Background: Medication adherence is associated with lower health care utilization and savings in specific patient populations; however, few empirical estimates exist at the population level. Objective: The main objective of this study was to apply a data-driven approach to obtain population-level estimates of the impact of medication nonadherence among Medicare beneficiaries with chronic conditions. Research Design: Medicare fee-for-service (FFS) claims data were used to calculate the prevalence of medication nonadherence among individuals with diabetes, heart failure, hypertension, and hyperlipidemia. Per person estimates of avoidable health care utilization and spending associated with medication adherence, adjusted for healthy adherer effects, from prior literature were applied to the number of nonadherent Medicare beneficiaries. Subjects: A 20% random sample of community-dwelling, continuously enrolled Medicare FFS beneficiaries aged 65 years or older with Part D (N=14,657,735) in 2013. Measures: Avoidable health care costs and hospital use from medication nonadherence. Results: Medication nonadherence for diabetes, heart failure, hyperlipidemia, and hypertension resulted in billions of Medicare FFS expenditures, millions in hospital days, and thousands of emergency department visits that could have been avoided. If the 25% of beneficiaries with hypertension who were nonadherent became adherent, Medicare could save $13.7 billion annually, with over 100,000 emergency department visits and 7 million inpatient hospital days that could be averted. Conclusion: Medication nonadherence places a large resource burden on the Medicare FFS program. Study results provide actionable information for policymakers considering programs to manage chronic conditions. Caution should be used in summing estimates across disease groups, assuming all nonadherent beneficiaries could become adherent, and applying estimates beyond the Medicare FFS population.

We develop a method that embeds signals about consumers' knowledge to evaluate prospective choice architecture policies. We analyze three proposals for U.S. Medicare prescription drug insurance markets: (i) menu restrictions, (ii) personalized information, and (iii) defaulting consumers to cheap plans. We link administrative and survey data to identify informed enrollment decisions that proxy for preferences of observationally similar misinformed consumers. Results suggest that each policy yields winners and losers, with the menu restrictions harmful to most but personalized information beneficial to most. These results are robust across signals of consumers' knowledge but differ from the benchmark that excludes such signals.

OBJECTIVES:To evaluate the effects of formulary restrictions on utilization and costs of oral hypoglycemic agents (OHAs), statins, and renin-angiotensin system (RAS) antagonists among low-income subsidy (LIS) recipients in Medicare Part D plans.STUDY DESIGN:We analyzed a 5% sample of 2012 Medicare data from the Chronic Conditions Data Warehouse together with a customized dataset capturing beneficiaries' histories of plan assignment.METHODS:We constructed 3 nonexclusive study cohorts comprising of users of OHAs, statins, and RAS antagonists. Eligible study subjects were LIS recipients randomized to benchmark plans. Formulary restrictions of interest were noncoverage, prior authorization, and step therapy. Study outcomes included generic dispensing rate (GDR), mean cost per prescription fill, and medication adherence based on proportion of days covered (PDC). Random intercept regression models were performed to estimate the effects of formulary restrictions on the study outcomes by drug class.RESULTS:After covariate adjustment, beneficiaries who were subject to formulary restrictions on brand name pioglitazone and single-source brand name dipeptidyl peptidase-4 inhibitors (saxagliptin, sitagliptin, and sitagliptin-metformin) had a GDR 3 percentage points higher and a cost per prescription fill $10.8 less, but similar PDC compared with those who faced no restrictions. Restricting access to brand name atorvastatin and single-source brand name statins (rosuvastatin and ezetimibe-simvastatin) was associated with a GDR 14.9 percentage points higher and a cost per prescription fill $29.6 less, but with no impact on PDC. Restricting use of single-source brand name RAS antagonists (olmesartan, valsartan, and valsartan-hydrochlorothiazide) was associated with a GDR 15.0 percentage points higher, a cost per prescription fill $27.2 less, and a PDC 1.3 percentage points lower.CONCLUSIONS:Placing formulary restrictions on brand name drugs shifts utilization toward generic drugs, lowers the overall cost per prescription fill, and has minimal impact on overall adherence for OHAs, statins, and RAS antagonists among LIS recipients.

Background: The improvement of medication use is a critical mechanism that accountable care organization (ACO) could use to save overall costs. Currently pharmaceutical spending is not part of the calculation for ACO-shared savings and risks. Thus, ACO providers may have strong incentives to prescribe more medications hoping to avoid expensive downstream medical costs.Methods: We designed a quasinatural experiment study to evaluate the effects of Pioneer ACOs on Medicare Part D spending and utilization. Medicare fee-for-service beneficiaries with Part D drug coverage who were aligned to a Pioneer ACO were compared with a random 5% sample of non-ACO beneficiaries. Outcomes included changes in Part D spending, number of prescription fills, percent of brand medications, and total Part A and B medical spending. We utilized a generalized linear model with a difference-in-differences approach to estimate 2011-2012 changes in these outcomes among beneficiaries aligned with Pioneer ACOs, adjusting for all beneficiary-level demographics, income and insurance status, clinical characteristics, and regional fixed effects.Results: Being in an ACO did not significantly affect Part D spending (-$23.52; P = 0.19), total prescriptions filled (-0.12; P = 0.27), and the percent of claims for brand-name drugs (0.06%; P = 0.23). The ACO group was associated with savings in Parts A and B spending of $345 (P < 0.0001) per person per year.Conclusions: We found that beneficiaries aligned to Pioneer ACOs were not associated with changes in pharmaceutical spending and use, but were associated with savings in Parts A and B spending in 2012.

Background: Although the adoption of e-prescriptions among physicians has increased substantially under the Medicare Improvements for Patients and Providers Act and Meaningful Use programs, little is known of its impact on patient outcomes. Objective: To examine the impact of e-prescribing on emergency visits or hospitalizations for diabetes-related adverse drug events (ADEs) including hypoglycemia. Design: This is a prospective, observational cohort study with patient fixed effects. Setting: 2011–2013 fee for service Medicare. Patients: In total, 3.1 million Medicare fee for service, Part D enrolled beneficiaries over age 66 with diabetes mellitus and at least 90 days of antidiabetic medications. Measurements: E-prescribing was measured as the percentage of all prescriptions a person received transmitted to the pharmacy electronically. The outcome measure was the occurrence of an emergency department (ED) visit or hospitalization for hypoglycemia or diabetes-related ADE. Results: Unadjusted results show that there were 21 ADEs per 1000 beneficiaries that had ≥75% of their medications e-prescribed. Beneficiaries with lower e-prescribing levels had significantly higher numbers of ADEs. We found a robust association between the greater use of electronic prescriptions in the outpatient setting and the lower risk of an inpatient or ED visit for an ADE event among Medicare beneficiaries with diabetes in our adjusted analysis. At the e-prescribing threshold of 75% and above, significant reductions in ADE risk can be seen. Limitations: As an observational study, the results show an association but do not prove causation. Conclusions: Use of e-prescribing is associated with lower risk of an ED visit or hospitalization for diabetes-related ADE.

The purpose of this rejoinder is to clarify key areas of agreement and disagreement with Abaluck and Gruber and address aspects of their reply to our comment, both of which appear in the December 2016 issue of the American Economic Review. Readers of our exchange may wonder how we can reach such divergent conclusions from analyzing the same data. In this rejoinder we show how. We demonstrate that Abaluck and Gruber’s criticism of our analysis is based on their mistaken claims about theory and empirics, their omission of key facts, and their emphasis on results that obscure our many areas of agreement.

Introduction: We examined trends in the use of observation services and the relationship between index service type (observation services, emergency department [ED] visits, inpatient stays) and both clinical outcomes and Medicare payments.Methods: We created a yearly cohort panel of Medicare beneficiaries with chest pain. We evaluate the relationships between index service type and 30-day clinical outcomes using a multinomial logit model and between index service type and Medicare payments using generalized linear models.Results: In 2009, 24% of patients with chest pain received observation services; this rose to 29% in 2011. Conversely, 20% were treated as hospital inpatients in 2009; this fell to 16% in 2011. In the adjusted analysis, the risk of 30-day return to the hospital was 7% less ( 95% confidence interval, 5%-8%) for those receiving observation services as compared with inpatients. Average Medicare payments ranged from $3032 for beneficiaries initially treated in the ED to $3885 for those initially treated in observation to $6545 for those initially treated as inpatients.Discussion: Patients treated in observation are less likely than those treated in the ED or as inpatients to have an adverse event within 30 days. Adjusted Medicare payments, including the index stay and the subsequent 30 days, were substantially less for those treated in observation as compared with those treated as inpatients, but more than for those treated and released from the ED. Higher rates of observation service use do not appear to be negatively affecting patient outcomes and may lower costs relative to inpatient treatment. Published by Elsevier Inc.

BACKGROUND:Medicare coverage of skilled nursing facility (SNF) care requires that beneficiaries have a 3-night inpatient stay in the prior 30 days to be eligible. Time spent by beneficiaries receiving hospital-based observation services does not count toward this requirement.OBJECTIVES:To examine the frequency of Medicare beneficiary discharge from hospital-based observation services to SNFs and its impact on Medicare coverage.STUDY DESIGN:Retrospective cohort study.RESULTS:In 2010, 195,068 community-dwelling beneficiaries received hospital-based observation services. Beneficiaries were overwhelmingly (96.5%) discharged back to the community without home health services. Only 1.2% (2319) were discharged to non-covered SNFs, while 0.6% (1196) were discharged to covered SNFs. Patients discharged to SNFs experienced longer lengths of stay (LOS) than those discharged back to the community (34.9 hours vs 25.5 hours; P<.01). Approximately one-fourth of beneficiaries discharged to SNFs had an observation LOS of 48 hours or more.CONCLUSIONS:While only a small minority of community-dwelling Medicare beneficiaries who received hospital-based observation services in 2010 were discharged to an SNF not covered by Medicare, the implications for these patients and the associated costs deserve attention. These findings have important implications for Medicare's observation service and 2-midnight policies.

We study whether people became less likely to switch Medicare prescription drug plans (PDPs) due to more options and more time in Part D. Panel data for a random 20 percent sample of enrollees from 2006--2010 show that 50 percent were not in their original PDPs by 2010. Individuals switched PDPs in response to higher costs of their status quo plans, saving them money. Contrary to choice overload, larger choice sets increased switching unless the additional plans were relatively expensive. Neither switching overall nor responsiveness to costs declined over time, and above-minimum spending in 2010 remained below the 2006 and 2007 levels.

Background: Information is limited regarding utilization patterns and costs for chronic kidney disease-mineral and bone disorder (CKD-MBD) medications in Medicare Part D-enrolled dialysis patients.Study Design: Retrospective cohort study.Setting & Participants: Annual cohorts of dialysis patients, 2007-2010.Predictors: Cohort year, low-income subsidy status, and dialysis provider.Outcomes: Utilization and costs of prescription phosphate binders, oral and intravenous vitamin D analogues, and cinacalcet.Measurements: Using logistic regression, we calculated adjusted odds of medication use for low-income subsidy versus non-low-income subsidy patients and for patients from various dialysis organizations, and we report per-member-per-month and average out-of-pocket costs.Results: Phosphate binders (similar to 83%) and intravenous vitamin D (77.5%-79.3%) were the most commonly used CKD-MBD medications in 2007 through 2010. The adjusted odds of prescription phosphate-binder, intravenous vitamin D, and cinacalcet use were significantly higher for low-income subsidy than for non-low-income subsidy patients. Total Part D versus CKD-MBD Part D medication costs increased 22% versus 36% from 2007 to 2010. For Part D-enrolled dialysis patients, CKD-MBD medications represented similar to 50% of overall net Part D costs in 2010.Limitations: Inability to describe utilization and costs of calcium carbonate, an over-the-counter agent not covered under Medicare Part D; inability to reliably identify prescriptions filled through a non-Part D reimbursement or payment mechanism; findings may not apply to dialysis patients without Medicare Part D benefits or with Medicare Advantage plans, or to pediatric dialysis patients; could identify only prescription drugs dispensed in the outpatient setting; inability to adjust for MBD laboratory values.Conclusions: Part D net costs for CKD-MBD medications increased at a faster rate than costs for all Part D medications in dialysis patients despite relatively stable use within medication classes. In a bundled environment, there may be incentives to shift to generic phosphate binders and reduce cinacalcet use. (C) 2014 by the National Kidney Foundation, Inc. Published by Elsevier Inc. All rights reserved.

Despite extensive use of prescription medications in ESRD, relatively little is known about the participation of Medicare ESRD beneficiaries in the Part D program. Here, we quantitated the sources of drug coverage among ESRD beneficiaries and explored the Part D plan preferences of ESRD beneficiaries with regard to deductibles, coverage gaps, and monthly premiums. We obtained data on beneficiary sources of creditable coverage, characteristics of Part D plans, demographics, and residence from the Centers for Medicare and Medicaid Chronic Condition Data Warehouse and identified beneficiaries with ESRD from the US Renal Data System. We found that a substantial proportion (17.0%) of ESRD beneficiaries lacked a known source of creditable drug coverage in 2007 and 64.3% were enrolled in Part D. Of those enrolled, 72% received the Medicare Part D low-income subsidy. ESRD beneficiaries who enrolled in standalone Part D plans without the assistance of the low-income subsidy tended to prefer more comprehensive coverage options. In conclusion, more outreach is needed to ensure that beneficiaries who lack coverage obtain the coverage they need and that ESRD beneficiaries join the best plans for managing their disease and accompanying comorbid conditions.

BACKGROUND:Cardiovascular disease (CVD) is a major source of mortality and morbidity in dialysis patients. Population-level descriptions of CVD medication use are lacking in this population.STUDY DESIGN:Retrospective cohort study.SETTING & PARTICIPANTS:Adult dialysis patients in the United States, alive on December 31, 2006, with Medicare Parts A and B and enrollment in Medicare Part D continuously in 2007.PREDICTOR:CVDs and demographic characteristics.OUTCOME:≥1 prescription fill during follow-up (2007).MEASUREMENTS:Average out-of-pocket costs per user per month and average total drug costs per member per month were calculated.RESULTS:Of 225,635 dialysis patients who met inclusion criteria during the entry period, 70% (n = 158,702) had continuous Part D coverage during follow-up. Of these, 76% received the low-income subsidy. β-Blockers were the most commonly used CVD medication (64%), followed by renin-angiotensin system inhibitors (52%), calcium channel blockers (51%), lipid-lowering agents (44%), and α-agonists (23%). Use varied by demographics, geographic region, and low-income subsidy status. For CVD medications, mean out-of-pocket costs per user per month were $3.44 and $49.59 and mean total costs per member per month were $124.02 and $110.32 for patients with and without the low-income subsidy, respectively.LIMITATIONS:Information was available for only filled prescriptions under the Part D benefit; information for clinical contraindications was lacking, information for over-the-counter medications was unavailable, and medication adherence and persistence were not examined.CONCLUSIONS:Most Medicare dialysis patients in 2007 were enrolled in Part D, and most enrollees received the low-income subsidy. β-Blockers were the most used CVD medication. Total costs of CVD medications were modestly higher for low-income subsidy patients, but out-of-pocket costs were much higher for patients not receiving the subsidy. Further study is warranted to delineate sources of variation in the use and costs of CVD medications across subgroups.

OBJECTIVE:We sought to evaluate several statistical modeling approaches in predicting prospective total annual health costs (medical plus pharmacy) of health plan participants using Pharmacy Health Dimensions (PHD), a pharmacy claims-based risk index. METHODS:We undertook a 2-year (baseline year/follow-up year) longitudinal analysis of integrated medical and pharmacy claims. Included were plan participants younger than 65 years of age with continuous medical and pharmacy coverage (n = 344,832). PHD drug categories, age, gender, and pharmacy costs were derived across the baseline year. Annual total health costs were calculated for each plan participant in follow-up year. Models examined included ordinary least squares (OLS) regression, log-transformed OLS regression with smearing estimator, and 3 two-part models using OLS regression, log-OLS regression with smearing estimator, and generalized linear modeling (GLM), respectively. A 10% random sample was withheld for model validation, which was assessed via adjusted r, mean absolute prediction error, specificity, and positive predictive value. RESULTS:Most PHD drug categories were significant independent predictors of total costs. Among models tested, the OLS model had the lowest mean absolute prediction error and highest adjusted r. The log-OLS and 2-part log-OLS models did not predict costs accurately as the result of issues of log-scale heteroscedasticity. The 2-part model using GLM had lower adjusted r but similar performance in other assessment measures compared with the OLS or 2-part OLS models. CONCLUSION:The PHD system derived solely from pharmacy claims data can be used to predict future total health costs. Using PHD with a simple OLS model may provide similar predictive accuracy in comparison to more advanced econometric models.