David C Grabowski’s research while affiliated with Harvard Medical School and other places

This study aims to fill the knowledge gap on the effect of psychotherapy on the quality of care of long-stay nursing home (NH) residents with mental illness or Alzheimer’s Disease and Related Dementias (ADRD). The 2017-2018 MDS and Medicare claims (e.g., Carrier file) and Residential History File were used to identify mental illness and ADRD diagnoses, patient characteristics, psychotherapy and quality of care. Fee-for-service Medicare beneficiaries aged 65 and older enrolled in Medicare Parts B who had six quarterly MDS assessments with/ without psychotherapy in the last four quarters were compared on two subgroups of residents: those with mental illness but no ADRD and those with ADRD. Five quarterly outcomes were examined: any 1) physical behavior; 2) verbal behavior; 3) other behavior; 4) hospitalization; 5) emergency-room visits. Covariates included predisposing, enabling characteristics and need factors. Propensity score matching was used to balance covariates between treatment and control groups. Difference-in-difference models of quarterly data, clustered within patients, were estimated for each outcome among each cohort. Analyses included 525,114/180,000 resident-quarter observations from 87,519/30,000 unique NH long-stay residents with ADRD/mental-illness-only. Among the ADRD/mental-illness-only group, 579/469 unique residents received psychotherapy. Psychotherapy was associated with lower likelihood of hospitalization (Odds ratio: 0.73, 95%CI: 0.53, 0.99) in residents with mental-illness-only. No significant results were found for the other outcomes nor the ADRD group. Findings suggest that psychotherapy may have limited effects on the quality of care for residents with ADRD but may contribute to lower hospitalization risk among residents with mental-illness-only.

Researchers have often used the nursing home Minimum Data Set (MDS) data to measure the rate of medication use, but little evidence exists on the accuracy of the MDS-based medication items. We compared quarterly rates of antipsychotic, antidepressant, and hypnotic use between the MDS and Part D Event file (PDE) in 2018 among long-stay nursing home residents enrolled in Medicare Part D who were ≥ 65 years old with psychiatric disorders or dementia identified in the MDS data. We used Cohen kappa to assess agreement in the share of residents using mental health medications during the quarter and used PDE data as a reference to calculate validity parameters. A total of 34.0%/73.4% of study participants were identified using antipsychotic/antidepressant medications in MDS assessments vs 34.3%/71.3% identified in PDE records (kappa value: 0.90/0.83). The sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of MDS data compared to PDE were 96.7%/84.5%, 92.8% /96.7%, 96.3%/91.1%, and 93.6%/94.0%. Only 3.7% of study participants identified in the MDS data, vs 32.6% PDE-based users (kappa value: 0.1), used hypnotics. By combining anxiety medications with hypnotics, the rates of MDS hypnotic or anxiety users increased to 35.0% (kappa value: 0.74). The sensitivity, specificity, PPV, and NPV were 89.5%, 85.6%, 92.8%, and 79.8%, respectively. Agreement between the MDS and PDE in antipsychotic and antidepressant use was high, suggesting that the MDS is a valid tool to measure antipsychotic and antidepressant use. Additional work is needed to understand the disagreements between MDS and PDE in hypnotic use.

Objective To quantify quality of care following an admission to a nursing home with low or high antipsychotic drug use. Background Misuse of antipsychotics in U.S. nursing homes is a huge concern for policymakers. Methods We utilized an instrumental variable approach to estimate the effect of facility-level antipsychotic use on patient outcomes. The instrument was the differential distance to the nearest low-use antipsychotic nursing home relative to the nearest high-use antipsychotic nursing home. Post-acute care short-stay and long-stay residents in U.S. nursing homes were identified using Medicare administrative claims and the Minimum Dataset 3.0 (2014–2019). Outcomes included hospitalizations, falls, pressure ulcers, physical restraint use, medication use, and diagnosis of schizophrenia, bipolar disease, anxiety, or depression. Results Among long-stay residents, receiving care from a low-use facility reduced the diagnosis of schizophrenia, use of restraints, and hospitalizations. There was also a reduction in the hospitalization rate [−0.9 percentage point (pp)], likelihood of long-stay status (−1.8 pp), and diagnosis of schizophrenia (−0.2 pp) at 90 days among short-stay residents. We also observed larger reductions among residents with dementia and serious mental illness. Conclusions Admission to a nursing home with a low use of antipsychotics led to decreased hospitalizations, restraint use, and diagnosis of schizophrenia. Curbing the high use of antipsychotics remains a priority of policymakers as the centers for medicare and medicaid services conducts off-site audits to assess whether nursing homes accurately code residents with schizophrenia. It will be important to monitor if centers for medicare and medicaid services downgrades any quality star ratings due to inappropriate coding and assess the implications on quality of care.

During the initial year of the COVID-19 pandemic, a disproportionate share of COVID-19–related deaths occurred among nursing home residents. Initial estimates of all-cause mortality rates also spiked in early and late 2020 before falling to near or below historical rates by early 2021. During the first 3 years of the pandemic, the US nursing home resident population also decreased by 18% (239 000 fewer residents) compared with pre-pandemic levels. After accounting for these population changes, the all-cause nursing home mortality rate has remained above pre-pandemic levels through the middle of 2023. The peak was in December 2020 at 5692 deaths per 100 000 residents, which was 19% higher than estimates not accounting for the population decrease.

Objective To estimate differences in facility‐level outcomes between nursing homes which reached Institutional Special Needs Plan (I‐SNP) maturity and those which never cared for I‐SNP enrollees. Study Setting and Design We used a difference‐in‐differences design to estimate the effect of I‐SNP maturity, defined as having at least 33.75% of Medicare long‐stayers in the nursing home enrolled in any I‐SNP. Our main outcome was the hospitalization rate in each nursing home‐year. Secondary outcomes included the share of residents with medication use, fall, urinary tract infection, catheter insertion, pressure ulcer, physical restraint use, increased need for help with activities of daily living (ADLs), and mortality. Data Sources and Analytic Sample This repeated cross‐sectional study used 100% Medicare claims, Minimum Data Set assessments, and publicly available Medicare Advantage (MA) plan characteristics data (2004–2021). We included all MA beneficiaries who resided in US nursing homes which reached I‐SNP maturity and those without I‐SNP enrollees. Principal Findings We identified 2530 nursing homes which reached I‐SNP maturity (treated) and 9830 nursing homes without I‐SNP enrollees (untreated). There were some differences observed between these nursing homes, including shares of residents who were White (76.42% vs. 84.84%) and on Medicaid (66.94% vs. 55.45%). These nursing homes were also larger on average (141.76 beds vs. 87.56 beds). From the difference‐in‐differences model, nursing homes which reached I‐SNP maturity experienced declines of 4.1 percentage points (pp) for hospitalizations, 1.0 pp for pressure ulcers, 1.3 pp for urinary tract infections ( p < 0.001) alongside increases in the need for help with ADLs, use of antipsychotics, falls, and physical restraints. Conclusions Nursing homes which reached I‐SNP maturity experienced fewer hospitalizations and pressure ulcers but a decline in function and increase in other negative outcomes. I‐SNPs may be a promising model to improve care for long‐stay residents, but more research is needed to understand potential adverse consequences.

Institutional Special Needs Plans (I-SNPs) are a type of Medicare Advantage (MA) plan designed for long-term residents of nursing homes or those who live in the community but need a nursing home level of care. I-SNPs provide for on-site nurse practitioners in nursing homes to help improve primary care, care planning, and care coordination. Our study was the first to undertake a nationwide descriptive analysis of the I-SNP market. The share of long-stay nursing home residents enrolled in I-SNPs quadrupled from 2006 to 2021, from 2.2 percent to 8.8 percent. Despite this growth, nearly 70 percent of nursing homes did not have any residents enrolled in I-SNPs in 2021, and in more than 60 percent of US counties, there were no I-SNPs available. Nearly 94 percent of long-stay nursing home residents who were enrolled in I-SNPs in 2021 were eligible for both Medicare and Medicaid. More research is needed to understand why I-SNPs enter certain markets and why long-stay nursing home residents choose to enroll in them; to assess I-SNP enrollment barriers; and to compare hospital admission rates and quality outcomes among long-stay nursing home residents by I-SNP enrollment status. Increased understanding of I-SNPs could inform policies to address long-standing concerns about the quality of care in nursing homes.