AE Sales, VA Puget Sound Health Care System; SC Hedrick, VA Puget Sound Health Care System; J Sullivan, VA Puget Sound Health Care System; SL Gray, University of Washington; M Curtis, California Department on Aging; J Tornatore, VA Puget Sound Health Care System; H Starks, VA Puget Sound Health Care System

Objectives: To examine factors influencing choice of one of three types of community residential care settings: Adult Family Home (AFH), Assisted Living (AL), and Adult Residential Care (ARC, also known as boarding home). These settings are relatively recent innovations in long term care. They are less restrictive overall than skilled nursing facilities, but offer a higher level of care than is typically available in home health. Washington state has been a national leader in developing access to these types of facility. VA is proposing to pilot a program of payment for care in these settings. Little is known about factors influencing choice of setting.

Methods: All interested residents (N= 349) entering three types of community residential care: AFH, ARC, and AL facilities on State (Medicaid) funding in a three county area in Washington state were enrolled in this study. Information on the residents, their informal caregivers, the providers, and settings was collected at entry into setting. Resident information included demographic variables, health status, functional status, behavior problems and cognitive status as well as presence or absence of an informal caregiver (usually a family member). We included facility characteristics: restrictiveness of policies, overall pleasantness of facility, and occupancy rate; and administrator characteristics: years of experience. We also included a market variable, supply of AFH beds, constructed from state data. We analyzed the data using multinomial logit for the three types of facility with AFH as the base category.

Results: Most study residents entered AFH (58%), followed by AL (27%) and ARC (15%). The selection model demonstrated good fit for both AFH (c-statistic=.98) and AL (.97), with worse fit for ARC (.90). Pseudo-R-squared was 0.63 for the model overall. Significant predictors of choice of AL over AFH were age (+, meaning that as age increased, residents were more likely to choose AL); functional status measured by ADLs (-; as functional status decreased, residents were more likely to choose AFH); problem behavior (-); and restrictiveness of facility (+). Both supply of AFH beds and administrator years of experience showed trends towards significance (p=0.053 in both cases, both negative). Significant predictors of choice of ARC over AFH were resident education (dichotomized as high school or not, -); ADLs (-); facility regulations (+); overall pleasantness of facility (-); and mental component score on the SF36 (-).

Conclusions: Factors affecting choice of facility vary by facility type. Both facility and resident characteristics contribute to choice of facility type, and market characteristics, often hypothesized to be determinant, appear to play a lesser role. The dominant role of functional status is notable, and the absence of cognitive status as an influence is also noteworthy. Of particular policy relevance is the influence of facility restrictiveness, which works consistently to lead residents to choose facilities other than AFHs.

Impact: The findings of this analysis will be of interest to decision-makers in long term care in VA as they craft policy to enable payment for these settings for veterans. These findings suggest the need for careful assessment of resident functional status and of facility regulatory status during placement.