Risk Factors of Skilled Nursing Facility Admissions and the Interrelation With Hospitalization and Amount of Informal Caregiving Received

Longitudinal Care Network Changes and Associated Healthcare Utilization Among Care Recipients

This study examines caregiver networks, including size, composition, and stability, and their associations with the likelihood of hospitalization and skilled-nursing facility (SNF) admissions. Data from the National Health and Aging Trends Study linked to Center for Medicare and Medicaid Services data were analyzed for 3855 older adults across five survey waves. Generalized estimating equation models assessed the associations. The findings indicate each additional paid caregiver was associated with higher adjusted risk ratios (aRR) for hospitalization (aRR = 1.24, 95% CI 1.10-1.41) and SNF admission (aRR = 1.28, 95% CI 1.06-1.54) among care recipients, a pattern that is also observed with the addition of unpaid caregivers (hospitalization: aRR = 1.13, 95% CI 1.06-1.20; SNF: aRR = 1.12, 95% CI 1.02-1.23). These results suggest that policies and approaches to enhance the quality and coordination of caregivers may be warranted to support improved outcomes for care recipients.

A Joint Model for Disability, Self-Rated Health, and Mortality Among Medicare Beneficiaries-Differences by Chronic Disease and Race/Ethnicity

OBJECTIVES

Quantifying interdependence in multiple patient-centered outcomes is important for understanding health declines among older adults.

METHODS

Medicare-linked National Health and Aging Trends Study data (2011-2015) were used to estimate a joint longitudinal logistic regression model of disability in activities of daily living (ADL), fair/poor self-rated health (SRH), and mortality. We calculated personalized concurrent risk (PCR) and typical concurrent risk (TCR) using regression coefficients.

RESULTS

For fair/poor SRH, highest odds were associated with COPD. For mortality, highest odds were associated with dementia, hip fracture, and kidney disease. Dementia and hip fracture were associated with highest odds of ADL disability. Hispanic respondents had highest odds of ADL disability. Hispanic and NH Black respondents had higher odds of fair/poor SRH, ADL disability, and mortality. PCRs/TCRs demonstrated wide variability for respondents with similar sociodemographic-multimorbidity profiles.

DISCUSSION

These findings highlight the variability of personalized risk in examining interdependent outcomes among older adults.

Review and evaluation of imputation methods for multivariate longitudinal data with mixed-type incomplete variables

Estimating relationships between multiple incomplete patient measurements requires methods to cope with missing values. Multiple imputation is one approach to address missing data by filling in plausible values for those that are missing. Multiple imputation procedures can be classified into two broad types: joint modeling (JM) and fully conditional specification (FCS). JM fits a multivariate distribution for the entire set of variables, but it may be complex to define and implement. FCS imputes missing data variable-by-variable from a set of conditional distributions. In many studies, FCS is easier to define and implement than JM, but it may be based on incompatible conditional models. Imputation methods based on multilevel modeling show improved operating characteristics when imputing longitudinal data, but they can be computationally intensive, especially when imputing multiple variables simultaneously. We review current MI methods for incomplete longitudinal data and their implementation on widely accessible software. Using simulated data from the National Health and Aging Trends Study, we compare their performance for monotone and intermittent missing data patterns. Our simulations demonstrate that in a longitudinal study with a limited number of repeated observations and time-varying variables, FCS-Standard is a computationally efficient imputation method that is accurate and precise for univariate single-level and multilevel regression models. When the analyses comprise multivariate multilevel models, FCS-LMM-latent is a statistically valid procedure with overall more accurate estimates, but it requires more intensive computations. Imputation methods based on generalized linear multilevel models can lead to biased subject-level variance estimates when the statistical analyses involve hierarchical models.