Publications

OBJECTIVE: To evaluate trends in outpatient visits among Medicare beneficiaries with established AF before and after the pandemic and to investigate differences in outpatient care utilization patterns based on baseline utilization levels.

RESULTS: We analyzed 2018-2022 Medicare data to assess outpatient visit trends among 124,483 beneficiaries with atrial fibrillation (AF), categorizing them into quartiles based on pre-pandemic in-person visit frequency. We found that in-person visits declined while telehealth use increased across all groups during the pandemic. After the pandemic, total visits remained below pre-pandemic levels for higher-utilizing groups after the pandemic.

BACKGROUND: Falls and related injuries (FRI) pose a large burden among older adults with depression. Proactively identifying individuals at high FRI risk enables timely and tailored interventions, reducing unnecessary health care resource utilization. However, prior prediction models relied on fixed time intervals and failed to capture dynamic changes in health status over time.

OBJECTIVES: To develop and validate machine-learning algorithms (i.e., elastic net, random forest, and gradient boosting machine) for predicting 3-month FRI risk among older adults with depression.

METHODS: This prognostic modeling study included fee-for-service Medicare beneficiaries aged 65 years or older with a depression diagnosis in 2017. Beneficiaries were followed in 3-month episodes from the first depression diagnosis until the earliest of death, hospice services or nursing facility utilization, switching to Medicare Advantage plans, or the end of the study period (i.e., December 31, 2019). A total of 261 time-varying predictors, spanning patient-, provider-, health system- and region-related factors, were updated every 3 months to predict incident FRI risk in the subsequent 3 months. We assessed prediction performance using c-statistics and stratified patients into different risk subgroups using the best-performing model.

RESULTS: Among 274,268 eligible beneficiaries, the mean age was 74.6 (standard deviation [SD] = 7.2) years, 32.0% were male, 85.2% were White, and 15.1% experienced at least one FRI event throughout the study period. Using the random forest model (c-statistics = 0.68), 68.9% of the actual FRI cases were captured in the top three deciles of predicted risk. Individuals in the bottom seven deciles had a minimal FRI incidence (< 1.7%). Key predictors included frailty, age, prior FRI history, and daily dose of antidepressants.

CONCLUSION: Using a nationally representative cohort and time-varying predictors, our model offers a practical approach for efficiently identifying older adults at high FRI risk, which can be updated over time. This approach can inform clinical decision-making and optimize the allocation of fall prevention resources.

IMPORTANCE: Angiotensin II receptor blockers (ARBs) are common treatments for hypertension, heart failure, and chronic kidney disease. From 2018 to 2019, hundreds of valsartan, losartan, and irbesartan products were recalled due to ingredient impurities.

OBJECTIVE: To estimate the impact of the 2018 to 2019 ARB shortages on medication adherence, switches to alternatives, and associated drug spending up to 18 months.

DESIGN, SETTING, AND PARTICIPANTS: This longitudinal cohort study with a difference-in-differences (DiD) analysis used pharmacy claims data from IQVIA's all-payer Formulary Impact Analyzer dataset from July 2017 to January 2020, comprising prerecall users of valsartan, irbesartan, and losartan vs similar nonrecalled medications (other ARBs, angiotensin-converting enzyme inhibitors [ACEIs]). Analyses were conducted from November 2023 to October 2025.

EXPOSURES: Use of the recalled drugs (valsartan, irbesartan, and losartan) at baseline vs comparison antihypertensives (nonrecalled ARBs, ACEIs).

MAIN OUTCOMES AND MEASURES: Mean proportion of days covered for ARBs and ACEIs, switches to alternatives, medication gaps of 30 or more days, and associated drug spending (insurer and patient out-of-pocket costs).

RESULTS: For 13.8 million ARB users (median [IQR] age in 2018, 66 [56-74] years; 54.8% female) vs 23.4 million comparison drug users (median [IQR] age in 2018, 62 [54-72] years; 46.0% female), mean proportion of days covered changed by 0.55 percentage points (pp; 95% CI, 0.34-0.76 pp) within 18 months. Relative changes in gaps of 30 or more days, insurer drug spending, and patient out-of-pocket drug spending changed by less than 5% (relative changes of -2.5%, 0.6%, and 3.7%, respectively). ARB users experienced an increase in medication switches in the 90 days after the valsartan recall (DiD estimate: 8.46 pp; 95% CI, 8.30-8.63 pp; 229.0% relative increase). Smaller increases in switching occurred after the first irbesartan and first losartan recalls (DiD estimate: 1.20 pp; 95% CI, 1.12-1.27 pp; 32.4% relative increase). The proportion of individuals switching was greater among those with Medicare (DiD estimate: 9.49 pp; 95% CI, 9.28-9.72 pp; 256.8% relative increase) or third-party insurance (DiD estimate: 7.81 pp; 95% CI, 7.57-8.04 pp; 210.8% relative increase) vs Medicaid fee-for-service insurance (DiD estimate: 2.54 pp; 95% CI, 2.31-2.77 pp; 43.1% relative increase) or among customers paying with cash (DiD estimate: 3.42 pp; 95% CI, 3.22-3.61 pp; 87.1% relative increase).

CONCLUSIONS AND RELEVANCE: This cohort study shows that access to alternatives may have mitigated gaps in treatment during the 2018 to 2019 ARB recalls and drug shortages. Potential disparate impacts among certain subgroups highlight the need for policies to mitigate financial and other systematic access barriers to receiving health care during drug shortages.

BACKGROUND: Drug shortages are a global concern that poses risks to clinical care and health systems. Lower prices and market pressures are often cited as drivers of drug shortages; thus, a commonly proposed policy lever is to increase drug prices. However, global evidence to evaluate the association between shortages and prices is lacking. This research aims to fill the gap by examining the global markets of drug shortages.

METHOD: We included 25 global shortage markets by reviewing publications on drug shortages from 2013 to 2023. We used quarterly pharmaceutical sales data across 74 countries/regions from the IQVIA MIDAS quarterly sales volume data database from Q3 2011 to Q3 2022. Our outcome of interest was shortage intensity, defined as the drug shortage duration between its onset and the end of meaningful shortages. To assess the impact of price on the occurrence and intensity of shortages, we used zero-inflated negative binomial regression, with price as the independent factor and included gross domestic product (GDP) per capita and the number of manufacturers as covariates. P values less than a Bonferroni-corrected significance level (0.00625) were considered statistically significant.

RESULTS: Out of 1096 subjects (each representing a drug market in a specific country), 65% (712 subjects) experienced meaningful shortages with a median shortage intensity of 10 quarters. We found price was not a significant predictor of either shortage odds (p=0.044) or intensity (p=0.066). Factors significantly associated with increased odds of a non-shortage occurrence included a unit increase in GDP per capita (adjusted OR (aOR): 1.707, 95% CI 1.428 to 2.040) and in the number of manufacturers (aOR: 4.038, 95% CI 3.045 to 5.354), corresponding to 70.7% and 303.8% higher odds, respectively. A unit increase in GDP per capita was significantly associated with a 10.4% decrease (adjusted rate ratios: 0.896, 95% CI 0.834 to 0.962) in the duration of shortage intensity.

CONCLUSION: Our study reveals global inequities in the impact of drug shortages, with countries with lower GDP per capita disproportionately affected. Persistent shortages of essential medications have been observed worldwide over the past decade, but are not evenly distributed across countries. Collectively, our findings suggest the need to consider creative and alternative policy strategies beyond pricing to address both critical drug shortages and global inequities in shortage experiences.

BACKGROUND: Hospitalized older adults are commonly discharged with changes to antihypertensive and glucose-lowering (cardiometabolic) medications. Though adverse drug events remain a leading cause of readmissions, there is little contemporary data on how medication discharge planning is communicated and how often medication errors occur post-discharge.

OBJECTIVE: To assess older adults' post-hospital medication use and ambulatory follow-up after receiving cardiometabolic medication changes during hospitalization.

DESIGN: Prospective cohort study from 11/2022 to 01/2024.

PARTICIPANTS: Adults aged 65 years or older from discharged home from an academic medical center with changes to pre-admission cardiometabolic medications.

MAIN MEASURES: Participants completed 7- and 90-day telephonic surveys on health status, medication use, and discharge planning. Self-report of medication use was compared to discharge summaries to identify medication errors (not initiating, not stopping, or taking incorrect dose). Multivariable regression models were used to identify characteristics associated with errors.

KEY RESULTS: The cohort included 151 participants (median [IQR] age 74 [70-78] years; 54% male; 17% Black, 82% White, 41% frail). Participants were admitted with a median (IQR) of 3 (2-4) cardiometabolic medications and discharged with a median (IQR) of 2 (1-4) medication changes. Of the 319 individual medications changed at discharge, 33% were further modified by 90 days. Participants reported comprehensive medication discharge planning for only 13% of medication changes. Though 93% of participants reported they understood the purpose of each of their medications at discharge, 39% had ≥ 1 medication errors at 7 days and 50% at 90 days. Use of ≥ 5 cardiometabolic medications was associated with higher rates of medication errors at 7 days (IRR 1.63; 95% CI 1.07-2.48) and 90 days (IRR 1.66; 95% CI 1.13-2.45).

CONCLUSIONS: Most hospitalized older adults discharged with cardiometabolic medication changes experienced medication errors or gaps in discharge planning. Steps to ensure all patients receive high-quality medication discharge planning are needed.

OBJECTIVES: It is unclear how long youth with opioid use disorder (OUD) should continue taking buprenorphine, and what adherence they should achieve. We identified patterns of duration/adherence and assessed associations with subsequent overdose, emergency department (ED) use, and hospitalization.

METHODS: This retrospective cohort analysis used 2014-2022 data from the Massachusetts Public Health Data Warehouse. We identified youth aged 13 to 26 years initiating buprenorphine and used group-based trajectory modeling to categorize youth into duration/adherence trajectories over 12 months. Using multivariable Cox regression, we examined associations between trajectories and time to fatal/nonfatal opioid overdose, all-cause ED use, and all-cause hospitalization during the subsequent 12-month period.

RESULTS: Among 11 649 Massachusetts youth initiating buprenorphine, most were aged 21 years or older (89.0%), male (60.3%), white non-Hispanic (85.9%), and enrolled in Medicaid (55.4%). We identified 4 patterns of medication use: (1) high adherence for 12 months (23.7%); (2) low adherence for 12 months (27.5%); (3) discontinuation in 3 to 9 months (16.4%); and (4) discontinuation in less than 3 months (32.5%). Trajectories included 580 (5.0%) and 774 (6.6%) youth switching to methadone and naltrexone, respectively. Compared with high adherence for 12 months, overdose risk was higher with low adherence for 12 months (adjusted hazard ratio [aHR], 1.46; 95% CI, 1.24-1.73), discontinuation in 3 to 9 months (aHR, 1.82; 95% CI, 1.52-2.17), and discontinuation in less than 3 months (aHR, 1.76; 95% CI 1.50-2.06). Compared with high adherence, low adherence and discontinuation in less than 3 months had higher risk of ED use, and all other trajectories had higher risk of hospitalization.

CONCLUSIONS: Medication adherence may prevent overdose, ED use, and hospitalization. Strategies to increase treatment duration/adherence likely avert harm.