Publications

Marketing payments from opioid manufacturers to outpatient physicians are associated with opioid prescribing. We examined the association between opioid-related payments to hospitalists and opioid prescribing at discharge from 2015 to 2023. We linked Medicare Part B Physicians & Other Practitioners public use files, Centers for Medicare & Medicaid Open Payments data, and Medicare Part D Prescribers Public Use Files to identify hospitalists' opioid-related payments and opioid prescribing. Of 67,408 hospitalists, 360 (0.53%) received payments, which declined from 581 in 2015 to 65 in 2023. Each payment was associated with a 0.94 percentage-point increase (95% confidence interval [CI], 0.60 to 1.28) in the proportion of outpatient prescription claims for opioids, equivalent to 112 additional opioid claims (95% CI, 70.78 to 154.73). Total payment value was not associated with opioid claims. Opioid-related payments to hospitalists are uncommon and declining, but even minimal exposure is associated with greater opioid prescribing.

BACKGROUND: Migraine attacks can be prevalent and debilitating, affecting 40 million adults in the United States. Patients can be prescribed prophylaxis (eg, β blockers, anticonvulsants, antidepressants) and rescue medications (eg, triptans, opioids [which are not recommended for use to treat migraine]) to treat and/or relieve migraine. Calcitonin gene-related peptide inhibitors (CGRPi) are the newest class of drugs approved for migraine prophylaxis. Patients taking CGRPi report reduced migraine frequency and severity, which should result in lower use of rescue migraine medications.

OBJECTIVE: To compare the use of triptans, rescue medications, and opioids for patients with migraines prescribed preventive CGRPi (pCGRi) and patients who are not prescribed pCGRPi.

METHODS: This retrospective cohort study used Komodo Health medical and pharmacy claims data from 2018 to 2023. Adults aged 18 to 60 years with a chronic migraine diagnosis and continuous enrollment for 12 months before and after index were included if they initiated a migraine medication (pCGRPi or non-CGRPi) during the index period. Outcomes included use of triptans, opioids, and other rescue medications in the post-index period compared with the pre-index period. Propensity score-matched cohorts were analyzed using adjusted logistic regression models with a treatment-by-time interaction term to estimate relative changes over time between cohorts.

RESULTS: In adjusted analyses, compared with the non-CGRPi cohort, the pCGRPi cohort experienced lower odds in triptan use (interaction odds ratio [OR] = 0.57; 95% CI = 0.53-0.61; P < 0.001) and in opioid use (interaction OR = 0.92; 95% CI = 0.85-0.99; P  =  0.033). Although rescue medication use increased over time, the relative pre-post change did not differ significantly between cohorts (interaction OR = 0.96; 95% CI = 0.89-1.04; P = 0.28). Sensitivity analyses restricted to neurologist-prescribed index medications showed consistent findings for triptans, but not for use of opioids or rescue medications.

CONCLUSIONS: In the cohort prescribed pCGRPi, there was a statistically significant association with decreased of triptan use over time. Comparatively, the non-CGRPi cohort had statistically significant increases in the use of triptans, opioids, and other rescue medications over time. These findings support that prophylactic use of pCGRPi may be associated with a reduced reliance on migraine-specific rescue medications.

BACKGROUND: The COVID-19 pandemic disrupted HIV prevention services, including reductions in pre-exposure prophylaxis (PrEP) prescription and HIV testing. However, evidence on the association between the pandemic with PrEP use, as well as HIV and sexually transmitted infection (STI) rates among PrEP users remained limited.

OBJECTIVE: To compare patterns of PrEP adherence, HIV incidence, and STI rates among adult men newly initiating PrEP before vs during the COVID-19 pandemic.

METHODS: We conducted a new-user retrospective cohort analysis of commercially insured PrEP users using the 2018-2021 MarketScan administrative claims database. Adult men aged 18-49 years in the MarketScan administrative claims database prescribed tenofovir disoproxil fumarate with emtricitabine or tenofovir alafenamide with emtricitabine for PrEP were included. Group-based trajectory modeling was used to identify PrEP adherence patterns. After balancing group demographic and clinical characteristics with inverse probability treatment weighting, time-varying Cox proportional hazards models were used to evaluate HIV risk and Poisson regression was used to assess STI rates.

RESULTS: Of 8,683 adult men (43% aged 25 to 34 years), 5,957 initiated PrEP before the pandemic and 2,726 during the pandemic. Prepandemic adherence trajectories included nonadherence (11.8% of the cohort) and rapidly declining (20.8%), gradually declining (21.1%), and consistently high (46.3%) adherence. Pandemic adherence trajectories included early discontinuation (35.7%) and rapidly declining (19.5%), gradually declining (14.8%), and consistently high (29.9%) adherence. Initiating PrEP during the pandemic was associated with higher risk of HIV acquisition (adjusted hazard ratio = 2.20 [95% CI = 1.08-4.47]). STI rates did not differ between groups (adjusted incidence rate ratio = 1.04 [95% CI = 0.88-1.23]).

CONCLUSIONS: Individuals initiating PrEP during the pandemic were associated with reduced PrEP adherence and higher HIV incidence. Targeted strategies to maintain access and continuity of HIV prevention services during public health emergencies are needed.

PURPOSE: Medications for opioid use disorder improve survival and health outcomes, yet access remains limited. Telemedicine bridge clinics offer rapid, low-barrier treatment initiation and retention. This study evaluated whether care through a telemedicine bridge clinic was associated with improved medication engagement, emergency department (ED) utilization, and Medicaid expenditures compared with usual office-based care.

METHODS: Pennsylvania Medicaid claims were linked with clinical data to identify adults aged 18-64 who received care from a telemedicine bridge clinic between April 2020 and July 2022. Bridge clinic patients were matched to up to four Medicaid enrollees initiating care with other opioid use disorder providers based on demographic and eligibility characteristics. Difference-in-differences models compared medication use, proportion of days covered (PDC) by medications for opioid use disorder, and ED utilization in the 6 months before and after the index visit. Medicaid expenditures were examined descriptively.

RESULTS: The sample included 552 bridge clinic patients and 2208 matched comparison patients. A greater proportion of bridge clinic patients filled a buprenorphine prescription in the month after qualifying encounter (91.3% vs. 73.6%). Relative to the comparison group, bridge clinic patients experienced a 6.9% increase in medication PDC and an 18.2% relative reduction in the proportion with any ED visit. Medicaid expenditures were lower among bridge clinic patients, driven primarily by lower inpatient spending.

CONCLUSIONS: Telemedicine bridge clinic care was associated with improved continuity of medication treatment and reduced ED utilization among Medicaid enrollees, supporting its role as an effective, low-barrier model for opioid use disorder care.

IMPORTANCE: Medicare Advantage (MA) plans use health care utilization management and other approaches that may be associated with different utilization patterns among beneficiaries from different racial and ethnic groups compared with traditional Medicare (TM).

OBJECTIVE: To measure the MA-TM utilization gap across racial and ethnic groups.

DESIGN, SETTING, AND PARTICIPANTS: This retrospective cohort study included Medicare beneficiaries 65 years or older continuously enrolled in MA and a 20% random sample of TM beneficiaries during 2021. TM beneficiaries were matched to MA beneficiaries by geographic location, age, and sex separately for Black, Hispanic, and White beneficiaries. Data analysis was conducted between August 2025 and January 2026.

EXPOSURE: MA vs TM enrollment.

MAIN OUTCOME AND MEASURES: Utilization rates per 1000 beneficiaries for 3 discretionary procedures (hip, knee replacement, and back surgery), 3 less discretionary cardiac procedures (cardiac catheterizations, percutaneous coronary interventions, and coronary artery bypass grafting), and 2 overall measures (emergency department visits and hospitalizations). MA utilization was measured using Healthcare Effectiveness Data and Information Set data; TM utilization measures were created following the exact same specifications.

RESULTS: The final matched samples included 4 366 556 MA beneficiaries (2 442 688 [56.0%] female; 1 254 260 [28.7%] aged 71-75 years; 2 84 774 [6.5%] Black, 226 798 [5.2%] Hispanic, and 3 854 984 [88.3%] White) and 4 366 556 TM beneficiaries (2 442 688 [56.0%] female; 1 254 260 [28.7%] aged 71-75 years; 284 774 [6.5%] Black, 226 798 [5.2%] Hispanic, and 3 854 984 [88.3%] White). The MA-TM gap was largest for White beneficiaries for discretionary procedures. For example, there were 2.51 fewer (95% CI, -2.52 to -2.49) knee replacements per 1000 beneficiaries for White beneficiaries in MA vs TM, 1.26 fewer (95% CI, -1.32 to -1.21) for Black compared with White beneficiaries (P < .001), and 0.60 fewer (95% CI, -0.66 to -0.55) for Hispanic compared with White beneficiaries (P < .001). Among Hispanic and White beneficiaries, rates of coronary artery bypass grafting were marginally higher in MA vs TM (1.84 [95% CI, 1.67 to 2.02] vs 1.41 [95% CI, 1.25 to 1.56] per 1000 beneficiaries [P < .001]; 2.22 [95% CI, 2.17 to 2.26] vs 2.12 [95% CI, 2.05 to 2.20] per 1000 beneficiaries [P = .004], respectively) but not significantly different among Black beneficiaries in MA vs TM (1.37 [95% CI, 1.23 to 1.52] vs 1.45 [95% CI, 0.69 to 2.20]; P = .44). Compared with White beneficiaries, the MA-TM gap was significantly different but of similar magnitude for Black (0.09 [95% CI, 0.08 to 0.10] vs -0.07 [95% CI, -0.15 to 0.00];P < .001) and Hispanic (0.09 [95% CI, 0.08 to 0.10] vs 0.44 [95% CI, -0.40 to 0.47]; P = .007) beneficiaries. Hospitalizations showed similar MA-TM gaps across race and ethnicity. Emergency department visit rates were lower in the MA group, with the greatest difference among Black beneficiaries.

CONCLUSIONS AND RELEVANCE: In this cohort study of MA and TM beneficiaries, exposure to MA was associated with a narrowing of utilization gaps for discretionary procedures with similar impacts on utilization across racial and ethnic groups for less discretionary procedures and overall utilization. Observed reductions in disparities in utilization possibly reflect reductions in discretionary use of procedures among White beneficiaries rather than increased access for Black and Hispanic beneficiaries.

OBJECTIVE: The COVID-19 pandemic exacerbated health disparities and increased alcohol use and its health consequences. Although prepandemic research identified racial and ethnic disparities in receipt of alcohol use disorder treatment, few studies have described postpandemic disparities or have assessed the role social determinants of health (SDOHs) may play in these inequities. The authors examined how rates of treatment for alcohol use disorder differ by race and ethnicity, SDOHs, and at the intersection of both.

METHODS: In this cross-sectional analysis, the authors used a nationally representative sample of 5,785 U.S. adults with past-year alcohol use disorder from the 2023 National Survey on Drug Use and Health (NSDUH). Multivariable logistic regression models were used to estimate the probability of any past-year alcohol use disorder treatment overall, by the intersection of race and ethnicity with 20 SDOH subcategories, and across treatment settings.

RESULTS: Only 9.6% of adults with alcohol use disorder in the NSDUH received any treatment in 2023. Compared with other racial and ethnic groups, non-Hispanic Black or African American adults had the lowest probability of receiving treatment for alcohol use disorder overall and in the stratified analyses of nine of 20 SDOH subcategories.

CONCLUSIONS: These findings highlight significant disparities in the receipt of treatment for alcohol use disorder among non-Hispanic Black or African American adults, generally concentrated within the more vulnerable SDOH categories. These results underscore a critical need for clinically and culturally nuanced policies and practices tailored to specific populations with health disparities.