IntroductionThe objective of this research is to provide national estimates of the prevalence of health condition diagnoses among age-entitled transgender and cisgender Medicare beneficiaries. Quantification of the health burden across sex assigned at birth and gender can inform prevention, research, and allocation of funding for modifiable risk factors.MethodsUsing 2009–2017 Medicare fee-for-service data, we implemented an algorithm that leverages diagnosis, procedure, and pharmacy claims to identify age-entitled transgender Medicare beneficiaries and stratify the sample by inferred gender: trans feminine and nonbinary (TFN), trans masculine and nonbinary (TMN), and unclassified. We selected a 5% random sample of cisgender individuals for comparison. We descriptively analyzed (means and frequencies) demographic characteristics (age, race/ethnicity, US census region, months of enrollment) and used chi-square and t-tests to determine between- (transgender vs. cisgender) and within-group gender differences (e.g., TMN, TFN, unclassified) difference in demographics (p<0.05). We then used logistic regression to estimate and examine within- and between-group gender differences in the predicted probability of 25 health conditions, controlling for age, race/ethnicity, enrollment length, and census region.ResultsThe analytic sample included 9,975 transgender (TFN n=4,198; TMN n=2,762; unclassified n=3,015) and 2,961,636 cisgender (male n=1,294,690, female n=1,666,946) beneficiaries. The majority of the transgender and cisgender samples were between the ages of 65 and 69 and White, non-Hispanic. The largest proportion of transgender and cisgender beneficiaries were from the South. On average, transgender individuals had more months of enrollment than cisgender individuals. In adjusted models, aging TFN or TMN Medicare beneficiaries had the highest probability of each of the 25 health diagnoses studied relative to cisgender males or females. TFN beneficiaries had the highest burden of health diagnoses relative to all other groups.DiscussionThese findings document disparities in key health condition diagnoses among transgender Medicare beneficiaries relative to cisgender individuals. Future application of these methods will enable the study of rare and anatomy-specific conditions among hard-to-reach aging transgender populations and inform interventions and policies to address documented disparities.

Background The Centers for Medicare & Medicaid Services implemented the National Partnership to Improve Dementia Care in Nursing Homes (the Partnership) to decrease antipsychotic use and improve care for nursing home (NH) residents with dementia. We determined whether the extent of antipsychotic and other psychotropic medication prescribing in AL residents with dementia mirrored that of long‐stay NH (LSNH) residents after the Partnership. Methods Using a 20% sample of fee‐for‐service Medicare beneficiaries with Part D, we conducted a retrospective cohort study including AL and LSNH residents with dementia. The monthly prevalence of psychotropic medication prescribing (antipsychotics, antidepressants, anxiolytics/sedative‐hypnotics, anticonvulsants/mood stabilizers, benzodiazepines, and antidementia medications) was examined. We used an interrupted time‐series analysis to compare medication prescribing before (July 1, 2010–March 31, 2012) and after (April 1, 2012–December 31, 2017) the Partnership in both settings. Results We identified 107,931 beneficiaries with ≥1 month as an AL resident and 323,766 beneficiaries with ≥1 month as a LSNH resident with dementia, including 1,923,867 person‐months and 4,984,405 person‐months, respectively. Antipsychotic prescribing declined over the study period in both settings. After the launch of the Partnership, the rate of decline in antipsychotic prescribing slowed in AL residents with dementia (slope change = 0.03 [95% CLs: 0.02, 0.04]) while the rate of decline in antipsychotic prescribing increased in LSNH residents with dementia (slope change = −0.12 [95% CLs: −0.16, −0.08]). Antidepressants were the most prevalent medication prescribed, anticonvulsant/mood stabilizer prescribing increased, and anxiolytic/sedative‐hypnotic and antidementia medication prescribing declined. Conclusions The federal Partnership to reduce antipsychotic prescribing in NH residents did not appear to affect antipsychotic prescribing in AL residents with dementia. Given the increase in the prescribing of mood stabilizers/anticonvulsants that occurred after the launch of the Partnership, monitoring may be warranted for all psychotropic medications in AL and NH settings.

Research Objective Americans who belong to racial/ethnic minorities, particularly those who are Black and/or Hispanic, face substantial inequities in access to care. To better understand the drivers of these inequities, it's important to quantify how care patterns differ between groups. There is limited research that describes patterns of post‐acute care (PAC) use of racial/ethnic minorities following hospitalizations. We used a national sample of hospitalized Medicare enrollees to compare first discharge location across racial/ethnic minority groups. Study Design We conducted a cohort study of Medicare enrollees admitted to the hospital for five conditions that often require post‐acute care (PAC): congestive heart failure (CHF), stroke, hip fracture, lower extremity joint replacement (LEJR), and chronic obstructive pulmonary disorder (COPD). We identified all traditional Medicare beneficiaries hospitalized for these conditions from 2007 through 2017 using the Medicare Provider Analysis and Review (MedPAR) file. We then used MDS, OASIS, and IRF‐PAI assessments and claims to identify which PAC locations patients were discharged to. Our multinomial outcome of interest was discharge to home without home health, home with home health (HH), skilled nursing facility (SNF), inpatient rehabilitation (IRF), other locations (long term care hospitals or other acute care), or death during hospitalization. Our primary explanatory variable was patient race/ethnicity. We first compared discharge locations by disease cohort and race ethnicity. We then used multinomial logit models to adjust for age, gender, Medicare disability entitlement, dual Medicare‐Medicaid eligibility, and zipcode and hospital fixed effects. We purposefully did not include chronic disease indicators as chronic disease coding may be endogenous with race/ethnicity, however in sensitivity analyses that included chronic disease flags, the results did not differ substantially. From these models, we calculated the adjusted percent discharged to each location and compared by race/ethnicity. Population Studied 2,782,117 Medicare beneficiaries who were admitted to a hospital for our target conditions from 2007 through 2017. Principal Findings The study included 313,735 patients admitted for CHF, 387,792 for stroke, 289,027 for hip fracture, 591,677 for LEJR and 1,199,886 for COPD. After adjustment, among patients with COPD, 26.9% of white patients were discharged to home without PAC, and 35.8% were discharged to SNF compared to 34.6% and 28.7% of Black and 33.9% and 27.5% of Hispanic patients. Among patients with CHF, 50.1% of white patients were discharged to home without PAC, and 19% were discharged to SNF compared to 57.1% and 11.4% of Asian and 55.3% and 11.9% of Hispanic patients. Among patients with stroke, 43.4% of white patients were discharged to home without PAC, and 19% were discharged to SNF compared to 31.3% and 19.0% of Black patients. Discharge patterns were similar for hip fracture and lower extremity joint replacement. Conclusions Even after accoun...