Real-World Survival, Healthcare Resource Utilization, and Costs Among U.S. Elderly Patients With Diffuse Large B-Cell Lymphoma (DLBCL) Treated With R-GemOx in the Relapsed/Refractory Setting.

BACKGROUND: Little recent real-world evidence exists on overall survival, healthcare resource utilization (HCRU), and costs among R/R DLBCL patients treated with the combination of rituximab, gemcitabine, and oxaliplatin (R-GemOx), a widely-used regimen for patients ineligible for stem cell transplant due to age or comorbidities. PATIENTS AND METHODS: This retrospective analysis used 2014 to 2019 U.S. Medicare claims. Individuals aged ≥66 years with a new DLBCL diagnosis between October 1, 2015 and December 31, 2018 and continuous fee-for-service Medicare Part A, B, and D coverage in the 12 months pre- and postindex were followed to identify the sample of patients with evidence of R-GemOx treatment in the second-line (2L) or third-line (3L) setting. Outcomes included overall survival, all-cause and DLBCL-related HCRU, and costs after R-GemOx initiation. RESULTS: The final sample included 157 patients who received treatment with R-GemOx in the R/R settings (mean (SD) age 77.5 (6.0) years, 39.5% age>80 years; 66.9% male; 91.1% White). Of these, 126 received R-GemOx in the 2L setting and 31 received R-GemOx in the 3L setting. Median overall survival from R-GemOx initiation was 6.9 months and 6.8 months in the 2L and 3L setting, respectively. Rates of all-cause hospitalization (68.1% [2L] and >90% [3L]) and hospice use (42.9% [2L] and 51.7% [3L]) were high in the 12 months after R-GemOx initiation. All-cause total costs were substantial ($144,653 [2L] and $142,812 [3L]) and approximately 80% of costs were DLBCL-related within 12 months of R-GemOx initiation. CONCLUSION: Elderly U.S. Medicare beneficiaries diagnosed with DLBCL who initiated R-GemOx treatment in the R/R setting have poor overall survival, high rates of HCRU, and substantial costs.

Health care resource utilization and costs associated with diagnosed medication overuse headache and potential acute medication overuse in individuals with migraine.

OBJECTIVE: Estimate health care resource utilization and costs associated with medication overuse headache and potential acute medication overuse. METHODS: A retrospective analysis was conducted with Clinformatics Data Mart data (1 January 2019-31 December 2019) that included continuously enrolled commercially insured adults with migraine (International Classification of Diseases, Tenth Revision, Clinical Modification [ICD-10-CM] code G43.xxx). Medication overuse headache was defined as ≥1 inpatient or ≥2 outpatient claims with an ICD-10-CM code G44.41/40 (drug-induced headache). Potential acute medication overuse was defined as possessing sufficient medication for >10 mean treatment days/month for ergots, triptans, opioids, or combination analgesics or >15 mean cumulative days/month for simple prescription analgesics (e.g., acetaminophen, aspirin, other non-opioid analgesics) for >6 consecutive months. All-cause and migraine-related health care resource utilization and costs were compared after adjusting for demographic and clinical characteristics. RESULTS: Among 90,017 individuals with migraine, the frequency of medication overuse headache/potential acute medication overuse was 12.6% (diagnosed medication overuse headache: 0.6%; potential acute medication overuse: 12.1%). Adjusted all-cause total costs ($31,235 vs $21,486; difference: $9,749 [P < 0.001]) and adjusted migraine-related total costs ($9,770 vs $6,207; difference: $3,563 [P < 0.001]) were higher in the medication overuse headache/potential acute medication overuse group versus those without medication overuse headache/potential acute medication overuse. CONCLUSIONS: Individuals with diagnosed medication overuse headache/potential acute medication overuse had higher all-cause and migraine-related health care resource utilization and costs versus individuals without medication overuse headache/potential acute medication overuse, suggesting that improved migraine management is needed to reduce associated costs.

Real-world analysis of adverse event rates after initiation of ibrutinib among Medicare beneficiaries with chronic lymphocytic leukemia.

BACKGROUND: The first-generation BTK inhibitor ibrutinib is a standard-of-care therapy in the treatment of chronic lymphocytic leukemia (CLL) despite potential side effects that often lead to discontinuation. METHODS: This study used 2013-2019 claims data to describe the incidence rate of adverse events (AEs) among elderly Medicare beneficiaries newly initiating ibrutinib for CLL. RESULTS: The final sample contained 11,870 Medicare beneficiaries with CLL (mean age 77.2) newly initiating ibrutinib, of whom 65.2% discontinued over mean follow-up of 2.3 years. The overall incidence rate of AEs was 62.5 per 1000 patient-months for all discontinuers and 32.9 per 1000 patient-months for non-discontinuers. Discontinuers had a higher incidence rate of AEs per 1000 patient-months compared with non-discontinuers for all AEs examined, including infection (22.8 vs. 14.5), atrial fibrillation (15.1 vs. 7.0), anemia (21.9 vs. 14.5), and arthralgia/myalgia (19.5 vs. 13.6). CONCLUSION: In this first real-world study of a national sample of elderly US patients treated with ibrutinib, we found a clear unmet need for improved management of ibrutinib-related AEs and/or new treatments to improve real-world outcomes in patients with CLL.

Estimating The Impact Of Out-Of-Pocket Cost Changes On Abandonment Of HIV Pre-Exposure Prophylaxis.

Oral HIV pre-exposure prophylaxis (PrEP) is highly effective for preventing HIV. Several different developments in the US either threaten to increase or promise to decrease PrEP out-of-pocket costs and access in the coming years. In a sample of 58,529 people with a new insurer-approved PrEP prescription, we estimated risk-adjusted percentages of patients who abandoned (did not fill) their initial prescription across six out-of-pocket cost categories. We then simulated the percentage of patients who would abandon PrEP under hypothetical changes to out-of-pocket costs, ranging from $0 to more than $500. PrEP abandonment rates of 5.5 percent at $0 rose to 42.6 percent at more than $500; even a small increase from $0 to $10 doubled the rate of abandonment. Conversely, abandonment rates that were 48.0 percent with out-of-pocket costs of more than $500 dropped to 7.3 percent when those costs were cut to $0. HIV diagnoses were two to three times higher among patients who abandoned PrEP prescriptions than among those who filled them. These results imply that recent legal challenges to the provision of PrEP with no cost sharing could substantially increase PrEP abandonment and HIV rates, upending progress on the HIV/AIDS epidemic.

Geographic variations of pre-exposure prophylaxis reversal and abandonment among United States counties.

OBJECTIVE: In the United States, one in five newly insurer-approved pre-exposure prophylaxis (PrEP) prescriptions are reversed with over 70% of those reversed, being abandoned. Given the Ending the HIV Epidemic (EHE) initiative's goals, we assessed geographic variations of PrEP reversal and abandonment across EHE and non-EHE counties in the United States. DESIGN: This was a cross-sectional analysis of secondary data. METHODS: Data were collected from Symphony Analytics for adults 18 years and older, with a newly prescribed PrEP claim. Using the proportion of PrEP prescriptions by county, hotspot analysis was conducted utilizing Getis Ord Gi∗ statistics stratified by EHE and non EHE counties. Multivariable logistic regression was used to identify factors associated with residing in hotspots of PrEP reversal or PrEP abandonments. RESULTS: Across 516 counties representing 36,204 patients, the overall PrEP reversal rate was 19.4%, whereas the PrEP abandonment rate was 13.7%. Reversals and abandonments were higher for non-EHE (22.7 and 17.1%) than EHE (15.6 and 10.5%) counties. In both EHE and non-EHE counties, younger age, less education, females, and an out-of-pocket cost of greater than $100, were significantly associated with greater likelihood of residing in hotspots of PrEP reversal or abandonment, while Hispanics, Medicaid recipients, and an out-of-pocket cost of $10 or less had lower likelihood of residing in hotspots of reversal and abandonment. CONCLUSION: Findings indicate the need for implementation of focused interventions to address disparities observed in PrEP reversal and abandonment. Moreover, to improve primary PrEP adherence, national PrEP access programs should streamline and improve PrEP accessibility across different geographic jurisdictions.

Association of race and ethnicity with opioid prescribing for Medicare beneficiaries following total joint replacements.

BACKGROUND: Profound racial and ethnic disparities exist in the use and outcomes of total hip/knee replacements (total joint replacements [TJR]). Whether similar disparities extend to post-TJR pain management remains unknown. Our objective is to examine the association of race and ethnicity with opioid fills following elective TJRs for White, Black, and Hispanic Medicare beneficiaries. METHODS: We used the 2019 national Medicare data to identify beneficiaries who underwent total hip/knee replacements. Primary outcomes were at least one opioid fill in the period from discharge to 30 days post-discharge, and 31-90 days following discharge. Secondary outcomes were morphine milligram equivalent per day and number of opioid fills. Key independent variable was patient race-ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic). We estimated multivariable hierarchical logistic regressions and two-part models with state-level clustering. RESULTS: Among 67,550 patients, 93.36% were White, 3.69% were Black, and 2.95% were Hispanic. Compared to White patients, more Black patients and fewer Hispanic patients filled an opioid script (84.10% [Black] and 80.11% [Hispanic] vs. 80.33% [White], p < 0.001) in the 30-day period. On multivariable analysis, Black patients had 18% higher odds of filling an opioid script in the 30-day period (odds ratio [OR]: 1.18, 95% confidence interval [CI]: 1.05-1.33, p = 0.004), and 39% higher odds in the 31-90-day period (OR: 1.39, 95% CI: 1.26-1.54, p < 0.001). There were no significant differences in the endpoints between Hispanic and White patients in the 30-day period. However, Hispanic patients had 20% higher odds of filling an opioid script in the 31- to 90-day period (OR: 1.20, 95% CI: 1.07-1.34, p = 0.002). CONCLUSIONS: Important race- and ethnicity-based differences exist in post-TJR pain management with opioids. The mechanisms leading to the higher use of opioids by racial/ethnic minority patients need to be carefully examined.

Antiretroviral Treatment Gaps and Adherence Among People with HIV in the U.S. Medicare Program.

Approximately one-quarter of people with HIV (PWH) in the U.S. receive coverage through the Medicare program; however, no prior real-world study has examined antiretroviral therapy (ART) gaps and adherence and associated factors in this population. This retrospective cohort analysis used 2013-2018 national Medicare fee-for-service claims data to identify all PWH initiated on a new ART regimen including protease inhibitors [PI], non-nucleoside reverse transcriptase inhibitors [NNRTIs], or integrase strand transfer inhibitors [INSTIs] between 1/1/2014 and 12/31/2017. Study outcomes included ART adherence (based on proportion of days covered [PDC]), continuous treatment gaps ranging from 1 to 6 days to ≥ 180 days, and discontinuation (continuous gap ≥ 90 days) in the 12-month follow-up period. Multivariable regressions were used to assess factors associated with ART adherence and discontinuation. The final sample included 48,627 PWH (mean age: 54.5 years, 74.4% male, 47.5% White, 89.8% disabled). Approximately 53.0% of PWH had a PDC ≥ 0.95, 30.2% had a PDC between 0.70 and < 0.95, and 16.8% had PDC < 0.70. Treatment gaps of at least ≥ 7-days (55.2%) and ≥ 30-days (26.2%) were common and 10.1% PWH discontinued treatment. Younger age, female sex, Black race, higher comorbidity score, mental health conditions, and substance use disorder were associated with higher odds of lower adherence and discontinuation (all p-values < 0.05). In conclusion, suboptimal adherence and treatment gaps in ART use were commonly observed among PWH in Medicare. Interventions and policies to mitigate barriers to adherence are urgently needed in this population to both improve their survival and increase the potential for ending the HIV epidemic in the US.

Real-world treatment patterns, survival, health resource use and costs among Medicare beneficiaries with diffuse large B-cell lymphoma.

Aim: To examine real-world treatment patterns, survival, healthcare resource use and costs in elderly Medicare beneficiaries with diffuse large B-cell lymphoma (DLBCL). Methods: 11,880 Medicare patients aged ≥66 years with DLBCL between 1 October 2015 and 31 December 2018 were followed for ≥12 months after initiating front-line treatment. Results: Two-thirds (61.2%) of the patients received standard-of-care R-CHOP as first-line treatment. Hospitalization was common (57%) in the 12-months after initiation of 1L treatment; the mean DLCBL-related total costs were US$84,416 during the same period. Over a median follow-up of 2.1 years, 17.8% received at least 2L treatment. Overall survival was lower among later lines of treatment (median overall survival from initiation of 1L: not reached; 2L: 19.9 months; 3L: 9.8 months; 4L: 5.5 months). Conclusion: A large unmet need exists for more efficacious and well-tolerated therapies for older adults with DLBCL.

Diffuse large B-cell lymphoma (DLBCL) is the most common form of Non-Hodgkin lymphoma, and it becomes more common with age. While researchers continue to develop newer, more effective treatments for DLBCL, it is important to understand how patients use existing treatments and the associated costs, particularly among the elderly. In our real-world analysis of nearly 12,000 older patients with DLBCL, we found high rates of hospitalization and hospice use, short length of life in later lines of therapy and substantial healthcare costs. Our findings suggest a large current unmet need for more effective and well-tolerated therapies for older adults with DLBCL in both the front-line and relapse/refractory settings.

Out-of-Pocket Costs for Long-Acting Injectable and Oral Antipsychotics Among Medicare Patients With Schizophrenia.

OBJECTIVE: The authors sought to describe out-of-pocket (OOP) costs among beneficiaries with schizophrenia differing in Medicare Part D low-income subsidy (LIS) status. METHODS: National 100% Medicare claims were used to identify all adult fee-for-service Medicare Part D beneficiaries with schizophrenia who used antipsychotics in 2019 (N=283,813). Proportions of patients by LIS status, OOP costs per prescription, and annual OOP costs were reported. Results were stratified by type of antipsychotic received (oral antipsychotic [OAP], first-generation long-acting injectable [FGA-LAI], or second-generation long-acting injectable [SGA-LAI]). RESULTS: In the final sample, 90.3% of beneficiaries had full LIS status, paying minimal copayments (29.6% institutionalized full LIS, paying $0; 42.2% noninstitutionalized full LIS, ≤100% federal poverty level [FPL], paying $1.25-$3.80; and 18.5% noninstitutionalized full LIS, >100% FPL, paying $3.40-$8.50). Only 0.9% of the sample received partial LIS status, and 8.8% had a non-LIS status. Non-LIS beneficiaries had the highest OOP costs, followed by partial LIS beneficiaries. Before entering catastrophic coverage, median OOP costs per prescription for generic OAPs, brand-name OAPs, FGA-LAIs, and SGA-LAIs were $10.85, $171.97, $26.09, and $394.28, respectively, for non-LIS beneficiaries and $3.69, $105.82, $9.35, and $229.20, respectively, for partial LIS beneficiaries. The annual total OOP costs varied substantially by LIS status (full LIS, $0-$130.79; partial LIS, $458.96; non-LIS, $998.81). CONCLUSIONS: Most Medicare beneficiaries with schizophrenia qualified for full LIS and faced minimal OOP costs for both OAPs and LAIs. The remainder (i.e., partial LIS and non-LIS beneficiaries) faced substantial OOP costs, both per prescription and annually, especially for SGA-LAIs.

Ibrutinib discontinuation and associated factors in a real-world national sample of elderly Medicare beneficiaries with chronic lymphocytic leukemia.

Prior studies evaluating ibrutinib discontinuation are limited to clinical trials and selected medical centers and hence may not reflect real-world practice. This study used Medicare claims (2013-2019) to examine ibrutinib discontinuation and associated factors among elderly patients with chronic lymphocytic leukemia (CLL)/small lymphocytic lymphoma (SLL). Over a median follow-up of 2.1 years, two-thirds (65.2%) of the 11,870 new ibrutinib initiators were discontinued, with half (45.1%) of patients discontinuing within 12 months of initiation. Factors such as advanced age, lack of Part D low-income subsidy, evidence of prior CLL/SLL treatment, and cardiovascular comorbidities (e.g. atrial fibrillation) were associated with higher risk of discontinuation. Over a median of 1.2 years from discontinuation, 40% of discontinuers initiated another CLL/SLL treatment after ibrutinib discontinuation; 25% of patients restarted ibrutinib treatment at some point over follow-up. Our findings point to a large unmet need with the widely used BTKi ibrutinib and underscore the importance of ongoing development of efficacious and well-tolerated CLL/SLL therapies.

Racial and Ethnic Disparities in Long-Acting Injectable Antipsychotic Use in a National Sample of Medicare Beneficiaries With Schizophrenia.

This cross-sectional study assesses the racial and ethnic disparities in long-acting injectable antipsychotic use in a national sample of Medicare beneficiaries with schizophrenia.

Real-world adherence and discontinuation among Medicare beneficiaries initiating venetoclax vs. BTKis in relapsed/refractory chronic lymphocytic leukemia.

The treatment landscape for chronic lymphocytic leukemia (CLL) has been transformed by the availability of Bruton's tyrosine kinase inhibitors (BTKis) and the B-cell lymphoma 2 (BCL-2) inhibitor venetoclax. Despite clinical trial data supporting these novel oral agents, evidence evaluating real-world adherence is limited. This study used 2015-2019 Medicare claims data for elderly patients with relapsed/refractory CLL to assess differences in real-world adherence and discontinuation in the 12 months after treatment initiation. In the final sample of 711 venetoclax patients and 1,566 BTKi patients, we found that those initiating venetoclax tended to be younger (mean age 75.6 [SD 6.0] vs 77.6 [SD 6.9] years, p < .001) but had poorer clinical characteristics. After risk-adjustment, the venetoclax group had higher adherence (61.9% vs. 45.4%, p < .0001) and lower discontinuation when compared to the BTKi group (28.5% vs. 47.4%, p < .001). These favorable real-world findings underscore the importance of developing well-tolerated novel combinations for older adults.

Assessing unmet need among elderly Medicare Beneficiaries with Mantle cell lymphoma: an analysis of treatment patterns, survival, healthcare resource utilization, and costs.

Studies evaluating real-world outcomes and health care utilization for mantle cell lymphoma are limited. We utilized national Medicare claims (2009-2019) to examine treatment patterns, healthcare resource utilization, costs, and survival in 3664 elderly patients receiving 1 L treatment for MCL. Over a median follow-up of 2.8 years, 40.3% received at least 2 L treatment. The most common 1 L regimen was bendamustine-rituximab (50.1%), with increased use of BTKi-based regimens observed in 2 L (39.4%). Half (51.8%) of patients had an all-cause hospitalization within 12 months of initiating 1 L; hospitalization rates were higher in later lines. Healthcare costs were substantial and most costs (>80%) were MCL-related. Overall survival was poorer among later lines of treatment (median OS from initiation of 1 L: 53.5 months; 2 L: 22.0 months; 3 L: 11.8 months; 4 L: 7.8 months). These results suggest a large unmet need and future work should evaluate whether novel therapies have improved outcomes among elderly patients with MCL.

Antipsychotic utilization, healthcare resource use and costs, and quality of care among fee-for-service Medicare beneficiaries with schizophrenia in the United States.

BACKGROUND: No research to date has examined antipsychotic (AP) use, healthcare resource use (HRU), costs, and quality of care among those with schizophrenia in the Medicare program despite it serving as the primary payer for half of individuals with schizophrenia in the US. OBJECTIVES: To provide national estimates and assess regional variation in AP treatment utilization, HRU, costs, and quality measures among Medicare beneficiaries with schizophrenia. METHODS: Cross-sectional descriptive analysis of 100% Medicare claims data from 2019. The sample included all adult Medicare beneficiaries with continuous fee-for-service coverage and ≥1 inpatient and/or ≥2 outpatient claims with a diagnosis for schizophrenia in 2019. Summary statistics on AP use; HRU and cost; and quality measures were reported at the national, state, and county levels. Regional variation was measured using the coefficient of variation (CoV). RESULTS: We identified 314,888 beneficiaries with schizophrenia. About 91% used any AP; 20% used any long-acting injectable antipsychotic (LAI); and 14% used atypical LAIs. About 28% of beneficiaries had ≥1 hospitalization and 47% had ≥1 emergency room (ER) visits, the vast majority of which were related to mental health (MH). Total annual all-cause, MH, and schizophrenia-related costs were $23,662, $15,000 and $12,109, respectively. Among those with hospitalizations, 18.4% and 27.3% had readmission within 7 and 30 days and 56% and 67% had a physician visit and AP fill within 30 days post-discharge, respectively. Overall, 81% of beneficiaries were deemed adherent to their AP medications. Larger interstate variations were observed in LAI use than AP use (CoV: 0.21 vs 0.02). County-level variations were larger than state-level variations for all measures. CONCLUSIONS: In this first study examining a national sample of Medicare beneficiaries with schizophrenia, we found low utilization rates of LAIs and high levels of hospital admissions/readmissions and ER visits. State and county-level variations were also found in these measures.

SARS-CoV-2 Sequelae and Postdischarge Health Care Visits Over 5 Months Follow-up Among Children Hospitalized for COVID-19 or MIS-C.

Although post-acute sequelae of COVID-19 among adult survivors has gained significant attention, data in children hospitalized for severe acute respiratory syndrome coronavirus 2 is limited. This study of commercially insured US children shows that those hospitalized with COVID-19 or multisystem inflammatory syndrome in children have a substantial burden of severe acute respiratory syndrome coronavirus 2 sequelae and associated health care visits postdischarge.

Impact of Updated Clinical Practice Guidelines on Outpatient Treatment for Clostridioides difficile Infection and Associated Clinical Outcomes.

Background: The 2017 Infectious Diseases Society of America/Society for Healthcare Epidemiology of America (IDSA/SHEA) Clostridium (Clostridioides) difficile infection (CDI) guideline update recommended treatment with fidaxomicin or vancomycin for CDI. We aimed to examine outpatient CDI treatment utilization before and after the guideline update and compare clinical outcomes associated with fidaxomicin versus vancomycin use. Methods: A pre-post study design was employed using Medicare data. CDI treatment utilization and clinical outcomes (4- and 8-week sustained response, CDI recurrence) were compared between patients indexed from April-September 2017 (preguideline period) and those indexed from April-September 2018 (postguideline period). Clinical outcomes associated with fidaxomicin versus vancomycin were compared using propensity score-matched analyses. Results: From the pre- to postguideline period, metronidazole use decreased (initial CDI: 81.2% to 53.5%; recurrent CDI: 49.7% to 27.6%) while vancomycin (initial CDI: 17.9% to 44.9%; recurrent CDI: 48.1% to 66.4%) and fidaxomicin (initial CDI: 0.87% to 1.63%; recurrent CDI: 2.2% to 6.0%) use increased significantly (P < .001 for all). However, clinical outcomes did not improve. In propensity score-matched analyses, fidaxomicin versus vancomycin users had 4-week sustained response rates that were higher by 13.5% (95% confidence interval [CI], 4.0%-22.9%; P = .0058) and 30.0% (95% CI, 16.8%-44.3%; P = .0002) in initial and recurrent CDI cohorts, respectively. Recurrence rates were numerically lower for fidaxomicin in both cohorts. Conclusions: Vancomycin use increased and metronidazole use decreased after the 2017 guideline update. Fidaxomicin use increased but remained low. Improved outcomes associated with fidaxomicin relative to vancomycin suggest benefits from its greater use in Medicare patients.

Assessment of US Pharmacies Contracted With Health Care Institutions Under the 340B Drug Pricing Program by Neighborhood Socioeconomic Characteristics.

This cross-sectional study assesses pharmacy participation in the 340B Drug Pricing Program following the 2010 expansion and the extent to which growth has occurred in socioeconomically disadvantaged neighborhoods.

Financial incentives tied to Medicare star ratings: impact on influenza vaccination uptake in Medicare beneficiaries.

OBJECTIVES: To evaluate the impact of the star rating bonus payment policy on annual influenza vaccination rates before and after the policy was adopted for Medicare Advantage (MA) plans in 2012. STUDY DESIGN: Observational study using data from the Medicare Current Beneficiary Survey from 2007 to 2015 to test whether the bonus payment policy led to higher flu vaccination rates in MA prescription drug (MAPD) plans vs fee-for-service prescription drug plans (PDPs), which were ineligible for bonus payments. METHODS: Mean preperiod (2007-2011) and postperiod (2012-2015) influenza vaccination rates were compared for enrollees in both types of plans using descriptive and multivariate difference-in-difference (DID) equations. The experimental effect of the MA bonus payment policy was estimated as the interaction between plan type (MAPD plan vs PDP) and period (pre- vs post period) controlling for the main effects of plan type (MAPD vs PDP), timing of the observation (pre- vs post period), and other potential confounders. RESULTS: The study sample included 40,369 person-years of data in the preperiod and 27,703 person-years of data in the post period. Vaccination rates increased by 3.8% in MAPD plans compared with 2.7% in PDPs, leading to a relative MAPD-favored difference that was nonsignificant (P = .31). However, the effect was statistically significant (odds ratio [OR], 1.12; P = .03) in the main multivariate DID model. A larger relative difference was observed among beneficiaries 75 years and older (OR, 1.18; P = .03). CONCLUSIONS: The Medicare bonus payment policy led to a small increase in beneficiaries' flu vaccination rates, suggesting that expanding the star measure set could be an effective way to increase uptake for other recommended adult vaccines.

dPQL: a lossless distributed algorithm for generalized linear mixed model with application to privacy-preserving hospital profiling.

OBJECTIVE: To develop a lossless distributed algorithm for generalized linear mixed model (GLMM) with application to privacy-preserving hospital profiling. MATERIALS AND METHODS: The GLMM is often fitted to implement hospital profiling, using clinical or administrative claims data. Due to individual patient data (IPD) privacy regulations and the computational complexity of GLMM, a distributed algorithm for hospital profiling is needed. We develop a novel distributed penalized quasi-likelihood (dPQL) algorithm to fit GLMM when only aggregated data, rather than IPD, can be shared across hospitals. We also show that the standardized mortality rates, which are often reported as the results of hospital profiling, can also be calculated distributively without sharing IPD. We demonstrate the applicability of the proposed dPQL algorithm by ranking 929 hospitals for coronavirus disease 2019 (COVID-19) mortality or referral to hospice that have been previously studied. RESULTS: The proposed dPQL algorithm is mathematically proven to be lossless, that is, it obtains identical results as if IPD were pooled from all hospitals. In the example of hospital profiling regarding COVID-19 mortality, the dPQL algorithm reached convergence with only 5 iterations, and the estimation of fixed effects, random effects, and mortality rates were identical to that of the PQL from pooled data. CONCLUSION: The dPQL algorithm is lossless, privacy-preserving and fast-converging for fitting GLMM. It provides an extremely suitable and convenient distributed approach for hospital profiling.

DLMM as a lossless one-shot algorithm for collaborative multi-site distributed linear mixed models.

Linear mixed models are commonly used in healthcare-based association analyses for analyzing multi-site data with heterogeneous site-specific random effects. Due to regulations for protecting patients' privacy, sensitive individual patient data (IPD) typically cannot be shared across sites. We propose an algorithm for fitting distributed linear mixed models (DLMMs) without sharing IPD across sites. This algorithm achieves results identical to those achieved using pooled IPD from multiple sites (i.e., the same effect size and standard error estimates), hence demonstrating the lossless property. The algorithm requires each site to contribute minimal aggregated data in only one round of communication. We demonstrate the lossless property of the proposed DLMM algorithm by investigating the associations between demographic and clinical characteristics and length of hospital stay in COVID-19 patients using administrative claims from the UnitedHealth Group Clinical Discovery Database. We extend this association study by incorporating 120,609 COVID-19 patients from 11 collaborative data sources worldwide.