Unbiased Analysis of Today's Healthcare Issues

Geographic Variation in the Quality and Cost of Care for Patients with Rheumatoid Arthritis

Written By: Jason Shafrin - Nov• 28•16

This is the title of my latest article in JMCP co-authored with Arijit Ganguli,  Yuri Sanchez Gonzalez, Jin Joo Shim, and Seth A. Seabury.  The paper’s abstract is below.

BACKGROUND: There is considerable push to improve value in health care by simultaneously increasing quality while lowering or containing costs. However, for diseases that are best treated with comparatively expensive treatments, such as rheumatoid arthritis (RA), there could be tension between these aims. In this study, we measured geographic variation in quality, access, and cost for patients with RA, a disease with effective but costly specialty treatments.

OBJECTIVE: To assess the geographic differences in the quality, access, and cost of care for patients with RA.

METHODS: Using large claims databases covering the period between 2008 and 2014, we measured quality of care metrics by metropolitan statistical areas (MSAs) for patients with RA. Quality measures included use of disease-modifying antirheumatic drugs (DMARDs) and tuberculosis (TB) screening before initiating biologic DMARD therapy. Access to care measures included measured detection and the share of patients with RA who visited a rheumatologist. Regression models were used to control for differences in patient demographics and health status across MSAs.

RESULTS: For the 501,376 patients diagnosed with RA, in the average MSA 64.1% of RA patients received a DMARD, and 29.6% of RA patients initiating a biologic DMARD appropriately received a TB screening. Only 17% (73/430) of MSAs comprised the top 2 Medicare Advantage star ratings for DMARD use. Measured detection was 0.59% (IQR = 0.47%-0.71%; CV = 0.355) on average, and 57.6% (IQR = 48%-69%; CV = 0.341) of RA patients visited a rheumatologist. MSAs with the highest DMARD use spent $26,724 (in 2015 U.S. dollars) annually treating patients with RA, $5,428 more (P < 0.001) than low DMARD-use MSAs, largely because of higher pharmacy cost ($5,090 vs. $7,610, P < 0.001). However, MSAs with higher DMARD use had lower RA-related inpatient cost ($1,890 vs. $2,342, P = 0.024).

CONCLUSIONS: There were significant geographic variations in the quality of care received by patients with RA, although quality was poor in most areas. Fewer than 1 in 5 MSAs could be considered high quality based on patient DMARD use. Access to specialist care may be an issue, since just over half of patients with RA visited a rheumatologist annually. Efforts to incentivize better quality of care holds promise in terms of unlocking value for patients, but for some diseases, this approach may result in higher costs.


Insurance value of multiple sclerosis therapies

Written By: Jason Shafrin - Nov• 27•16

An interesting paper from Shih et al. in this month’s edition of AJMC finds that–due to risk aversion–healthy individuals actually place a high value on the development of new treatments for multiple sclerosis.

Results: In the baseline model, Avonex, Tysabri, and Tecfidera generated $46.2 billion of total value to consumers, almost one-third of which accrued to those without MS. The total value to consumers was double manufacturer revenue. Results were qualitatively robust to the use of alternate epidemiological and economic parameters. We found that value to the healthy is positively related to disease severity, and that value to both the sick and the healthy are larger when costs are shared via health insurance.

Conclusions: Theory predicts that treatments for severe disease provide “peace of mind” value to the healthy. Avonex, Tysabri, and Tecfidera have generated significant social value, a large majority of which accrues to consumers. Future economic valuations of medical technology should consider both the potential value to the healthy and the effects of insurance.

Specifically, the authors find that when patients bear the full cost of treatment, the value of the 3 disease modifying treatments (DMTs) to patients with MS totals $11.1 billion, whereas the value to the healthy is $8.9 billion. In other words, 44.5% of health care value from DMTs accrues to healthy individuals.


Precision medicine will increase drug prices and that’s a good thing

Written By: Jason Shafrin - Nov• 22•16

Check out my latest article at MedCity News looking at how the advent of precision medicine will affect patients, regulators and payers.

The article describes how precision medicine will be good for patients.  Precision medicines will likely be more expensive that regular therapies, but–because they are targeted to more specific populations–aggregate spending on pharmaceuticals could increase or decrease.  The article also discusses challenges regulators face when determining the efficacy and safety of precision medicines.


How to regulate precision medicines

Written By: Jason Shafrin - Nov• 21•16

Currently, pharmaceutical treatments that are used in the U.S. need to gain an approval from the FDA.  The FDA’s approval is contingent on a demonstration of efficacy and safety in a randomized controlled trial (RCT).  However, precision medicine makes the standard FDA approval problematic.

As described in Breckenridge et al. (2016), in the precision medicine paradigm, patient response predicted by a specific patient gene or genes.  Thus, while RCTs approach has the advantages of strong internal validity and the establishment of causal inference, RCTs may not be sufficiently powered for treatments targeting rarer genes.

Alternatives include using real-world data such as data from patient registries and electronic medical records.  Alternatively, researchers can used a an adaptive clinical trial design based on Bayesian rather than frequentist statistics.

Patient reported outcomes could even be included in precision medicine trial designs.

Rather than considering patients as a large heterogeneous group, precision medicine provides the possibility of working with smaller, more homogeneous and better-informed patient populations with potentially unique urgencies and needs that are clearly expressed. Direct involvement of patients in regulatory activities has brought to the fore the need to assess symptomatic results using patient-reported outcomes (PROs) even in the development of drugs for the treatment of diseases such as cancer that have clear and highly relevant clinician-reported outcomes.

Another potential problem with precision medicines is their high price tags. If drugs are priced to value, precision medicines could generate significant value for a smaller subset of patients.  Thus, precision medicine is likely to increase the cost per user price to sufficiently reimburse the innovator, but the effect on overall drug spending is unclear as these precision medicines will be used on a smaller patient population.

Precision medicines may be better adapted to “learn and confirm” regulatory approval processes such as the FDA’s Breakthrough Therapy and EU’s PRIME designation.

Finally, patient risk tolerances will likely play a role in whether certain precision medicines are viable in the market.

When medicines have to be taken long before patients are confronted with the morbidity and mortality of their condition, their tolerance for the uncertainties about benefits and harms of new products may be limited, requiring new approaches to long-term monitoring of benefits and risks.

In short, precision medicines offer the opportunity to significantly improve patient outcomes, but regulatory and reimbursement challenges loom.

Optimal Matching Techniques

Written By: Jason Shafrin - Nov• 20•16

In randomized controlled trials, participants are randomized to different groups where each group receives a unique intervention (or control). This process insures that any differences in the outcomes of interest are due entirely to the interventions under investigation.   While RCTs are useful, they are expensive to run, are highly controlled and suffer from their own biases (e.g., attrition bias).

Use of real-world data are an alternative to randomized controlled trials to measure the effect of an intervention. Statistical analysis of the effect of an intervention on patient outcomes must demonstrate that the assignment of the intervention is unrelated to outcomes. In the RCT, this ignorability of treatment assignment occurs due to randomization. In the real world, techniques such as difference in differences, instrumental variables analysis, regression discontinuity, and propensity score patching are often used to eliminate or minimize any endogeneity bias in treatment assigned when using real world data. However, not all methods are the same.

Consider the case of propensity score matching. An article by Fullerton et al. (2016) examines which approaches to propensity score matching work best. The consider three dimensions of propensity score matching:

  • Number of variables included in the match. One approach is to include all covariates in the match, whereas other approaches only match on selected variables likely to influence the outcome of interest. One could select these variables based on expert opinion or previous literature. The Fullerton paper uses a backward selection approach where they run a logistic regression for the effect of the patient characteristic on receipt of treatment. All treatments with a p-value >0.10 are included in the match.
  • Functional form. Most often, a propensity score is calculated based on the probability of receiving treatment conditional on patient characteristics using a logistic regression. The logistic regression is widely used due to its simplicity but all interactions or higher order terms would have to be explicitly specified as functional terms in logistic regressions, which is rarely done for a large number of covariates. Fullerton and co-authors also use a general boosted regression (GBR), a multivariate nonparametric regression technique that can flexibly include nonlinear relationships between the propensity score and a large number of covariates.
  • Number of matches. In this simplest case, each individual receiving the treatment of interest is matched to one and only one individuals based on proximity of propensity score. However, each “treated” individual could also be matched to multiple individuals. Matching on more individuals increases the bias—as the people matched by definition have a larger difference in propensity score—but more power.

Further, there are other types of matching that could be considered.

  • ESLID Matching. In this approach, patient characteristics are categorized into mandatory and optional. Patients must match on the mandatory characteristics and then among those matched, an individual (or individuals) is selected based on closest match on the optional characteristics.
  • Exact matching. This matching occurs when treated individuals are matched to all controls with exactly the same values for the covariates of interest. Clearly this creates a good match, but power decreases.
  • Coarsened exact matching (CEM). In this specification, “the number of matching dimensions is reduced by creating intervals for continuous variables and possibly redefining categorical variables.” For instance, one could match on age deciles which would allow the age levels themselves not to match but age relative to ones peers would be similar.

Using this approach applied to data from Germany on disease management programs for type 2 diabetes (DMPDM2). The authors found the following:

Exact matching methods performed well across all measures of balance, but resulted in the exclusion of many observations, leading to a change of the baseline characteristics of the study sample and also the effect estimate of the DMPDM2. All PS-based methods showed similar effect estimates. Applying a higher matching ratio and using a larger variable set generally resulted in better balance. Using a generalized boosted instead of a logistic regression model showed slightly better performance for balance diagnostics taking into account imbalances at higher moments.

Nevertheless, the authors wisely recommend applying multiple matching techniques to ensure the results are robust to matching specification choice.


Friday Links

Written By: Jason Shafrin - Nov• 17•16

The AMA on value-based drug pricing

Written By: Jason Shafrin - Nov• 16•16

The American Medical Association released a statement yesterday in support of value-based pricing of pharmaceuticals.  However, AMA claims that value-based pricing should follow the following core principles.

  • Value-based prices of pharmaceuticals should be determined by objective, independent entities. They also should be evidence-based and the result of valid and reliable inputs and data that incorporate rigorous scientific methods including clinical trials, clinical data registries, comparative effectiveness research and robust outcome measures that capture short- and long-term clinical outcomes.
  • Processes to determine value-based prices of pharmaceuticals must be transparent, easily accessible to physicians and patients, and provide practicing physicians and researchers a central and significant role. They should also limit administrative burdens on physicians and patients.
  • Those same processes should incorporate affordability criteria to help assure patient affordability and limit system-wide budgetary impact.
  • Value-based pricing of pharmaceuticals should allow for patient variation and physician discretion.

Not only are payers and policymakers getting on board with value-based drug pricing, but providers–like the AMA–and even life sciences companies themselves are getting on board.

In my role as Director of Research at the Innovation and Value Initiative, questions such as how to measure value and the identification of the best methods for tying reimbursement to value are the exactly the questions we are interested in answering with rigorous scientific research.

It’s exciting that there is a growing interest in this area across a broad set of health care stakeholders.

Are quality bonus payments based on hospital readmissions reliable?

Written By: Jason Shafrin - Nov• 15•16

Maybe not.  That is the answer from a study by Thompson et al. (2016).  Using data from the Healthcare Cost and Utilization Project (HCUP) State Inpatient Databases (SID) for six states (AR, FL, IA, MA, NY,WA)  from 2011 to 2013, the authors measure hospital performance reliability for the Hospital Readmission Reduction Program (HRRP).  The define reliability as follows:

Measure reliability is determined by the between-hospital variation in event rates (the signal ) and the within-hospital variation in event rates (the noise ) (Adams 2009; Adams et al. 2010). Reliability (R ) is then estimated as R = signal /(signal + noise ), and it ranges from zero to one, with reliability increasing as this ratio approaches one.

Is reliability a big issue?  Isn’t this just random noise that gets evened out over time?  Statistically the answer is yes, but if reimbursement based on readmissions is tied mostly to random chance rather than quality of care, there is little point in tying reimbursement to readmissions.  Of course, smaller hospitals are less likely to have reliable measures due to a smaller sample size of admissions.

The conditions evaluated by the study were acute myocardial infaction (AMI), congestive heart failure (CHF), and pneumonia (PN); more recently, it has expanded to include readmissions for chronic obstructive pulmonary disease (COPD), total hip and/or knee arthroplasty (THKA), and coronary artery bypass graft surgery (CABG).  For each of these conditions, the authors followed CMS’ approach for measuring readmission rates:

…we used hierarchical logistic regression models to estimate the predicted to expected number of readmissions for each hospital and condition, given their hospital case mix (i.e., the P/E ratio or excess readmission ratio); this ratio was then multiplied by the observed readmission rate for the entire condition-specific cohort. All models were adjusted for the comorbidities used in the CMS risk-adjustment models.

Reliability was measured as the ratio of the between-hospital standard devaiation in average readmission rates divided by the sum of the between-hospital standard deviation in average readmission rates and the within-hospital standard deviation in readmission rates.  The within-hospital variance will vary across hospitals due to both sample size and the proportion of readmissions (which affects the variance as readmissions follow a binomial distribution).

Using a benchmark reliaability of 0.70, the authors found that HRRP reliability was mixed:

  • AMI (R = 0.58), 18.6% of hospitals exceed R>0.70
  • CHF (R = 0.61), 16.6% of hospitals exceed R>0.70
  •  COPD (R = 0.65), 28.1% of hospitals exceed R>0.70
  • PN (R = 0.68), 40.2% of hospitals exceed R>0.70
  • CABG (R = 0.78), 86.6% of hospitals exceed R>0.70
  • THKA (R = 0.94), 100% of hospitals exceed R>0.70

Overall, the authors found that readmission measures for surgical admissions are more reliable than those for medical admissions.  They conclude the following:

For measures to be reliable for group-level comparisons, they must have sufficient between-group variation (e.g., physicians, hospitals, accountable care organizations) and adequate sample size. The absence of either of these elements limits the measure s usefulness in hospital performance profiling. For instance, the condition with the largest sample size, CHF, has among the worst reliability because hospital-level RSRRs do not vary substantially (range: 16.1 36.2 percent). Conversely, RSRRs for CABG have much higher reliability, despite its small sample size, because the variation in RSRRs is much larger (range: 11.4 47.1 percent).

A very interesting study that won a Best of AcademyHealth Annual Research Meeting Award for 2016.

Cancer drug pricing in Europe

Written By: Jason Shafrin - Nov• 14•16

How do Euroepan countries reimburse for pharmaceuticals? A paper by Pauwels et al. (2014) provides an nice summary. I review that article today.

With the exception of Germany, most countries had a national and/or regional drug budget.  Germany is also unique in that only Germany and the UK allow for free pricing, whereas other countries use reference pricing or price negotiation.

Note that the hospital drug market differs from the outpatient pharmacy setting.  In the inpatient setting, most countries except the Netherlands all for price negotiation and only the Netherlands and Germany do not use tendering. According to the Better Pharmacare Coalition:

Tendering is a process in which government, as payer, negotiates the lowest price for a pharmaceutical drug1. In exchange for this low price, the supplier of the drug gets their product listed on the public drug plan. Often, the lowest bidder becomes the sole tender meaning their drug is the only one available in an entire class of drugs (such as statins) to patients on the public drug plan2.


Belgium evaluates drugs based on  therapeutic value, cost-effectiveness, budget impact and positioning in clinical practice, although orphan drugs are exempt from pharmaco-economic investigation.  Reimbursement decisions are made nationally. Unauthorized drug use should be paid by the pharmaceutical company in a compassionate use program.

In France, prices are set through negotiations between the Economic Committee on Health Care Products and the pharmaceutical companies.  Cost effectiveness is used to determine reimbursement eligibility, but in some cases of very severe disease, treatments with modest benefits may be covered.  “Payback mechanisms are imposed in case of unexpected sales per therapeutic indication based on the budget impact evaluation. New drugs are exempt from these agreements for a period of time which depends on the level of medical benefit.”

Germany grants full reimbursement for every prescription drug that has received marketing authorization.  “Whether or not a drug provides added benefit determines if the drug falls within the internal reference system or can be the subject of price negotiations. Also patented drugs can be included in reference groups. Orphan drugs are exempt from this early benefit assessment if annual sales in the preceding 12 months did not exceed €50 million.” Germany relies on its Institut für Qualität und Wirtschaftlichkeit im Gesundheitswesenis (IQWiG) to conduct health technology assessments.   The Pauwels paper states “performance-based risk-sharing arrangements exist in the form of rebate contracts between sickness funds and pharmaceutical companies to accelerate market access of drugs with doubtful value.”

Italy finances cancer pharmaceuticals based on a system that has national, regional and local components. “Although pricing and reimbursement are a national competence, regions can charge copayments to patients, resulting in price differentiation across the country.” Drug prices are negotiated, but Italy also has fixed drug spending ceilings: “The expenditures to pharmacological products in primary care cannot exceed 11.35 %, while pharmaceutical expenses in the hospital have to remain below 3.5 % of the national and regional health care budget.   In case of overspending, companies have to pay back the complete budget excess in primary care and half of the budget excess in hospitals.”

The Netherlands uses external reference pricing is used to set a maximum price for drugs based on the average price in other European countries.   The pricing considers both both patented and generic drugs and generally compares drugs to therapeutic equivalents.  If therapeutic equivalents are not available, cost effectiveness methods are used to determine price with a willingness to pay threshold of €20,000 to 80, 000/QALY; orphan drugs, however, can apply for an exception.  The Netherlands has introduced performance-based risk-sharing arrangements (e.g., coverage with evidence development, performance-linked reimbursement).

Poland sets prices based largely on internal reference prices (i.e., the price of branded and unbranded products in the same therapeutic class).  HTA is used and cost-effectiveness thresholds are set by law, but the ministry of health can ignore the Polish HTA’s advice.  Beginning in 2012, a new Reimbursement Act will allow for performance-based risk-sharing.

Sweden has  21 county councils all having the responsibility over their local health care budget and drug formularies.  Although reimbursement is generally determined nationally, local councils can cover additional treatments not covered nationally.  Sweden uses cost effectiveness analysis to inform price, but “Sweden is one of the few countries that apply a societal perspective to consider costs and benefits of healthcare in cost-effectiveness analysis.”

The United Kingdom sets prices of branded pharmaceuticals based on the Pharmaceutical Price Regulation Scheme (PPRS).  “In England, the National Health Service (NHS) is entrusted with allocation of the budget to the Primary Care Trusts (PCT), which are grouped within regional Strategic Health Authorities (SHA) [68]. The PCTs are responsible for delivering health care and health improvements within the local area: they have their own budgets and set their own priorities.”  England and Wales rely on the National Institute for Health and Care Excellence (NICE) to determine cost effectiveness with a cost/QALY threshold of generally around £20,000–30,000 (€23, 600–35,400) per QALY.  In Scotland, the HTA agency is the Scottish Medicines Consortium (SMC) and in Wales there is the All Wales Medicines Strategy Group (AWMSG).  Almost all drugs allowed by the NHS to be prescribed are completely reimbursed.  The UK is exploring including more value-based pricing schemes starting in 2014.  The UK does use performance based pricing: “Performance-based risk-sharing arrangements are designed for oncology drugs such as bortezomib for multiple myeloma: the NHS will continue funding after the fourth cycle when a reduction of 50 % in serum plasmaprotein M is obtained.”


Cost-Effectiveness in Health and Medicine

Written By: Jason Shafrin - Nov• 13•16

How do you do cost effectiveness the right way?  Peter Neumann–a colleague of mine at Precision Health Economics–edits a book to explain how to do just that.  Neumann and Gillian D. Sanders, Louise B. Russell, Joanna E. Siegel, and Theodore G. Ganiats have produced a second edition of their classic text Cost-Effectiveness in Health and Medicine.  The website’s description is as follows:

Twenty years after the first edition of COST-EFFECTIVENESS IN HEALTH AND MEDICINE established the practical benchmark for cost-effectiveness analysis, this completely revised edition of the classic text provides an essential resource to a new generation of practitioners, students, researchers, and policymakers.

Produced by the Second Panel on Cost-Effectiveness in Health and Medicine–a team of 13 experts from fields including decision science, economics, ethics, psychology, and medicine–this new edition is a comprehensive guide to the use of cost-effectiveness analysis as an evaluative tool at the institutional and policy levels. As health care systems face increasing pressure to derive maximum value from expenditures, the guidelines in this new text represent not just the best information available, but a vital guide to health care decision-making in a challenging new era.

Completely revised and enriched with examples and expanded coverage, this second edition of COST-EFFECTIVENESS IN HEALTH AND MEDICINE builds on its predecessor’s excellence, offering required reading for both analysts and decision makers.

I have not read this book yet, but look forward to doing so.