve the patient’s clinical status, and enhance the patient’s overall sense of well-being. In addition, beta blockers can reduce the risk of death and the combined risk of death or hospitalization. Use of 1 of the 3 beta blockers proven to reduce mortality (e.g., bisoprolol, carvedilol, and sustained-release metoprolol succinate) is recommended for all patients with current or prior symptoms of HF with reduced ejection fraction (<40 percent), unless contraindicated, to reduce morbidity and mortality. [v] Prior research studies, meta-analysis and randomized clinical trials have established the efficacy of beta blockers in terms of beneficial impact on clinical outcomes in patients with current or prior symptoms of heart failure[vi],[vii],[viii],[ix]  Despite these documented benefits, adherence to beta-blocker medication is found to be lowest amongst all medications used in HF.[x]

To explore actual compliance with these guidelines and their impact on inpatient and emergency department (ED) stays, we analyzed a MedInsight client data for three calendar years (2012-2014). The target population was members aged 18 years and older, diagnosed with Congestive Heart Failure (CHF) and receiving beta-blocker therapy.  Medication adherence was measured as the percentage of prescribed doses of beta- blocker or beta-blocker combination medication effectively taken by the patient over a specified period. This population was grouped into three medication adherence categories i.e. below 50 percent, 50-75 percent, and above 75 percent.

Results

Figure 1 presents the demographic findings of our analysis. The data indicated that elderly members (ages 65 +) were found to be more adherent to beta-blockers therapy as compared to the younger population (ages 19-64 years). No substantial variation in medication adherence was observed between male and female members.

We performed further statistical analysis to explore the relationship of beta blockers adherence to all cause emergency department (ED) and inpatient (IP) utilization. The ED and IP Utilization rates (per 1,000 patient years) were adjusted using Milliman Advanced Risk Adjusters (MARA) software. Figure 2 and Figure 3 highlight the following observations on ED and IP utilization rates in relation to adherence:

Observations by Age Band:

  • The relationship between adherence and adjusted ED rates was a strong inverse relationship for working age members (19-64) and a more moderate inverse relationship for members aged 65+. Adjusted inpatient rates for members 19-64 with adherence above 75% were about 60% of the rate seen for members with adherence rate below 50%. Adjusted inpatient rates for members aged 65+ with adherence above 75% were similar to working age members, but higher for aged 65+ members with poor adherence.

Observations by Gender:

  • The relationship between adherence and adjusted ED rates and adjusted IP rates was a strong inverse relationship both males and females. Male members with adherence below 50% had adjusted IP rates higher than those of females with poor adherence. Adjusted IP rates for males with adherence of 50%+ were similar to female members matched for adherence.

Observations by Line of Business:

  • The relationship between adherence and adjusted ED rates was not consistent by line of business. Commercial members showed little difference in adjusted ED rates with better adherence, while adjusted ED rates for Medicare members did show a decline with better adherence. Medicaid members only showed reduced adjusted ER rates in the above 75% adherence group.
  • The relationship between adherence and adjusted IP rates was similar for each line of business. There was little difference between members with poor adherence and moderate adherence, but members in the above 75% adherence group had substantially lower adjusted IP rates in each line of business.

Discussion

The above results show a statistically significant relationship between adherence to beta blocker therapy and all-cause ED and IP utilization. The results suggest that continued focus on beta-blocker therapy for members diagnosed with HF may result in reductions in ED and IP utilization. Although the analysis was based on limited administrative data, the statistical significance of the results highlight adherence to beta blockers as a target for interventions to reduce emergency department and inpatient utilization.

[i] Savarese G, Lund LH. Global Public Health Burden of Heart Failure. Cardiac Failure Review.2017;3(1):7-11.

Avaiable at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5494150/#

[ii] Heart Disease and Stroke Statistics-2016 Update: A Report from the American Heart Association. Writing Group Members, Mozaffarian D, Benjamin EJ, Go AS, Arnett DK, et al. American Heart Association Statistics Committee; Stroke Statistics Subcommittee. Circulation. 2016 Jan 26; 133(4):e38-360. Available at: http://circ.ahajournals.org/content/133/4/e38.long

[iii] Heidenreich PA, Albert NM, Allen LA, et al. Forecasting the impact of heart failure in the United States: a policy statement from the American Heart Association. Circulation: Heart Failure. 2013;6(3):606-619. Available at: http://circheartfailure.ahajournals.org/content/6/3/606.full.pdf?download=true

[iv] Jondeau G, and Milleron O. Beta-Blockers in Acute Heart Failure- Do They Cause Harm? JACC: Heart Failure Aug 2015, 3 (8) 654-656. Available at: http://www.heartfailure.onlinejacc.org/content/3/8/654

[v] Yancy CW, et al. American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. 2013 ACCF/AHA guideline for the management of heart failure: a report of the American College of Cardiology Foundation/American Heart Association Task Force on practice guidelines. Circulation. 2013 Oct 15; 128(16):e240-327. Available at: http://circ.ahajournals.org/content/128/16/e240.full.pdf+html

[vi] Bristow MR. Treatment of chronic heart failure with β-adrenergic receptor antagonists: a convergence of receptor pharmacology and clinical cardiology. Circ Res2011; 109:1176-94. Available at: http://circres.ahajournals.org/content/109/10/1176.full

[vii] Manurung D, Trisnohadi HB. Beta blockers for congestive heart failure. Acta Med Indones. 2007 Jan-Mar; 39(1):44-8. Available at: http://www.inaactamedica.org/archives/2007/17297209.pdf

[viii] Heidenreich P, Lee T, Massie B. Effect of Beta-Blockade on Mortality in Patients with Heart Failure: A Meta-Analysis of Randomized Clinical Trials 1. J Am Coll Cardiol. 1997; 30(1):27-34. Available at: http://content.onlinejacc.org/article.aspx?articleid=1124060

[ix] Hernandez AF, et al. Clinical effectiveness of beta-blockers in heart failure: findings from the OPTIMIZE-HF (Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure) Registry. J Am Coll Cardiol 2009; 53:184-192. Available at: http://content.onlinejacc.org/article.aspx?articleid=1139342

[x] Viana M, Laszczynska O, et al. Medication adherence to specific drug classes in chronic heart failure. J Manag Care Spec Pharm. 2014 Oct;20(10):1018-26. Available at: http://www.amcp.org/WorkArea/DownloadAsset.aspx?id=18572






Milliman and MedInsight Twitter Chat on Healthcare Waste

StickyEvidence Based Decision Making, Health Waste Calculator, Value-Based Care, Waste

Last week on Twitter, we hosted a Twitter chat with Milliman about healthcare waste in the industry.

MedInsight’s Rich Moyer and Marcos Dachary were joined by Harvard Medical School’s Dr. Michael Chernew and VBID Health’s Dr. Mark Fendrick to answer questions posed by Milliman. Other healthcare professionals joined the discussion asking questions and offering ideas using the hashtag #healthwaste.

This was the first time Rich Moyer or Marcos Dachary had ever participated in an event like this. And over the course of an hour, we saw a rich discussion about healthcare waste developing. This was a great opportunity for others in the industry to be able to ask questions directly to the healthcare experts on this panel.

If our #healthwaste Twitter chat interested you, be sure to check out Milliman’s Critical Point podcast on healthcare waste to hear more from Marcos Dachary and MedInsight’s Dr. David Mirkin.

We had a great time discussing healthcare waste with the community and look forward to doing another Twitter chat on a new topic in the future!






Annual Physical: An Endless Debate

Sticky, Analytics, Healthcare Analytics, Preventive Care, Value-Based Care

There has been an ongoing debate in United States (US) on the value of the annual comprehensive physical examination. One school of thought recommends elimination of the annual comprehensive physical exam due to lack of value. Choosing Wisely 2014 states that healthy people often do not need annual comprehensive physicals[i], and systematic reviews[ii] back this up by finding these examinations do not significantly reduce morbidity or mortality. In addition to the fee for the physician, billions of dollar are spent on potentially unnecessary tests ordered during annual comprehensive physical exams and follow up diagnostics.

A second school of thought defends the importance of annual comprehensive physical exams considering these critical to maintain and reaffirm a patient-physician relationship. Arguments have been made to improve the annual comprehensive physical exams instead of eliminating it by using a multidisciplinary team based approach, providing physicians with time for a more personalized in-depth review and a more comprehensive and satisfying overall experience for the patient.[iii]

We wanted to understand if this ongoing debate has had an impact on the rate of annual physical exams. Using data from a MedInsight client with 12 million members, we reviewed recent trends in annual physical exams[*] over a five (5) year period 2012-2016. We restricted our analysis to healthy members between 18 and 64 years of age who did not have a chronic condition in a given year (defined as being in a non-chronic CCHG (Chronic Conditions Hierarchical Group)).

We had the following findings:

  • About 7 percent of members aged 18-64 years are getting annual physical exams.
  • Rate of annual physical visits over the five-year period was relatively flat for all age groups; the 51-64 age group had a slightly increasing trend in the rate of annual physical visits.
  • No substantial difference in annual physical visits was observed between male and female members.

Our analysis shows a recent decline in the number of people getting an annual physical examination. The reason for this decline could be the possible harms associated with these health checks such as fear of false positive results, over diagnosis, unwarranted treatment, follow-ups, and adverse emotional stress due to false labelling.2 Some individuals also find it a waste of their productive time waiting for the care; which could have been utilized better elsewhere. We assume these likely factors may be facilitating drop in numbers of annual physical exam visits.[i]

Systematic review and meta-analysis have also provided evidence that annual physical examination does not reduce morbidity or mortality for any cause, although they increased the number of new diagnoses. Replacing the annual physical examination with a more personalized health review involving more discussion and limited examination would be more acceptable to the patients. This would not only suffice the purpose of preventive health review; but also help in improving the physician-patient relationship.[ii]

Limitations

The results and interpretations are solely based on the administrative claim data.

 

References: 

[*] An annual physical exam was identified using procedure codes 99395 and 99396 when reported along with ICD9 diagnosis code V700 or ICD10 diagnosis code Z0000.

[i] Consumer Reports Health, 2014. Choosing wisely.Health checkups, when you need them and when you don’t. Available at: http://www.choosingwisely.org/patient-resources/health-checkups/

[ii] Krogsboll LT, Jorgensen KJ, Gronhoj Larsen C, Gotzsche PC. General health checks in adults for reducing morbidity and mortality from disease: Cochrane systematic review and meta-analysis. BMJ 2012;345: e7191. Available at: http://www.bmj.com/content/345/bmj.e7191

[iii] Allan H. Goroll, M.D., 2015. Toward Trusting Therapeutic Relationships — In Favor of the Annual Physical. Available at: http://www.nejm.org/doi/full/10.1056/NEJMp1508270?query=featured_home

[iv] Mehrotra, Ateev and Prochazka, Allan. Improving Value in Health Care — Against the Annual Physical. New England Journal of Medicine. October 2015; 373:1485-1487. Available at : https://www.nejm.org/doi/full/10.1056/NEJMp1507485

[v] Bloomfield HE, Wilt TJ. Evidence Brief: Role of the Annual Comprehensive Physical Examination in the Asymptomatic Adult, VA-ESP Project #09-009; 2011. Available at: https://www.ncbi.nlm.nih.gov/books/NBK82767/






Join Us for Our Upcoming Twitter Chat!

StickyHealth Waste Calculator, Healthcare Analytics, Value-Based Care, Waste

In recent years, few issues have been spoken about more than the need to curb unnecessary healthcare spending across the United States. As discussions continue at various parts of the government, Milliman invites you to join our #healthwaste Twitter chat. The chat features Milliman MedInsight consultants Rich Moyer (@RichCMoyer) and Marcos Dachary (@MarcosDachary), who will discuss ways in which the healthcare industry can identify wasteful spending and what can be done to address the underlying issues. Rich and Marcos will be joined by two of the premier experts in this area, Dr. Mark Fendrick of Value-Based Insurance Design Health (VBID), and Dr. Michael Chernew of Harvard Medical School.

What: #HealthWaste

Where: Twitter

When: Thursday, August 2, at 10 a.m. PST/1 p.m. EST

Topic: Identifying wasteful spending in healthcare

Moderator: @MillimanHealth

Rules of engagement

  • To participate in the chat, follow the hashtag #healthwaste.
  • Answer Q1, Q2, Q3… with A1, A2, A3….
  • Remember to include the hashtag #healthwaste in all your tweets.
  • If you are new to Twitter chats, considering using TweetChat.com.






Anesthesia in Screening Colonoscopy

Sticky, , Drug Utilization, Evidence Based Decision Making, Healthcare Analytics

The U.S. Preventive Services Task Force recommends colorectal cancer screening for men and women aged 50–75.[i] For colorectal cancer (CRC) screening, colonoscopy has been accepted as the most effective method, with the adenoma detection rate (ADR) remaining one of the most important measures of a quality colonoscopy. Factors shown to affect adenoma and polyp detection rates (PDR) include: use of sedation along with other factors like adequacy of bowel preparation, cecal intubation rate, withdrawal time, image enhancements, and the performing endoscopist.[ii]

Traditionally, conscious (moderate) sedation (CS) using midazolam and an opioid has been used in screening colonoscopies. Until recent Medicare reimbursement changes, for most colonoscopy procedures CS has been considered an inherent part of the procedure, not to be reported and billed separately except for the situation when moderate sedation is provided by a second physician in a facility setting. However over the past decade, newer anesthesia options have become available, and based on a nationwide survey distributed in 2004 to the members of the American College of Gastroenterology, it was found that approximately one quarter of patients now undergo deep sedation with propofol (propofol sedation (PS)).[iii] Although Propofol sedation (PS) has led to detection of more advanced polyps, a retrospective analysis of 699 consecutive patients who underwent inpatient screening colonoscopies at one academic inpatient centre found no significant difference between ADR or location of detected adenomas between the CS and PS groups.2

An upward trend in separate anesthesia by anesthesiologist for screening colonoscopy was found in those with higher comorbidity according to a study conducted on National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database.3 To get an insight into how frequently separate anesthesia services are used in screening colonoscopy, and the associated comorbidities of the patient, we performed an analysis on Commercial plan members aged 50-75 years using a MedInsight data for year 2012. We identified the number of comorbidities the patient had, using diagnosis codes for the conditions used in a similar study.[iv] A comorbidity is assigned only when the member had at least two claims 30 days apart with the corresponding diagnosis code in the two years preceding colonoscopy.

A total of 25,055 screening colonoscopies were performed in the year 2012 with the following characteristics.

Separate anesthesia was given in 3,465 (13.83%) of screening colonoscopies with a breakdown by comorbidity count and gender as given below:

Out of all screening colonoscopies performed under separate anesthesia (3,465) a staggering 62.11% (2,152) had no associated comorbidities and the additional costs for anesthesia services alone in such cases amounted to $ 796,134.

Although separate anesthesia for screening colonoscopies does not improve the ADR and most of the times has not been covered by the insurance plan, it is associated with increased patient satisfaction and reduced pain levels.   As a result it has been increasingly used for screening colonoscopy. The Centers for Medicare and Medicaid Services, as a provision of the Affordable Care Act revised the definition of “colorectal cancer screening tests” beginning January 1, 2017 to include anesthesia that is separately furnished in conjunction with screening colonoscopies.[i] It will be interesting to look at the utilization and financial impact of this ruling on commercial and Medicare payers in the near future.

[i] United States Preventive Services Task Force. Recommendations for colorectal cancer screening, 2008. Available at:  https://www.uspreventiveservicestaskforce.org/Page/Document/RecommendationStatementFinal/colorectal-cancer-screening2#tab

[ii] Nakshabendi R, Berry AC, Munoz JC, John BK. Choice of sedation and its impact on adenoma detection rate in screening colonoscopies. Annals of Gastroenterology : Quarterly Publication of the Hellenic Society of Gastroenterology. 2016;29(1):50-55. Available at: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4700847/

[iii] Khiani VS, Soulos P, Gancayco J, Gross CP. Anesthesiologist Involvement in Screening Colonoscopy: Temporal Trends and Cost Implications in the Medicare Population. Clinical gastroenterology and hepatology : the official clinical practice journal of the American Gastroenterological Association. 2012;10(1):58-64.e1. doi:10.1016/j.cgh.2011.07.005. Available at: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3214600/#R6

[iv] Comorbidity Measures for Use with Administrative Data Author(s): Anne Elixhauser, Claudia Steiner, D. Robert Harris and Rosanna M. Coffey Source: Medical Care, Vol. 36, No. 1 (Jan., 1998), pp. 8-27. Available at: http://czresearch.com/dropbox/Elixhauser_MedCare_1998v36p8.pdf

[v] Federal Register, The daily Journal of the United States Government. Medicare Program; Revisions to Payment Policies Under the Physician Fee Schedule Clinical Laboratory Fee Schedule, Access to Identifiable Data for the Center for Medicare and Medicaid Innovation Models & Other Revisions to Part B for CY 2015. Available at: https://www.federalregister.gov/articles/2014/11/13/2014-26183/medicare-program-revisions-to-payment-policies-under-the-physician-fee-schedule-clinical-laboratory






Prescription Claims Data: Can It Be Used To Understand Opioid Prescription Trends?

Sticky, Drug Utilization, Opioids

Since 2000, the rate of deaths from drug overdoses has increased 137%, including a 200% increase in the rate of overdose deaths involving opioids (i.e. opioid pain relievers and heroin).1 In fact, almost 40% of all opioid overdose deaths involved a prescription opioid.2 In addition to the loss of life, opioid overdoses have resulted in an estimated $504 billion in economic cost during 2015.3

Historically, opioid addiction was an issue discussed in the context of illegal drugs such as heroin and other synthetic opioids. However, in the last few years this issue has expanded into the realm of prescription opioids and become a nationwide crisis that can no longer be ignored.

The key question for policy makers and other health care stakeholders is, “What can we do to mitigate this crisis?” The solution needs to address patient behaviors that lead to overdoses and the prescription of unnecessary opioids by healthcare professionals. There isn’t a single right answer that will address both of these underlying issues regarding patient behavior and unnecessary prescriptions. Each and every stakeholder including policy makers, payers, providers, patients, family members, and others (e.g., law enforcement) have a role to play in the solution.

Data analytics can be a key resource in understanding and responding to this crisis. For example, the analysis of claims can help payers and providers gain an in-depth understanding of the utilization patterns of prescribed opioids. Specifically, claims data for prescription opioids should be analyzed across three dimensions:

  1. Providers: What are the key characteristics of providers who are potentially overprescribing opioids and who are these providers?
  2. Prescriptions: What is the trend and utilization pattern including potential unnecessary use of opioid prescriptions in a given population?
  3. Members: What do we understand about the medical condition(s) and other key characteristics of individuals who are prescribed opioids with a specific focus on members/patients potentially receiving unnecessary opioid prescriptions?

We used the Milliman MedInsight platform to analyze sample prescription claims from 2012 to 2016 to demonstrate the analyses for answering these critical questions. The claims data was analyzed at an aggregated population level as well as at the individual member, provider, and claim level. Exploratory analyses were conducted within a plan population to:

  • Determine if the overprescribing of opioids is concentrated and isolated within a specific geographic region or network.
  • Understand spikes or changes in prescribing rates for specific opioids over time.
  • Identify specific providers prescribing opioids at a higher than average prescription rate.
  • Identify at-risk members and providers for proactive outreach and education.
  • Evaluate the appropriateness of utilization given the use of pain management in selected co-morbidities.

The outcomes from the exploratory analyses were used to guide the in-depth evaluation and identify potential members and providers for targeted outreach.

Table 1. Opioid Prescribing Trends by Line of Business

One of the first places to start on the path to understanding opioid misuse is to track the ongoing utilization pattern for members across different lines of business (e.g., Medicare, Medicaid) over time as illustrated in Table 1. When a health plan has multiple lines of business, it is crucial to understand if the problem is occurring at an aggregate plan level, or within a specific line of business. Using a normalized metric such as scripts per thousand members provides an effective baseline metrics for comparison rather than simply comparing the sheer volume of opioid prescriptions. These baseline metrics can be further stratified by geographic region, product type (e.g., HMO, PPO), or even isolated by specific types of prescription opioids. We stratified the sample data set by lines of business and determined that the majority of the opioid scripts were being written within the Medicare population.

Table 2. Medicare – Top Prescribed Opioids in 2017

From here, we investigated the opioid prescriptions for the Medicare population to understand what types of opioids were being prescribed. Table 2 on the left shows that the majority of the opioids being prescribed within the sample Medicare population were short-acting opioids such as oxycodone/acetaminophen and hydrocodone/acetaminophen. We also analyzed the prescription rate over time to determine if there was a change in prescribing patterns associated with these two types of opioid medications. We found that the prescribing rate for hydrocodone/acetaminophen increased 133% over the last four years. This information can be useful for care managers as they think through mitigation strategies associated with different types of prescription opioids. The ability to analyze the prescription claims data from such different dimensions helps to hone in on a member population to be targeted for intervention.

Once a baseline understanding of opioid use is established, it is critical to identify sub-populations for effective targeting. It is not only important to identify the members who are receiving these prescriptions, but also understand why and how they are receiving them.  A detailed member level drill down is essential to understand the underlying conditions and diagnoses related to opioid use. Moreover, the member and claim level details also provide insights regarding the providers prescribing these opioids. These findings are helpful to understand if members received multiple opioid prescriptions from more than one provider. Within our sample dataset, we isolated a member’s opioid prescriptions in 2016. We were then able to analyze the total prescriptions they had per month, and how many distinct providers wrote opioid prescriptions in that month. If a plan has a large proportion of members who are receiving opioid prescriptions from more than one provider, they can develop mitigation strategies such as a pharmacy lock-in policy, which “locks-in” members to specific providers in order to monitor and reduce unnecessary distribution.

The ability to conduct such detailed analyses helps develop a deep understanding of prescription opioid misuse which is essential to design an appropriate response. In order to battle the ongoing crisis, payers and providers can utilize their claims data and sophisticated analytics to pinpoint areas for interventions.

For further information please contact rahul.ekbote@milliman.com or christine.castle@milliman.com

(1)Accessed on February 8th 2018 from the website of CDC (https://www.cdc.gov/mmwr/volumes/65/wr/mm655051e1.htm)

(2) Accessed on February 8th 2018 from the website of CDC (https://www.cdc.gov/drugoverdose/data/overdose.html)

(3) Accessed on February 8th 2018. ‘The Underestimated Cost of the Opioid Crisis’ by the Council of Economic Advisers – November 2017 (https://www.whitehouse.gov/sites/whitehouse.gov/files/images/The%20Underestimated%20Cost%20of%20the%20Opioid%20Crisis.pdf)






Healthcare Waste: A Conversation with Dr. Mark Fendrick

StickyAnalytics, Evidence Based Decision Making, Evidence-Based Measures, Health Waste Calculator, Value-Based Care, Waste

Dr. Mark Fendrick conceptualized and coined the term Value-Based Insurance Design and is the director of the Value-Based Insurance Design Center (VBID) at the University of Michigan. His research focuses on how clinical payment and consumer engagement initiatives impact access to care, quality of care, and healthcare costs. He has authored over 250 articles and book chapters and has received numerous awards for the creation and implementation of value-based insurance design. He spoke with Milliman MedInsight about healthcare waste and his work with the MedInsight Health Waste Calculator.

Question: You spent your career trying to improve access to high-value care. Why have you gotten into the low-value care arena?

Well, it turns out that most of the services that people consider high value – I like to say the things I beg my patients to do – while they improve individual and population health, they tend to add to overall healthcare costs. Often in modest ways, sometimes in high ways.

So given that most public and private payers are very concerned about the rate of healthcare cost growth, it has been incumbent upon us who have been pushing payers to use more of high-value care to identify potential opportunities to reduce spending on those services that don’t make Americans any healthier to allow us some head room to buy more of those high-value services that are currently underutilized. Early Value-Based Insurance Design, or VBID, just included reductions in consumer out of pocket costs for high-value services. As people used them more often costs went up.

Now we’re trying to identify those low-value services, make them more difficult for people to access, take advantage of those immediate and substantial savings, and hopefully have much of those savings be put back into services that improve Americans’ health.

Question: Where do you think a payer should start when thinking about low-value care?

Low-value care is a little bit tricky in the fact that when you start removing certain services or making it more difficult to access certain services there is going to be special interest and concern, whether it be on the provider side or on the patient side. So we decided to get a number of stakeholders from across the country together and identify services that were easy to find in claims that were almost always low-value when they were used, and that if we got rid of them no one would really mind.

We wanted these to be below the radar, or as I like to say fruit below the ground. This national multi-stakeholder Task Force on Low-Value Care selected five commonly overused services ready for action for purchasers to identify and eliminate. They include: diagnostic testing and imaging prior to low-risk surgery, population based vitamin D screening, prostate specific antigen testing for men over the age of 65, imaging in the first six weeks after muscular-skeleto back pain injury, and the use of branded drugs when a chemically identical generic was available.

It’s our hope that tools like the Health Waste Calculator can help payers identify these and many other services. Services that if they were successfully removed would pretty much not be noticed and would allow payers millions, if not billions, of dollars to spend on those high-value services that are currently out of reach for many people in America.

To read more about our conversation with Dr. Fendrick click here.






Stick to Those Asthma Meds

Sticky, Avoidable Admissions, Drug Utilization, Value-Based Care

Asthma accounted for 3,651 deaths, 2 million emergency department (ED) visits, and about 11 million physician office visits for 2014 in United States (US).[i] The estimated total cost of asthma, including medical expenses, loss of productivity resulting from missed work days, and premature death, was $56 billion in 2007.[ii] Daily long-term control therapy is recommended for patients who have persistent asthma to reduce airway inflammation, control chronic symptoms, and prevent acute asthma attacks.[iii],[iv] Studies have shown that non-adherence to asthma control medications leads to an increase in exacerbations of asthma and subsequent hospitalizations [v],[vi]

To explore the impact of adherence to controller medications on utilization of healthcare services, Milliman conducted an analysis on a large database using MedInsight, Milliman’s healthcare analytics platform. The analysis targeted MARA (Milliman Advanced Risk Adjuster) adjusted utilization of emergency department services related to asthma and any diagnosis (all cause) in members with persistent asthma. The results were compared among three different groups based on adherence to controller medications (adherence below 50 percent, 50-75 percent, and above 75 percent).

Findings:

  1. A statistically significant inverse correlation was found between the adherence rate (proportion of days covered (PDC)) and the number of ED visits.
  2. Asthmatic members enrolled with Medicare having adherence rate above 75 percent had a higher cost utilization as compared to below 50 and 50-75 percent. To some extent this could be explained by the fact that the members with adherence rate above 75 percent had a high comorbidity risk score as indicated by MARA.
  3. Asthmatic members enrolled with Medicaid had the highest overall ED utilization amongst all payer types.

Males generally had a higher ED utilization as compared to females.

Although analysis was based on administrative data, the statistical significance of the results highlights adherence to asthma controller medications as a target for interventions to reduce utilization and improve health outcomes for members with asthma.

[i] Centers for Disease Control and Prevention. Asthma Facts—National Center for Health Statistics, 2014.

Available at: https://www.cdc.gov/nchs/fastats/asthma.htm

[ii] Barnett SBL, Nurmagambetov TA. Costs of asthma in the United States: 2002–2007. J Allergy Clin Immunol 2011; 127(1): 145–52.

Available at: http://www.jacionline.org/article/S0091-6749(10)01634-9/fulltext

[iii] National Asthma Education and Prevention Program Expert Panel Report 3, 2007. Guidelines for the diagnosis and management of asthma.

Available at: http://www.nhlbi.nih.gov/files/docs/guidelines/asthsumm.pdf

[iv] Centers for Disease Control and Prevention. Asthma Stats. Use of long-term control medication among persons with active asthma.

Available at: http://www.cdc.gov/asthma/asthma_stats/longterm_medication.htm

[v] Piecoro LT, et al. Asthma Prevalence, Cost, and Adherence with Expert Guidelines on the Utilization of Health Care Services and Costs in a State Medicaid Population. Health Services Research. 2001; 36(2): 357-71.

Available at: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1089228/pdf/hsresearch00003-0066.pdf

[vi] Stern L, et al. Medication compliance and disease exacerbation in patients with asthma: a retrospective study of managed care data.  Ann Allergy Asthma Immunol.  2006; 97(3): 402-8.

Available at: http://www.sciencedirect.com/science/article/pii/S1081120610608083






Society of Actuaries Podcast: Machine Learning Tools Overview

Stickymachine learning

Machine learning is a hot topic in healthcare analytics today. Shea Parkes and Anders Larson, two of Milliman’s machine learning experts, started a podcast focusing on the subject for the Society of Actuaries (SOA) in 2015. The latest episode was posted May 1, 2018 and is called “Machine Learning Tools Overview.” In the episode, they discuss the range of tools available to help implement machine learning.

“Tools are anything that helps you get the job done. You should do some planning, you should talk some stuff out. But at a certain point, especially with machine learning, someone is going to need to interface with a computer and get something done.” Shea and Anders discuss a range of tools for machine learning, drilling deeper on Graphical User Interface (GUI) modeling tools such as RapidMiner and KNIME. Many of the prior sessions focused on the concepts and theory, only hinting at specific tools and implementations. This episode kicks off a whole new series that will focus on languages, libraries, frameworks, and cloud providers.

Click here to listen to the latest podcast!

Previous episodes include a 12 part series on machine learning from an actuarial perspective. The episodes start with fundamental concepts, then move on to many of the common machine learning algorithms.

Shea Parkes has been in many different roles in his time at Milliman. Now as a Principal he helps lead PRM Analytics with products that focus on Accountable Care and other Organizations. He also participates in many volunteering opportunities with the Society of Actuaries (SOA), mostly within the SOA’s Predictive Analytics and Futurism division. He served on the council of that division for a term and participated in planning many webinars, conference sessions, and newsletter articles.

Anders Larson is a consulting actuary with the Milliman Indianapolis office. He provides actuarial consulting services to commercial and Medicaid health plans, self-funded groups, and provider organizations.

All the episodes of the podcast can be found on the official SOA page here and you can find it on iTunes!

If you would like to learn more about the podcast and or request topics for discussion email Shea Parkes at shea.parkes@milliman.com.






Pharmacy Prescription – Is It Generic Enough?

Sticky, Drug Utilization, Healthcare Analytics

Prescription drugs comprise the third-largest component of U.S. national health expenditures, behind only hospital care and physician services, according to the Centers for Medicare and Medicaid Services.[i] According to the QuintilesIMS Institute, invoiced sales of prescription drugs amounted to $450 billion in 2016, representing 13.4 percent of all U.S. health spending. Prices for prescription drugs in the U.S. are far higher than they are in other industrialized countries. In 2014, on an invoice price basis, the U.S. spent $1,327 per capita on prescription drugs, as compared to non-U.S. members of the Organisation for Economic Co-operation and Development, which had a median per capita drug spending of $489.[ii]

Generic drugs represent an effective and affordable method to decrease pharmacy costs. Generic drugs have saved the U.S. healthcare system $1.67 trillion in the last decade, generating $253 billion in savings in 2016 alone.[iii]

To explore the extent of generic drug usage in pharmacy prescriptions we conducted a focused analysis using drug categories in one of MedInsight’s client databases. Key findings of this analysis were:

  • The proportion of generic prescriptions dispensed has increased consistently over the last 7 years, accounting for 89% of total prescription events in 2016.
  • Generic prescriptions account for only 27% of total prescription costs.
  • Branded drug prescription events decreased by 48% between 2010 to 2016,
  • Branded spending increased by 42% over that same period.

Our observations were based on our analysis of a single customer’s data, however the results are in line with the analysis presented by the Generic Pharmaceutical Association (GPhA) for 2015.[iv]

We further analyzed the prescription events and prescription costs for the top 5 drug categories in terms of total cost in the year 2016. The key observations were:

  • All five categories were found to be directly or indirectly related to a chronic condition.
  • In four out of five categories, the generic prescription rate was greater than 50%.
  • The exception was anti-asthmatic drugs. These were prescribed largely in the form of branded drugs (78%) with a generic prescription rate of only 22% percent.
  • More than 90% of the total prescription costs were attributable to branded drugs in all top five drug categories.

The analysis confirmed market perception and published trend of increasing generic prescription rates despite the bulk of pharmacy expense still driven by the expensive branded drugs.

[i] Centers for Medicare and Medicaid Services. National Health Expenditures 2015 Highlights. Available at: https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/NationalHealthExpendData/Downloads/highlights.pdf

[ii] Avik Roy,2017.  A Market-Based Plan for Affordable Prescription Drugs. Available at:  https://freopp.org/a-market-based-plan-for-affordable-prescription-drugs-931e31024e08

[iii] Association for Accessible Medicines. 2017 Generic Drug Access and Savings in the U.S. Report. Available at: https://www.accessiblemeds.org/resources/blog/2017-generic-drug-access-and-savings-us-report

[iv] Generic Pharmaceutical Association, 2015. Generic Drug Savings In The U.S. Seventh Annual Edition:2015. Available at: http://www.gphaonline.org/media/wysiwyg/PDF/GPhA_Savings_Report_2015.pdf

 






Clinical Economics of Asthma Controller Medication: Adherence & Outcomes

StickyDrug Utilization, Evidence Based Decision Making, Medical Management

Asthma is a lifelong disease. In addition to limiting a person’s quality of life, medically it is also associated with significant healthcare utilization and costs.[i] According to a survey in 2015, there were 5.9 million physician office visits and 2.0 million Emergency Department visits in the US because of asthma.[ii] [iii] The Centers for Disease Control and Prevention (CDC) estimated that in 2007 asthma alone cost the U.S. about $56 billion in medical costs, lost school and work days, and early deaths.

Good adherence to long-term asthma controller medications, defined as 50% or greater, has been shown to be associated with reduced acute asthma exacerbations in patients with persistent asthma and is one of the key outcome measurements for programs focused on improving persistent asthma management. Despite these facts, actual adherence is only about 30 to 40%.[iv] [v] One likely barrier to achieving higher asthma controller medication adherence is cost. Generic medications, which are usually less expensive than branded medications, were found to have better overall adherence rates.[vi]

To explore asthma controller medication adherence rates, treatment outcomes, and any potential association with branded or generic drugs, Milliman conducted a focused analysis using MedInsight’s healthcare analytic platform on a large database containing claims for over 3.9 million Medicaid, Medicare Advantage and commercial members. We compared the incidence of acute asthma exacerbations and status asthmaticus events among asthma patients (defined as having intermittent or persistent asthma) who were exclusively on branded asthma controller medications to patients who were exclusively on generic asthma controller medications.  We also researched the proportion of generic and branded asthma controller medications and utilization patterns in patients that achieved a high adherence rate. We defined our goal for high adherence as 80% or greater for the proportion of days covered (PDC) with asthma controller medications for patients under treatment for more than three months.

Findings:

  • Asthmatic patients were prescribed 52 types of asthma long-term controller medication drug compositions within the study period. This compares with a total of 69 drug compositions considered to be asthma long-term controller medications (HEDIS AMR-A table).[vii]
  • Forty-two out of the 52 prescribed compositions were branded compositions.
  • Twenty-two percent of asthma patients using generic controller medications exclusively had high adherence rates (80% or higher) compared with 14% of asthma patients who were using branded controller medications exclusively.
  • Asthma patients using generic controller medications exclusively had much lower rates for acute exacerbations and status asthmaticus (34%) than asthma patients using branded controller medications exclusively (62%). Both groups had high adherence rates of over 80%.

  • We also found that patients using generic controller medications exclusively and having adherence rates greater than 80% had 24% lower Emergency Department visits and 32% lower inpatient admissions per thousand member months compared with patients using branded controller medications exclusively.

Our observations were based on limited administrative claims data; however, the results are in line with the analysis presented by association for accessible medicines depicting higher compliance among patients on generic drugs and other studies.vi While potential reasons for this finding such as demographic differences or formulary composition were not investigated, the most likely reason for this observation is differences in asthma severity between the population using generic controller medications and the population using branded controller medications.

References:

[i] American Academy of Allergy, Asthma and Immunology. Allergy and Asthma Drug Guide. Available at: http://www.aaaai.org/conditions-and-treatments/drug-guide.

[ii] Centers for Disease Control and Prevention. National Center for Health Statistics. Last updated: March, 2017. Available at: https://www.cdc.gov/nchs/fastats/asthma.htm.

[iii] National Ambulatory Medical Care Survey: 2015 State and National Summary Tables. Available at:https://www.cdc.gov/nchs/data/ahcd/namcs_summary/2015_namcs_web_tables.pdf

[iv] Asthma Controller Medication Adherence, Risk of Exacerbation, and Use of Rescue Agents Among Texas Medicaid Patients with Persistent Asthma. Published: 2015. Available at: http://www.amcp.org/WorkArea/DownloadAsset.aspx?id=20461.

[v] Mika J. Mäkelä, Vibeke Backer, Morten Hedegaard, Kjell Larsson, Adherence to inhaled therapies, health outcomes and costs in patients with asthma and COPD, Respiratory Medicine, Volume 107, Issue 10, 2013, pages 1481–1490, ISSN 0954-6111. Available at: http://dx.doi.org/10.1016/j.rmed.2013.04.005.

[vi] Generic Drug Access & Savings in the U.S. Association for Accessible Medicines (AAM). 2017 Report. Available at: https://www.accessiblemeds.org/sites/default/files/2017-07/2017-AAM-Access-Savings-Report-2017-web2.pdf.

[vii] HEDIS 2017 Volume 2 NDC Code List. Available at: http://www.ncqa.org/hedis-quality-measurement/hedis-measures/hedis-2017/hedis-2017-ndc-license/hedis-2017-final-ndc-lists.






MedInsight Health Waste Calculator Featured in Health Affairs’ Third Most-Read Article of 2017

StickyHealth Waste Calculator, Healthcare Analytics, Waste

Milliman’s MedInsight Health Waste Calculator – a leading analytic platform that provides actionable data around healthcare quality, efficiency, and effectiveness – was featured in Health Affairs’ third most-read article of 2017! The article, “Low-Cost, High-Volume Health Services Contribute The Most To Unnecessary Health Spending,” highlights the Commonwealth of Virginia’s use of the MedInsight Health Waste Calculator, which found more than $586 million in unnecessary costs in the Virginia All-Payer Claims Database.

Every year the healthcare industry wastes an estimated $750 billion[1]. Milliman created the MedInsight Health Waste Calculator to root out low-value, wasteful services to help significantly reduce healthcare waste for payers, providers, APCDs, and other healthcare systems.

The Virginia Center for Health Innovation (VCHI), Virginia Health Information (VHI), and Value-Based Insurance Design (VBID) Health partnered with MedInsight to support Virginia’s statewide initiatives to identify and reduce low value, unnecessary medical tests and procedures in their healthcare systems. VCHI brought together Virginia’s leading health providers, payers, and employer groups to root out low-value care. Working with the VHI team, they used the MedInsight Health Waste Calculator to aid their measurement, monitoring, and reporting on services that provide no net health benefit.

The Health Affairs article outlines the forty-four low-value health services analyzed in the 2014 data, why they were so costly, and how the MedInsight Health Waste Calculator identified them.

To read full press release or download the full article click here.

 

[1] Institute of Medicine, Best Care at Lower Cost: The Path to Continuously Learning Health Care in America. National Academies Press, 2013.






Overused Ophthalmology Imaging: More May Not Be Better

Sticky, Care Coordination, Episodes of Care, Evidence Based Decision Making, Value-Based Care

Globally, policy makers and payers are focusing on services that are regarded medically unnecessary and do not yield useful results. In the United States, these are referred to as low- or no-value services and are considered wasteful. A significant percentage of the services considered wasteful are medical imaging when done without appropriate indications. Inappropriate use of imaging services contributes to unjustifiable healthcare costs and may lead to irrelevant incidental findings, exposing individuals to unnecessary radiation doses.1

We set out to research wasteful eye imaging services because there is very limited literature or analytical research in this area. According to the American Academy of Ophthalmology Preferred Practice Pattern® guidelines, in the absence of symptoms or significant pathology, a comprehensive history and physical examination should suffice and imaging is not indicated. Advanced eye imaging (optical coherence tomography, visual field testing, fundus photography, or eye photography) should be reserved for malignant neoplasms of the eye, retinal detachments, and injury to eye, optic nerves, or optic chiasm.2

Similar views are advocated by Choosing Wisely® from the American Academy of Ophthalmology, which states: “Don’t routinely order imaging tests for patients without symptoms or signs of significant eye disease in the absence of symptoms or signs of significant disease pathology.”3

We used Milliman MedInsight’s Health Waste Calculator (HWC) to identify wasteful eye imaging services. The HWC ‘Imaging Tests for Eye Disease’ measure identifies eye imaging as wasteful unless the patient has symptoms or signs of significant eye disease such as neoplasms of the eye, choroidal detachment, optic atrophy, glaucoma, diabetic retinopathy, macular degeneration, etc., where imaging is considered medically necessary. We calculated the HWC ‘Imaging Tests for Eye Disease’ measure for one of our clients, a commercial health plan with approximately 2,091,517 members, using claim data from calendar year 2015. As shown in Figure 1, 67% of eye imaging services measured were wasteful and cost nearly $8,198,169.

Figure 1: Pattern of Eye Imaging Services

Service TypeNumber of ServicesPercentage of ServicesAggregate Allowed AmountPercentage of Allowed Amount
Total Eye Imaging157,308100%$12,736,798100%
Necessary Eye Imaging51,88833%$4,538,62934%
Wasteful Eye Imaging105,42067%$8,198,16966%

Figure 2: Percentage Distribution of Wasteful Eye Imaging 

Figure 2 profiles the distribution of the various advance eye imaging modalities that have been identified as wasteful. For each modality, the figure identifies the percentage of services that are wasteful and the corresponding percentage of wasteful total allowed dollars. Optical computed tomography has the highest rate of wasteful services and wasteful costs in the eye imaging service waste category. Analysis highlighted that it was performed for an array of inappropriate diagnoses including but not limited to headache, dizziness, and routine eye exam visits. This supports the proposition that eye imaging is being overused and is often done inappropriately in the absence of significant underlying disease pathology.

Another interesting finding revealed that 41.9% of members with wasteful services had a repeat service for eye imaging in a year, which was also determined wasteful.

It is important to note that claim data alone allows only an approximate identification of wasteful eye imaging. Even so, we were able to confirm a high prevalence of wasteful eye imaging for our client that allowed them to target potentially avoidable utilization and costs.

For more information on how to identify wasteful services, visit the MedInsight Health Waste Calculator web page. [Link to URL: http://www.medinsight.milliman.com/MedInsight/Products/Medinsight-Tools/?prid=71832]

1 Sorenson C, Drummond M, Bhuiyan Khan B. Medical technology as a key driver of rising health expenditure: Disentangling the relationship. ClinicoEconomics and Outcomes Research, 2013:5 223–234.

2 American Academy of Ophthalmology Preferred Practice Patterns Committee. Preferred Practice Pattern Guidelines. Comprehensive Adult Medical Eye Evaluation. Ophthalmology. January 2016. Volume 123, Issue 1, Pages P209–P236.

3      American Association Ophthalmology. Choosing Wisely. Five things physicians and patients should question. February 21, 2013. Accessed at http://www.choosingwisely.org/clinician-lists/american-academy-ophthalmology-routine-imaging-for-patients-without-symptoms-or-signs-of-eye-disease/






Reducing Healthcare Waste – A Three Pronged Approach

Sticky, Waste

Experts and stakeholders across the industry agree that there is an immediate need to reduce wasteful healthcare spending in the US. A 2012 Institute of Medicine report estimated that one third of US healthcare spending, or $750 billion is wasted annually (Washington Post, September 2012). Of that amount, $210 billion is spent on unnecessary services attributable to overtreatment. Overtreatment can result from a number of varying factors. However, regardless of the underlying reasons, the end result is that patients are exposed to treatments and tests that do not result in optimal care and outcomes. In other words, these are healthcare services we shouldn’t be paying for regardless of the cost. More alarming than the huge amounts of money spent on overtreatment, is that many of these tests and treatments and their downstream effects can be harmful to patients.

There are many stakeholders motivated to reduce healthcare waste. Regardless of whether you’re a payer or a provider, the fundamental steps to actually reduce healthcare waste are similar:

  • Identify and quantify wasteful services;
  • Implement strategies to reduce utilization of unnecessary services; and
  • Track and evaluate the success of waste reduction strategies and adjust efforts accordingly.

To identify and quantify wasteful services, Milliman and VBID Health collaboratively developed a tool called the MedInsight Health Waste Calculator. The MedInsight Health Waste Calculator categorizes claims data using logic based on national initiatives such as Choosing Wisely, to quantify services that research has proven add no value in specific clinically nuanced scenarios. These wasteful services are flagged at the claim line level allowing the Health Waste Calculator to quantify utilization and costs associated with wasteful services. With over 400 clinical measures in the development pipeline, the latest version of the MedInsight Health Waste Calculator has 42 measures to quantify waste related to about 60 Choosing Wisely measures from claims data. Quantifying waste allows organizations to prioritize which services to target with waste reduction strategies by aligning internal priorities with metrics on waste prevalence and cost, and factors such as risk of member harm.

There are a number of high profile initiatives and collaborative partnerships aimed at identifying specific wasteful or unnecessary medical services. Two of the most prominent of these collaborative initiatives are Choosing Wisely and the US Preventative Services Task Force (USPSTF). One example of a service that is potentially unnecessary is cardiac imaging in certain situations. Specifically, the American College of Cardiology identified as unnecessary stress cardiac imaging or advanced non-invasive imaging in the initial evaluation of members where there are no cardiac symptoms or no high risk markers present. As noted on the Choosing Wisely website, “asymptomatic, low risk patients account for up to 45 percent of unnecessary screening.” (Choosing Wisely, American College of Cardiology) A more common and familiar category of services that has generated a lot of publicity over the years has been unnecessary prescription of antibiotics for non-bacterial infections. Among the harmful consequences of these services are increased antimicrobial resistance causing severe infections, complications, and longer hospital stays. (Antimicrobial resistance: risk associated with antibiotic overuse and initiatives to reduce the problem)

Once wasteful services have been identified within a population, the obvious question is, what can be done to reduce these unnecessary services. Organizations can leverage several different strategies to address and reduce wasteful services. Strategies are generally centered on one or more levers such as analytics and reporting, education and promotion, claim adjudication, provider network management, medical management, purchasing criteria, and/or benefit design. Strategies will vary based on cost and resources available to implement, and how much they disrupt the healthcare delivery and payment cycle. One of the MedInsight Health Waste Calculator’s earliest clients used the tool’s output to mature their waste reduction strategies and has realized significant savings by reducing wasteful services. By utilizing practice consultants in each of their accountable care organizations and embarking on member awareness campaigns among other interventions, this payer has reduced per member per month costs on ten waste measures ranging from 1.6% savings on unnecessary imaging for uncomplicated rhinosinusitis to 30.2% savings on unnecessary cervical cancer screening in women aged 13 to 20.

Less disruptive strategies, such as member awareness campaigns, will likely have the least amount of impact on reducing wasteful services. Initiating utilization management requirements for potential wasteful services can be costly and disruptive, but would have a high impact on reducing wasteful services. Quantifying unnecessary services in a population can provide guidance on where to focus an organization’s often limited resources. By categorizing identified waste into high volume, low cost services versus high cost, low volume services an organization can better predict expected savings against the cost of various intervention strategies. This article describes analysis of 44 wasteful health services in the Virginia All Payer Claims Database, which revealed that low-cost, high-volume services contributed the most to unnecessary health spending (Mafi, et al., 2017).

Regardless of the strategy, the ultimate objective of reducing wasteful services is to change the behavior of the providers who are ordering the potentially unnecessary services. With this key concept in mind, a group of researchers conducted a study to identify a framework that would be most favorable towards changing physicians’ behavior to reduce low-value care (Parchman, M.L., Healthcare (2016), http://dx.doi.org/10.1016/j.hjdsi.2016.10.005). The researchers conducted interviews with an eight-member stakeholder advisory committee that included patients, providers and health care leaders.  Among their findings, the research group found that providers were most inclined to change their behavior when wasteful services were presented as not only wasteful, but also harmful or potentially harmful to their patients. The overall framework that the researchers found to be most effective in creating the culture for change included:

  • Prioritize addressing low-value care;
  • Build a culture of trust, innovation and improvement;
  • Establish a shared language and purpose; and
  • Commit resources to measurements

Once a strategy has been implemented, the final step is to track and evaluate the effectiveness of the intervention strategy. The same tools and methods used to identify the wasteful or unnecessary services should be used to track these same services over time to determine if there has been an actual reduction in the services that were targeted for intervention. In closing, there is no debate that measurement of waste is critical for any serious approach to reduce unnecessary healthcare services. Integral to a three-pronged approach to reduce wasteful and unnecessary services is identifying, quantifying and then tracking the wasteful services over time.






Wasteful Pre-Operative Cardiac Testing Prior to Low-Risk Surgery

Sticky, Evidence Based Decision Making, Health Waste Calculator, Medical Management, Uncategorized, Value-Based Care

Anesthesiologists perform pre-operative assessment on all patients undergoing surgery in order to identify any disease or risks to the surgery and to plan perioperative anesthetic care to mitigate surgical risk. In patients with known cardiac disease, diagnostic cardiac tests such as stress tests and echocardiograms are sometimes recommended as part of this assessment.

A recommendation from the American Society of Anesthesiologists notes that pre-operative cardiac stress testing is only appropriate for identifying extremely high-risk patients, in whom the results would change management prior to surgery, change the decision of the patient to undergo surgery, or change the type of procedure that the surgeon will perform. 1 The American College of Cardiology and American Heart Association recommends that there is no benefit for routine pre-operative cardiac testing in low risk surgery for patients with no cardiac disease. 2

For this reason, Milliman MedInsight’s Health Waste Calculator identifies pre-operative cardiac testing for patients undergoing low- or moderate-risk non-cardiac procedures (e.g., surgery cataract, laparoscopic cholecystectomy, corneal transplant, and removal of tonsils) as wasteful. We applied the Health Wast Calculator pre-operative testing measure for one of our clients, a commercial health plan with approximately 850,000 members, using claims data from calendar year 2012. As shown in Figure 1, 2.2% of the 5,120 pre-operative cardiac testing services were wasteful, at a cost of $88,000. Hence, it is recommended that clinicians perform pre-operative evaluation appropriately and reduce distress among patients.

Service TypeNumber of ServicesPercentage of ServicesAggregate Allowed AmountPercentage of Allowed Amount
Total Preoperative Cardiac Testing5,120100%$2,513,987100%
Wasteful Preoperative Cardiac Testing1122.19%$88,0413.50%

It is important to note that claims data alone allows only an approximate identification of wasteful pre-operative cardiac testing. Even so, we were able to confirm instances of pre-operative testing for our client that could help it potentially avoid healthcare costs in the absence of benefit.

For more information on how to identify wasteful services, visit the MedInsight Health Waste Calculator web page.






Pre-operative Laboratory Testing for Low Risk Surgeries

Sticky, Evidence Based Decision Making, Health Waste Calculator, Population Health, Value-Based Care, Waste

Routinely performing pre-operative testing is a common practice prior to surgery. The American Society of Anesthesiologists (ASA) Task Force defines routine tests as those done in the absence of any specific clinical indication or purpose (i.e. tests intended to discover a disease or disorder in an asymptomatic patient).1   Routine pre-operative testing typically consists of a panel of blood and urine tests, chest X-rays and electrocardiogram.  The purpose is to identify any condition that would add risk to the surgery, and to adjust perioperative anesthetic care to mitigate surgical risk or even postpone or cancel surgery.[1] Although pre-operative testing has long been a common practice there is little or no evidence that supports any benefit for “routine testing” before low risk surgery.[2]

Unnecessary pre-operative testing may cause a patient to be subjected to increased cost of surgical care,   false positive, or borderline results that lead to further follow-up investigations.2 Follow-up investigations due to false positive results can cause unnecessary psychological and economic burdens, postponement of surgery, and even morbidity and mortality.2 In a cost effectiveness study by the Agency for Healthcare Research and Quality (AHRQ), it was found that routine pre-operative investigations resulted in a delay or cancellation of the planned surgery in about 2% of cases, some changes to anesthetic management in up to 11% of cases, or resulted in a change in surgical procedure in under 1% of cases.2

The components of preoperative laboratory testing considered in this blog were metabolic, general health, and electrolyte panel; urinalysis; blood glucose; serum albumin, creatinine, potassium, and sodium; bilirubin; complete blood count (CBC), hemoglobin, bleeding time, prothrombin time; alanine aminotransferase (ALT) and aspartate aminotransferase (AST).

To study the practice pattern of this pre-operative laboratory testing in low risk surgery cases we used the MedInsight Health Waste Calculator’s evidence-based algorithm which includes the Choosing Wisely guidelines for identifying unnecessary testing for low risk surgical procedures. The analysis was conducted on claims from a large Midwestern commercial health plan for the year 2015. Low-risk surgeries considered in this study include various endoscopic and laparoscopic surgeries, minor gynecology, orthopedic, ophthalmology, and urological procedures including superficial surgeries.

We identified 84 percent of pre-operative laboratory testing services associated with low risk surgical procedures as wasteful, meaning that only 16 percent were necessary (Figure 1). The total aggregate allowed cost for pre-operative laboratory testing was $80,445,931, of which wasteful laboratory testing associated with low risk surgical procedures accounted for $1,471,672, which is 2% of the total cost for pre-operative laboratory testing. These results show that wasteful pre-operative laboratory investigations add significant costs for patients and payers.

On further analysis of wasteful laboratory testing services, the top five wasteful pre-operative tests based on total cost were comprehensive metabolic panel, complete CBC, general health panel, urinalysis, hemoglobin (Hb) and prothrombin time (PT). Collectively these five common pre-operative lab tests contributed to 92% of the total wasteful laboratory testing. The profiling of these top five pre-operative tests is shown in Figure 2.

It is important to note that claims data alone is not sufficient to identify wasteful services with absolute certainty. However, even with conservative estimates, these findings confirm that a significant percentage of pre-operative laboratory testing done for low risk surgeries are wasteful.

[1] Apfelbaum JL, Connis RT, Nickinovich DG et al. Practice advisory for preanesthesia evaluation: an updated report by the American Society of Anesthesiologists Task Force on Preanesthesia Evaluation. Anesthesiology. 2012 Mar; 116(3):522–38.

[2] Balk EM, Earley A, Hadar N, Shah N, Trikalinos TA. Benefits and Harms of Routine Preoperative Testing: Comparative Effectiveness. Comparative Effectiveness Review No. 130. (Prepared by Brown Evidence-based Practice Center under Contract No. 290-2012-0012-I.) AHRQ Publication No. 14-EHC009-EF. Rockville, MD: Agency for Healthcare Research and Quality; January 2014






Is Risk Adjustment Enough in Population Analytics?

StickyPopulation Health, Risk Adjustment

Most analysts conducting population health data analysis understand that no two populations are completely the same. Even within a population with similar attributes – people with congestive heart failure, for example – there will be variances in the population subgroups. Many factors contribute to these variances, including age and gender, comorbidities and geographic differences in treatment patterns as well as provider unit pricing and efficiency.

Analysis of cost and utilization data, even with the use of “per member per month” and “per 1,000” calculations, does not produce a fully accurate comparison because of variations in risk or severity of the population. For a more “apples to apples” comparison, a common approach is the application of risk or severity adjustment methodologies to the data to normalize utilization and cost across the populations.

Risk adjustment methodologies, such as the Milliman Advanced Risk Adjuster (MARA), Optum’s Episode Risk Groups, or Verisk’s DxCGs, are common tools used to risk adjust results to account for the variations in populations’ risk and severity.

Consider the example below. Four PCP groups are each treating a population of patients who have Diabetes without Coronary Artery Disease as identified using Milliman’s Chronic Condition Hierarchical Groups. Each medical groups’ population is of varying size and consists of a mix of commercial and Medicare patients. In Figure 1 we see that the total medical and pharmacy Allowed PMPM cost for all services received by the identified patients is highest for Group 3 at $1,063 PMPM. (All services, regardless of diagnosis.)

Figure 1  
Medical Member MonthsAllowed PMPM
Group141,441$828
Group 220,975$893
Group 36,503$1,063
Group 44,869$843
Total73,788$868

Next, using MARA, we take the risk of each population into consideration. As shown in Figure 2, the population attributed to Group 2 has the highest average MARA concurrent risk score at 3.11, and the Group 3 population’s average concurrent risk is slightly lower at 3.06. A higher overall risk score means that the population’s medical conditions are more severe. Adjusting the cost experience for the population severity allows for a more comparable analysis of the cost of the two populations. We see that, even with the risk adjustment, Group 3’s allowed PMPM is well above the other three groups.

Figure 2    
Medical Member MonthsAllowed PMPMAvg MARA Concurrent RiskRisk Adj All PMPM
Group 141,441$8282.97$279
Group 2 20,975$8933.11$287
Group 3 6,503$1,0633.06$348
Group 44,869$8432.89$291
Total73,788$8683.01$288

Now, if we want to compare the efficiency of resource utilization in the treatment of these similar populations, an additional adjustment is needed to separate out unit cost versus resource use. Milliman’s GlobalRVUs are a unit value system that covers the entire range of healthcare services, including physician, hospital, DME, and pharmacy.

In Figure 3 we see that additional calculations have been added to our analysis. The Risk Adjusted RVUs PMPM is the total RVUs (RVUs) for each group’s diabetic population divided by the population’s average MARA concurrent risk score, giving us an overall measure of utilization. Allowed per RVU is the total allowed amount (not adjusted) divided by the total number of RVUs associated with the delivery of services. Utilization Efficiency is calculated as the ratio between the PCP group’s risk adjusted RVUs to the overall average.

Using these data, we see that the main driver of the higher risk-adjusted allowed PMPM is that Group 3’s Utilization Relativity Ratio (based on Risk Adjusted RVU PMPM) is 1.15 or 15% higher than the aggregate for the groups and 19% higher than the next highest group (0.97). We also see that relative price, as measured by Allowed Per RVU, is a relatively small contributor to the difference in risk-adjusted allowed PMPM as the Allowed Per RVU for Group 3 ($45.14) is only 2% higher than the aggregate for the group ($44.39).

Figure 3       
Medical Member MonthsAllowed PMPMAvg MARA Concurrent RiskRisk Adj All PMPMAllowed per RVURisk Adj RVUs PMPMUtilization Relativity Ratio
Group 141,441$8282.97$279$436.460.97
Group 220,975$8933.11$287$466.240.93
Group 36,503$1,0633.06$348$457.701.15
Group 44,869$8432.89$291$466.300.94
Total73,788$8683.01$288$446.681.00

Applying risk adjustment methodologies to cost of care analysis is the approach many analysts use. But is risk adjustment enough? As shown in the example above, leveraging an RVU methodology in conjunction with risk adjustment can provide additional critical insights into variances due to both population severity and provider efficiency.

 






Task Force on Low Value Care – Fall Convening

StickyValue-Based Care

The excitement in the room was palpable as 28 attendees descended on the Westin’s Earhart Room in Detroit, Michigan. Stakeholders from across the nation gathered on September 13 with a common goal—to reduce wasteful healthcare services in the United States. Our charge for the day was to prioritize 29 healthcare services that, in particular clinical scenarios, should not be offered even if they are free. These low-value healthcare services account for an estimated $210 billion or 27% of wasteful healthcare spending in the United States.1 The Task Force on Low Value Care narrowed the list to six measures, meant to serve as a starting point to reduce wasteful care.

Attendees represented a variety of stakeholders. Policy executives from Pfizer, Sanofi, J&J, and Amgen rubbed elbows with large employers such as Caterpillar, Walmart, GE, Comcast, and Hewlett Packard Enterprise. Employer coalitions from the Midwest and Pacific regions joined representatives from the Mid-America Coalition on Healthcare, the National Coalition on Healthcare, Network Strategic Initiatives, and the Independent Colleges and Universities Benefits Association. States were generously represented with attendees from the State Comptroller of Connecticut, the New York City Mayor’s Office of Labor Relations, and the Virginia Center for Health Innovation, which is embarking on the creation of a Virginia Health Value Dashboard. Health plan representation included epidemiology specialists from Kaiser Permanente Washington and Blue Cross Blue Shield of Massachusetts. The National Patient Advocate Foundation brought the patient perspective to the table, and representatives from MedInsight and VBID Health, collaborators on the MedInsight® Health Waste Calculator, provided insight on measuring and acting on the Task Force’s recommendations.

The Task Force used seven key criteria to identify the best services to start with. The first involved considering the potential for patient harm, whether physical, financial, psychological, or due to downstream procedures resulting from over-treatment. Recognizing the need for short-term wins to motivate continued action on this large problem, the Task Force also considered ease of implementation and identified measures where successful examples exist. For example, by requiring providers to select one or more of five evidence-based rationales, a health system in Alberta has seen a 92% reduction in unnecessary vitamin D screening. In order to make a significant impact on the issue, the Task Force also prioritized measures with high unit price, high prevalence, high aggregate spend, and high Waste Index (the proportion of services that are wasteful in each nuanced clinical scenario). Last, but equally as important, the Task Force considered political feasibility, selecting services that affect a variety of conditions in order to minimize consumer and provider backlash.

The list of six selected measures will be released in official communications by the Task Force. Five of the six measures are in production, allowing reliable quantification and trending with the MedInsight Health Waste Calculator.

 

  1. Smith, M., Saunders, R., Stuckhardt, L., & McGinnis, J. (2013). Best Care at Lower Cost: The Path to Continuously Learning Health Care in America. Washington, DC: The National Academies Press.






What a (Low Back) Pain

Sticky, Avoidable Admissions, Avoidable Events, Uncategorized, Value-Based Care

Low back pain is a common problem: more than one in four adults surveyed report having had low back pain during the past three months.[1] Though most low back pain resolves on its own, without medical intervention, it remains one of the most common reasons for physician visits.[2]

Imaging tests of the low back, such as plain X-rays, CT scans, or MRIs, commonly show abnormalities—even in asymptomatic people. Evidence of disk abnormalities, such as degeneration, bulging, protrusion, and annular fissure, is present in over 50% of asymptomatic people in their 30s, and in nearly 90% of people aged 60 or older.[3]

This combination creates a situation where people who undergo imaging tests for low back pain may have abnormalities identified that are unrelated to the pain. Additional tests, consultations, drugs, therapies, and procedures—sometimes even major surgery—may cascade from the initial decision to perform imaging. All this for low back pain that may resolve on its own: what a waste!

Milliman’s MedInsight® Health Waste Calculator can identify wasteful imaging practices for low back pain, using an evidence-based algorithm to analyze claims data in order to classify imaging services as “necessary,” “likely wasteful,” or “wasteful.” For one of our clients, a large commercial health plan, we looked at services performed in 2012. Forty-six percent of services were wasteful; 45% were likely wasteful; and only 9% were necessary. The total aggregate allowed cost for low back pain imaging was $3.3 million, of which wasteful imaging accounted for $2.0 million.

With an evidence-based and conservative approach to measuring the cost of wasteful imaging services—one that does not include the costs of the cascade of additional services—the MedInsight Health Waste Calculator identified an important opportunity to address cost and care for low back pain for our client.

For more information on how the MedInsight Health Waste Calculator can identify wasteful services, click here.

[1] Centers for Disease Control and Prevention (2015). Health, United States, 2015: Table 41: Severe Headache or Migraine, Low Back Pain, and Neck Pain Among Adults Aged 18 and Over, by Selected Characteristics: United States, Selected Years 1997-2014. Retrieved December 2, 2016, from http://www.cdc.gov/nchs/data/hus/2015/041.pdf.

[2] Centers for Disease Control and Prevention. National Ambulatory Medical Care Survey: 2012 State and National Summary Tables: Table 11: Twenty Leading Principal Reasons for Office Visits, by Patient’s Sex: United States, 2012. Retrieved December 2, 2016, from http://www.cdc.gov/nchs/data/ahcd/namcs_summary/2012_namcs_web_tables.pdf.

[3] Brinjikji, W., Luetmer, P.H., Comstock, B., Bresnahan, B.W., Chen, L.E., Deyo, R.A., et al. (April 2015). Systematic literature review of imaging features of spinal degeneration in asymptomatic populations. AJNR Am J Neuroradiol. 2015 Apr;36(4):811-6. doi: 10.3174/ajnr.A4173. Epub 2014 Nov 27. AJNR Am J Neuroradiol. 2015 Apr;36(4):811-6. doi: 10.3174/ajnr.A4173.






Methodology for Identifying Inpatient Admission Events

Sticky, Evidence-Based Measures, Healthcare Analytics

Evidence-based measures (EBMs) and many other health insurance data analytics require counts of inpatient admission events and associated costs. There are multiple ways to identify and count inpatient admission events, each of which may produce a different result. As per the California Department of Health Services, the methodologies used to identify and count individual inpatient admission events vary significantly across different organizations, due to unique data structures and availability.[i] An inpatient event can be identified by various data points like an inpatient Evaluation and Management (E&M) Current Procedure Terminology (CPT) code; an inpatient Uniform Billing (UB) revenue code; or an inpatient bill type. However, when it comes to utilization or counting of inpatient events, the facility claims that contain inpatient Uniform/Universal Billing form (UUB) revenue codes or inpatient bill types are the most credible, and therefore should be the ones used to identify inpatient admissions.

It is important to highlight that an inpatient admission event typically produces more than one claim line, and all of these individual claim lines must be grouped together to constitute the complete inpatient admission event. Various methods are available to do this, with specific advantages and limitations in each method. We have reviewed several of these methods with their potential advantages and shortcomings and have explored the impact of applying these different methods in a representative claims data set. Inpatient claims can come from different types of facilities, including acute hospitals, long-term acute care (LTAC) centers, acute rehabilitation centers, and skilled nursing facilities (SNFs), where there is evidence that the insured stayed overnight.[ii]

Methods

Method 1: All claim lines for a member with same admission date and discharge date constitute a single inpatient admission event when any one of the claim lines have a UB revenue code for inpatient services.

Method 1 accurately identifies an inpatient facility claim line but it fails to account for overlapping dates on claim lines. A single inpatient admission event can have individual claim lines with different admission and discharge dates. This is referred to as “overlapping dates.” Method 1 counts each claim line with a different admission or discharge date as a unique inpatient admission event and calculates the length of stay (LOS) separately for each.

Method 2: Method 1’s logic is used as the first step to identify admission events, followed by a second step grouping all claim lines with overlapping admission and discharge dates into one inpatient admission event.

Method 2 assumes that claim lines with overlapping admission and discharge dates are part of the same continuous stay and counts them as one inpatient admission event.

Method 3: All claim lines for a member with the same admission date and discharge date constitute a single inpatient admission event when any one of the claim lines has either a UB revenue code for inpatient services, or a bill type for inpatient services.

Method 3 casts a wide net to identify inpatient admission events, as it employs more claim types and service codes than Methods 1 or 2. However, similar to Method 1, claim lines with overlapping admissions or discharge dates are not grouped into a single inpatient admissions event. Expanding the data used to identify inpatient admission events increases the probability of having claim lines with overlapping admission or discharge dates.

Method 4: A step to combine claim lines with overlapping admission or discharge dates is added to Method 3.

Method 4 assumes that claim lines with overlapping admission and discharge dates are part of the same continuous stay and counts them as one inpatient admission event.

Method 5: All claim lines for a member with same the claim/encounter ID are grouped into one inpatient admission event when any of the claim lines have a UB revenue code for inpatient services or a bill type for inpatient services.

Method 5 assumes that the claim/encounter identification (ID) remains the same throughout an inpatient admission event, and combines all claim lines with the same claim/encounter ID. This method counts all claim lines with the same claim/encounter ID, even when there is no overlap between admission and discharge dates.

Variation in Results

We applied these different methods for counting inpatient stays on a test database composed primarily of commercial plan members with over 2.6 million lives and 18,551,986 member months for the year 2012 to compare inpatient utilization results for a Healthcare Effectiveness Data and Information Set (HEDIS) EBM: Inpatient Utilization General Hospital/Acute Care (IPU). Results of this comparison are provided in the table in Figure 1.

Figure 1: Comparison of Results Using Different Inpatient Stay Methodologies

Figure 1: Comparison of Results Using Different Inpatient Stay Methodologies

Inpatient admission event counts varied up to 25% across the different methodologies. These results clearly confirm that counts of inpatient admission events can vary significantly depending on the methodology used. Average LOS is also influenced considerably by the method used to identify inpatient admission events.

The outcomes of this analysis emphasize the importance of using a comprehensive methodology when defining inpatient admissions events. It also highlights the need to know the underlying methodology when comparing results for EBMs, for other inpatient admissions event-related measures between organizations, or for different time periods in order to understand if variations in results may be due to the methodology used to define inpatient admissions event rather than operational or quality differences.

[i] California Department of Healthcare Services. Methodology for Identifying Inpatient and Emergency Room Encounters, Appendix A. Retrieved September 7, 2016, from http://www.dhcs.ca.gov/provgovpart/Documents/Appendix%20A.%20Methods.pdf.

[ii] Health Care Cost Institute (September 2012). Health Care Cost and Utilization Report: 2011: Analytic Methodology. Retrieved September 7, 2016, from http://www.healthcostinstitute.org/files/HCCI_HCCUR2011_Methodology.pdf.






Covered California: Leveraging Telemedicine for Network and Quality Management

StickyHealthcare Reform, , ,

Healthcare purchasers, including state-based marketplaces, are making a push to move healthcare payments from volume to value and to transform the delivery system to promote quality of care. Covered California is seeking to do this through its contracting health plans and the qualified health plan (QHP) solicitation process. The most recent solicitation lays out key milestones for the 2017 to 2019 plan years. Beginning in 2017, during the annual QHP certification, a contracting plan will be required to “report on its strategies to ensure that contracting Providers are not charging unduly high prices and for what portion of its entire enrolled population it applies each strategy.” Strategies may include various tactics to move toward value-based payments, drive enrollees to high-value providers, and engage enrollees in decision-making and their own health management.

Milliman MedInsight® can assist health plans and providers with developing appropriate strategies, identifying analytic requirements, and conducting analysis to support them in their network management, quality management, and transformation efforts. We will be releasing a series of articles on these related topics.

To request the first article, “Leveraging Telemedicine and Digital Health Tools as a Strategy for Network and Quality Management,” click here.






Milliman MedInsight Chosen by Two State-Based Organizations and a Leading Community Coalition to Identify Unnecessary Costs and Procedures in Healthcare Delivery

StickyAnalytics, Value-Based Care

Milliman announced today that its healthcare tool, the MedInsight® Health Waste Calculator, has been chosen by two state All Payer Claims Databases (APCDs) and a leading Regional Health Improvement Collaborative to aid in their identification of unnecessary costs and procedures in healthcare delivery.

The Virginia Center for Health Innovation (VCHI), Oregon Health and Hospitals, and the Washington Health Alliance (WHA) have all licensed the MedInsight Health Waste Calculator to support their statewide initiatives to identify and eliminate waste in their healthcare systems.

“In Virginia, an important component of our state health innovation plan is to reduce utilization of low value medical tests and procedures so that we can free up needed resources to advance high value care. Our medical and business communities want to do what they can to better ensure that patients and providers are choosing wisely when it comes to healthcare,” said Beth Bortz, President and CEO of VCHI.

The Health Waste Calculator is an analytic tool that is powered by Milliman’s MedInsight software and encapsulates Value-Based Insurance Design (VBID) Health’s market knowledge on wasteful healthcare spending. The tool identifies and quantifies the use of unnecessary or potentially harmful clinical services, including those defined by national initiatives such as the U.S. Preventive Services Task Force and Choosing Wisely. Using the Health Waste Calculator, each state will be able to report on countless aspects of healthcare but the typical starting points are reports by county, region, or even health system, using these metrics:

• Percentage of individuals exposed to at least one potentially wasteful service.
• Percentage of potentially wasteful services in each market, what is called the Waste and Quality indices.
• Cost metrics of the waste impact as a baseline for them to work from as they implement strategies to realize greater efficiencies.

Mark Fendrick, Director of VBID, stated, “the achievement of Triple Aim in healthcare is to improve quality, enhance patient experience, and lower costs, which requires us to spend more on evidence-based services and less on those that do not produce health. The MedInsight Health Waste Calculator can identify, and potentially reduce, the use of low value services to free up needed resources to reduce important gaps in care.”

MedInsight is used by over 300 health plans, employers, at-risk providers/ACOs, state governments, community health coalitions, and third party administrators.

For more information about Milliman’s MedInsight products, click here.






Potential Overuse of Coronary Angiography Amongst Asymptomatic Population

Sticky, Avoidable Events, Healthcare Analytics

Coronary angiography is a key diagnostic tool in the management of patients with coronary artery disease. Coronary angiography is used to identify narrowing in coronary arteries and is used for decision support to ascertain the need for revascularization to minimize the risk of myocardial infarction. Various research studies have evaluated trends in the use and results of coronary angiography as key contributory factors to the variations seen in rates for coronary revascularization across the US. For example, a large study by Chan, et al. highlighted that inappropriate revascularization rates ranged from 0% to 55% in different facilities[i]. The use of diagnostic angiography has been advocated in different guidelines to ensure that the revascularization is warranted, however, this approach does not address inappropriate selection of patients for a diagnostic angiography.

According to the American Heart Association (AHA), coronary angiography is not recommended in patients who are at low risk according to clinical criteria and who have not undergone prior non-invasive risk testing. Coronary angiography is also not recommended in asymptomatic patients with no evidence of ischemia or non-invasive testing.[ii]  However, various studies observe that current practice patterns are not consistent with these recommendations. Abdallah, et al. in their review of the National CathPCI registry found that, out of 790,601 elective coronary angiographies performed on patients with no history of Coronary artery disease (CAD), approximately 291,586 angiographies were not preceded by any stress test. Amongst these patients with no stress test prior to coronary angiographies, approximately 38.5% did not have symptoms of angina.[iii]

Other studies have highlighted the variable diagnostic yield of elective diagnostic angiographies. A 2010 Patel et al. study of the cardiovascular national registry found that, in patients without known heart disease who underwent elective invasive angiography, 37.6% had obstructive coronary artery disease.[iv] This implies that nearly two-thirds of patients had a negative result for obstructive coronary artery disease. Amongst the group with no obstructive heart disease, 30.0% were noted to have no symptoms, including no angina.

False positive findings in inappropriate diagnostic tests can trigger subsequent treatment. Referred to as “diagnostic-therapeutic cascade,” inappropriate angiography in asymptomatic patients increases the likelihood of performing inappropriate coronary artery revascularizations and other associated medical utilization. Bradley, et al. studied 544 hospitals that performed more than 1 million elective coronary angiograms and more than 200,000 elective percutaneous coronary interventions (PCI) between 2009 and 2013, and found that 25.1% of patients were asymptomatic at the time of angiography. The proportion of angiographies performed on asymptomatic patients by individual hospitals ranged from 1% to 73.6%. Hospitals with higher rates of angiography performed on asymptomatic patients also had higher rates of inappropriate PCIs and lower rates of appropriate PCIs.v

Considering the volume and cost of these procedures, it is important to reinforce appropriate patient selection for coronary angiography.

To study the practice pattern of coronary angiography and associated services, we reviewed administrative claims data of a Midwestern US Health plan for the July 2011-June 2012 time period. MedInsight’s Health Waste Calculator tool was used to identify necessary, likely wasteful, and wasteful angiography service units. Angiography services for members who had administrative data evidence for cardiac symptoms or high risk diagnosis (ischemic and coronary heart disease, diabetes, and peripherally artery disease) were assigned as necessary services. The analysis in Table 1 shows that 99% of angiography services (n=11,110) met the necessary criteria, while 1% of services were determined to be either likely wasteful or wasteful (n=95), having no evidence for the presence of any the underlying symptoms or diagnosis that would classify these as necessary.

Further analysis highlighted that, amongst the necessary services, 37% were conducted for acute conditions (acute coronary syndrome, angina, and myocardial infarction), while 63% of services were conducted for non-acute conditions. Amongst the inappropriate angiography services, 37% of members received a cardiac stress test prior to the angiography, whereas the remaining 63% of members did not. Due to the absence of clinical data, such as from an Electric Medical Records (EMR) system, we assigned the non-necessary services with a prior stress test as “Likely Wasteful” and the remaining services with no prior stress test as “Wasteful.”

Table 1: Profile of Angiography ServicesProfile of Angiography Services

To study if members with wasteful angiography had other associated costs of follow-up and cardiac revascularization, we reviewed the utilization experience of members with wasteful angiography services for a period of six months after the wasteful service date. Table 2 summarizes the results of this analysis. We restricted our analysis to utilization for cardiac-related procedures and encounters. We found that cardiac revascularizations (including CABG and PCI) and other cardiac procedures comprised 11% of all cardiac related services, with a combined cost of $217,840 (35% of all allowed cardiac related costs). Follow-up diagnostic angiographies and cardiac imaging comprised another 27% of cardiac related services, with a combined cost of $89,290 (14% of all allowed cardiac related costs).

Table 2: Post-Angiography Utilization by Members with Wasteful Services

Post-Angiography Utilization by Members with Wasteful Services

While the results of this analysis are consistent with the low end of what has been reported for inappropriate coronary angiographies (1% to 73.6%)v, it raises concern over the need to improve patient selection for coronary angiography. It is important to note that claims data alone is not sufficient to identify wasteful angiography services with absolute certainty. However, even with conservative estimates, these findings confirm the significant resource use and expenses after angiography and the potential medical waste in the subsequent period.

Sources

[i] Chan et al, Appropriateness of Percutaneous Coronary Intervention JAMA. 2011;306(1):53-61. Available at: http://jama.jamanetwork.com/article.aspx?articleid=1104058

[ii] Jeffery L Anderason, Jonathan L. Haperin. “Guideline for the Diagnosis and Management of Patients with Stable Ischemic Heart Disease”; Journal of the American College of Cardiology. 2012; 60(24):2564-2603. Available at http://content.onlinejacc.org/article.aspx?articleid=1391403

[iii] Mouin S. Abdallah, John A. Spertus. “Symptoms and Angiographic Findings of Patients
Undergoing Elective Coronary Angiography Without Prior Stress Testing”; American Journal of Cardiology. Available at http://www.ajconline.org/article/S0002-9149(14)01112-6/pdf

[iv] Manesh R. Patel, Eric D. Peterson. “Low Diagnostic Yield of Elective Coronary Angiography”; New England Journal of Medicine 2010 March 11; 362(10): 886–895. Available at http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3920593/pdf/nihms540259.pdf

[v]Bradley SM, Spertus JA. “Patient selection for diagnostic coronary angiography and hospital-level percutaneous coronary intervention appropriateness: insights from the National Cardiovascular Data Registry”; JAMA; 2014;174 (10):1630-1639. Available at http://archinte.jamanetwork.com/article.aspx?articleid=1898877






Supporting Patient-Centered Medical Homes through Healthcare Analytics

StickyPCMH, Preventive Care

Although the concept of Patient-Centered Medical Homes (PCMH) has been around for more than 50 years 1, the last decade has seen a revitalization of the PCMH model and an increase in its presence across the nation. The model’s popularity hinges on an approach to providing comprehensive primary care and redesigning healthcare delivery processes. This is accomplished through an emphasis on team-based care delivery, a whole-person approach to patient care, collaborative relationships between individuals and their physicians, and the use of evidence-based medicine and clinical decision support tools.

In 2007, four nationally recognized physician organizations identified seven principles considered foundational to the PCMH2 model:

  1. Personal physicians
  2. Physician directed medical practices
  3. Whole person orientation
  4. Coordinated/integrated care
  5. Quality and safety
  6. Enhanced access
  7. Payment reform

Although the foundational principles of the PCMH have been largely agreed upon, there is no clear mode for how to create a successful PCMH. One of the most widely recognized models in place today is sponsored by the National Committee for Quality Assurance (NCQA), though there are numerous different demonstration and pilot projects in process across the country. As stated by Stange et al. in their Journal of General Internal Medicine article, “…the context for operationalizing the PCMH is still evolving based on what is being learned in many ongoing demonstrations,”2 underscoring the importance of evaluating and incorporating unique geographic, demographic, and economic considerations into the design of any new care model.

Successful care delivery transformation projects, especially PCMH implementation and sustainment activities, require significant emphasis on healthcare analytics to inform quality improvement activities in addition to managing cost and utilization control efforts. The use of structured and routine analysis of healthcare claims-based information enables both established organizations and newly developed PCMHs to receive ongoing feedback on process effectiveness and health outcomes, facilitating rapid-cycle process improvement across the organization.

PCMHs typically focus their analytic resources on operational process improvements and patient outcomes, with the goal of driving improvements in support of the Triple Aim. Successful organizations understand that routine and actionable information is the key to driving improvements. Examples of PCMH-focused analytic approaches being used across the country, which typically focus on cost, utilization, and quality, include but are not limited to the following:

  • Increased Use of Generic Pharmaceuticals: Pharmacy claims data is analyzed to identify areas of opportunity for transitioning members to generic equivalents as a cost-reduction initiative.
  • Appropriate Emergency Department (ED) Usage: Utilization patterns in the ED are evaluated to identify high utilization members (and attributed providers) and high frequency conditions. High rates of ED visits may be related to access and availability issues with primary care physicians, underutilization of urgent care services, or lack of understanding that many conditions are more appropriately addressed by a physician office visit rather than an ED visit.
  • Reduced Hospital Admissions and Readmissions: The disease burden of a population is assessed and abnormal utilization rates are identified to assist primary care providers in focusing on specific conditions and/or groups of patients so they can better manage healthcare status and prevent unnecessary hospitalizations and readmissions using outpatient resources.
  • Consistent Practices across Providers: Provider service patterns are compared to identify variations in care practices that could affect all three categories – cost, utilization, and quality. This information can be used to drive adherence to clinical guidelines and identify re-education opportunities for providers who are more expensive than their peers when treating specific cases (e.g., compare costs to treat diabetes patients across providers within a practice to identify discrepancies in care practices and cost savings opportunities). This type of benchmarking analysis can also be used to identify providers who may have high rates of ED usage, or hospital admissions and readmissions within their patient panel.
  • Targeted Care Management Services: Service consumption and diagnosis patterns are analyzed to identify members in need of care management services, for example, members with multiple chronic conditions, mental health issues, or poly-pharmacy usage. Risk scores and clinical drivers can be used to understand individual patient needs and priorities.
  • Preventive Care and Evidence-Based Measures: Claims data is used to assess whether members are consistently receiving preventive care, such as depression screenings, cancer screenings, diabetic eye exams, HbA1c testing, lipid profiles, and blood pressure screenings. Evidence-based care guidelines are implemented for chronic conditions, such as diabetes and cardiovascular disease, and compliance is tracked through claims data.

Experts in the field cite “a growing body of evidence…that the patient-centered medical home is an effective model to transform primary care and serve as a foundation”3 for other accountable care models. Leveraging data analysis is a crucial component of success for organizations pursuing this type of delivery model transformation.

Sources

1Deborah Piekes et al. Early Evidence on the Patient-Centered Medical Home. Agency for Healthcare Research and Quality, U.S. Department of Health and Human Services 2012; 12-0020-EF.

2Kurt C Stange et al. Defining and Measuring the Patient-Centered Medical Home. J Gen Intern Med 2010; 25(6):601-12.

3Harbrecht, Marjie G, and Lisa M. Latts. Colorado’s Patient-Centered Medical Home Pilot Met Numerous Obstacles, Yet Saw Results Such as Reduced Hospital Admissions. Health Affairs 2012; 31(9):2010-2017.






Top Five Analytic Trends

StickyAnalytics, Benchmarking, Healthcare Analytics, Trend Analytics

It goes without question that the U.S. health insurance industry is in a state of flux.  Americans are buying individual products through health insurance marketplaces, new insurance carriers have entered the market, and Medicaid has been expanded in 29 states and the District of Columbia. These market changes, in addition to other reform provisions already introduced and others just starting to take hold, have subjected the market to an unprecedented level of change.

It is said that insurers like risk but hate uncertainty.  What is for certain today is that the old strategies of accepting good risks and repelling poor risks is no longer a recipe for success.  To thrive in this new environment, health insurers must make smart decisions using data to keep ahead of the competition.

Within that context, here are five areas where Milliman clients are using data and analytics in innovative ways to bring some order to the chaos:

  1. Provider Network Optimization. Despite bending the cost curve, one of the great lessons of the HMO era was that consumers value choice. For years, PPOs competed on network size; employers cared more about network disruption affecting their employees than the cost/volume trade-off. In the face of cost pressures, employers and consumers are now starting to accept that smaller networks may be worth the disruption. To meet this need, plans are deploying sophisticated modeling that combines traditional network access and adequacy measures with reimbursement and quality analytics to develop new “smart” networks.
  1. Value-Based Incentive Programs. It’s widely accepted that fee-for-service reimbursement rewards volume over value. As a replacement for FFS, many payers are promoting value-based incentive strategies that shift reimbursement from fee schedules to bonus pools that pay additional incentives when quality and/ or cost targets are met. Analytics are key to selecting measures, setting thresholds, and assessing provider performance. They also aid providers trying to operate under these new risk arrangements, identifying gaps in care, and benchmarking peer performance.
  1. New Trend Dynamics. While predicting the actual numbers requires the proverbial “crystal ball,” the health insurance industry has a reasonably mature understanding of the drivers of health care cost trend. But things are getting more complicated as physician practice patterns change, populations age but live longer, millions of new consumers flood into the individual and Medicaid markets, and burgeoning innovation (e.g., telemedicine/ telehealth, wearables, smartphones, home visits, retail clinics, etc.) disrupts how and where care is provided. Analytics are key to understanding the “trends in trend” in this new world.
  1. Transparency. The healthcare market has earned a reputation for opaqueness. Consumers are more likely to rely on word-of-mouth when selecting a physician, the price of services depends on who’s paying and has little relationship with the actual cost of services, and information on outcomes and quality is kept locked away from prying eyes. Not so in a post-reform world; consumers can now shop on the basis of price and quality, they can go online and find out how much an appendectomy costs at hospital A or B and which one has a higher success rate, and health plan quality ratings are there for all to see when selecting an exchange plan. Big data and analytics make all of this possible.
  1. Care Management Efficiency. Gone are the days when health insurers had unlimited funding for care management programs. Today, plans must make judicious use of limited administrative dollars to meet medical loss ratio minimums while still managing complex and catastrophic cases.  Analytics help plans optimize their care management programs, prospectively identifying those members most likely to benefit from care management, and then enrolling them in the right program.

With many of their traditional performance management tools neutralized by reform, health insurers have had to get smart about how they leverage data and information: they use analytics to design benefit plans, develop marketing strategies and consumer segmentations, select network providers, develop reimbursement strategies, improve clinical quality, and optimize their remaining cost and quality management tools. In today’s market, how a plan leverages analytics, turning data into actionable information, will make the difference between survival and demise.