Skip to main content
. Author manuscript; available in PMC: 2022 Dec 1.
Published in final edited form as: J Am Geriatr Soc. 2021 Sep 30:10.1111/jgs.17464. doi: 10.1111/jgs.17464

FIGURE 1.

FIGURE 1

Difference-in-difference regressions with individual- and nursing home–fixed effects to predict schizophrenia reporting among nursing home residents with and without ADRD documentation, pre- and post-partnership, by race. This figure is a visual representation of the fully adjusted model in Table 2. The fully adjusted model adjusts for age, gender, CHESS, ADL, and behavioral expressions of dementia. ADRD, Alzheimer’s disease and related dementias. ADL, activities of daily living: A scale score ranging from 0, complete independence to 5, total dependence, created by summing seven ADL items on a scale from 0–4 (with 0 = complete independence and 4 = total dependence for each item).25 CHESS, Changes in Health, End-stage disease and Symptoms and Signs: A predictive scale score ranges from 0, most stable to 5, or least stable.26 To measure the behavioral expressions of dementia, we use the Agitated and Reactive Behavior Scale (ARBS) to evaluate resident agitation and resistance to care.27 The ARBS score ranges from 0 to 12, where 0 represents no behaviors and 12 represents severe behaviors. We estimated a difference-in-differences regression model with a triple interaction for race, dementia, and the postpartnership period and present the p-value from the triple interaction coefficient. Pre- and postpartnership predicted values were obtained using predictive margins following the estimation of the model