Innovation in Aging, 2018, Vol. 2, No. S1 entire follow-up period. There were no differences in health services variables between the programs until the final year. SESSION 2495 (PAPER) LONG TERM CARE THE EFFECT OF PRIMARY CAREGIVERS’ EMPLOYMENT STATUS ON OLDER PEOPLE’S CARE RECEIPT FROM DIFFERENT SOURCES P. Doty1, A. Janus2, 1. Office of the Assistant Secretary for Planning and Evaluation, U.S. DHHS, 2. University of Edinburgh Social gerontologists have predicted that increased female employment would adversely affect the ability of disabled older people to rely on unpaid family caregivers. This paper is one of only a few that focuses on not only whether disabled older people receive fewer hours of help from employed primary caregivers but also whether other members of the care network compensate for any reduction in primary caregiver hours. We conduct a pooled cross-sectional analysis of data on 2,268 older people and their primary caregivers from the National Study of Caregiving I and II, which was administered as part of rounds 1 and 5 of the National Health and Aging Trends Study. Consistent with previous studies, we find that employed primary caregivers provide fewer hours of help per week compared to non-working primary caregivers (27 versus 40 hours, p < .01). Secondary caregivers’ hours do not change in response to the employment status of GSA 2018 Annual Scientific Meeting primary caregivers. However, older people with an employed primary caregiver make greater use of paid help compared to older people with a non-working primary caregiver (8 versus 5 hours, p < .01). Therefore, while older people partially compensate for the reduction in hours from employed primary caregivers through greater use of paid help, older people with an employed primary caregiver still receive fewer total hours of help overall (-10 hours, p < .01). We also find that the compensatory effect of paid help is greater among older people with 2+ ADLs, dementia, and higher incomes. CHARACTERIZING STAFF INTERACTIONS IN A SAMPLE OF HIGH PERFORMING NURSING HOMES A. Snow1, V. Clark2, S. Zhao2, P. Nash3, R. Allen4, R. Bolton5, C. Hartmann6, 1. Tuscaloosa VA Medical Center & University of Alabama, 2. Center for Healthcare Organization and Implementation Research, Edith Nourse Rogers Memorial Veterans Hospital, 3. Tuscaloosa VA Medical Center, 4. The University of Alabama, 5. Bedford VA Medical Center, 6. Center for Healthcare Organization and Implementation Research, Edith Nourse Rogers Memorial Veterans Hospital & Boston University Increasing evidence indicates nursing home staff interactions with residents and each other contribute to care quality. We characterized interaction in a sample of top performing person-centered care (PCC) adopters in the Veterans Administration (VA) system of 134 Community Living Centers (CLCs– i.e., nursing homes). We identified 8 topperforming CLCs using a combination of quality and PCC indices. At these 8, staff participated in a validated survey of relational coordination. Researchers also conducted structured observations of staff and resident behavior in public areas. Analyses were descriptive. 392 staff responded to the survey. Researchers conducted 2,459 observations. Across all CLCs, staff reported the quality of their interactions within (WI) their workgroup consistently lower than the quality of their interactions between (BW) their and other workgroups. BW ratings: 24% strong, 34% moderate, 42% weak. WI ratings: 92% weak. Strong ratings varied by workgroup type and by WI and BW. Nursing was more likely to receive strong ratings than other workgroups but less likely to assign strong ratings to other groups. 31% of observations were of staff performing direct resident care while communicating with the resident and 39% were of staff having no communication with the resident during care. Results in this topperforming sample indicate potential room for continued improvement. Given our findings of rather low relational coordination and interaction quality, quality improvement interventions directly targeting staff communication and staff-resident interaction might represent opportunities for positive impact not only in high PCC performers but across the PCC performance spectrum. ENHANCING STAFF-RESIDENT INTERACTIONS THROUGH A FRONTLINE STAFF QUALITY IMPROVEMENT INTERVENTION C. Hartmann1, J. Palmer2, C. Pimentel3, R. Allen4, S. Zhao5, N. Wewiorski6, K. Dillon6, A. Snow7, 1. Center for Healthcare Organization and Implementation Research, Downloaded from https://academic.oup.com/innovateage/article/2/suppl_1/791/5170632 by guest on 09 January 2024 WHAT DOES EXPLAIN THE REGIONAL VARIATION IN LIVING AT HOME OR IN INSTITUTION WHEN DEPENDENT IN FRANCE? A. Carrere1, E. Cambois2, 1. INED, 2. French Institute for Demographic Studies (INED) Population aging raises the question of the long-term care (LTC) organization for the dependent elderly. In France, “in home” LTC is promoted to meet both individual and policy (partly cost-related) preferences. However, LTC management is decentralized at a county level inducing variations in the type of LTC supply, partly due to (unequal) local resources. This article aims to identify county variations in living in institution for dependent elderly people (60+) and how much different LTC supply contributes to these variations. We used different data sources to estimate the probability of living in institution versus at when having disabilities (i.e. bathing difficulties): survey EHPA 2015 for population living in institution (n=308,312) and survey VQS 2014 for population living at home (n=161,419). We complemented these sources with data on contextual LTC supply and socioeconomic status (SES) indicators. Multilevel regression models are computed to assess how much of the county variance of institutionalization decreases when controlling for in LTC supply and SES and identify the share of variance explained by the differences in LTC supply and SES. We found that county variation explained 21% of the institutionalization risk: of which 40% is explained by SES contextual differences and 7% by difference in the LTC supply. Further researches are needed to determine whether these differences reflect county’s unequal resources to supply LTC or county’s different preferences. 791 792 SOCIODEMOGRAPHIC DIFFERENCES IN LONGTERM CARE USE DURING THE LAST 2 YEARS OF LIFE IN SWEDEN S. Kelfve1, J. Wastesson2, B. Meinow3, 1. Linkoping University, 2. Aging Research Center, Karolinska Institutet, 3. Aging Research Center, Karolinska Institutet & Stockholm University Most people experience a period of poor health and dependency on long-term care (LTC) use in the end of life. However, little is known about how the length of this period and the type of LTC used before death varies between sociodemographic groups in Sweden. Using register data covering all individuals 65 years and older who died in Sweden in November 2015 (n=6405), we aimed to investigate the association between sociodemographic factors (sex, civil status, education and migration) and LTC use (home-help service or institutional care) during the last 2 years of life. Overall, men died without having any LTC to larger extent than women (32% and 18%, respectively) and a larger proportion of women (44%) died in institutional care compared to men (28%). Married people died without having any LTC more frequently than nonmarried, (38% vs 26% among men and 33% vs 14% among women). Persons born outside of Sweden used slightly less LTC than Swedish born people. Among immigrant men, 38% died without having any LTC, but only 18% among immigrant women. Corresponding numbers among Swedish born was 32% among men and 23% among women. Finally, the results show that persons with higher educational level more frequently died without LTC and this difference was more pronounced among women than men. Results from this study show substantial differences in LTC use between sociodemographic groups in Sweden. Future studies should investigate if the “aging-in-place policy” that has been increasingly put into practice has affected the socioeconomic differences in LTC use. SESSION 2500 (SYMPOSIUM) ADVANCES IN LIFE COURSE RESEARCH ON AGING: NEW OPPORTUNITIES AND FINDINGS FROM LARGE REPRESENTATIVE PANEL STUDIES Chair: J. Smith, University of Michigan, Ann Arbor, Michigan Discussant: D. Dannefer, Case Western Reserve University, Cleveland, Ohio There is increasing multidisciplinary interest in understanding the longterm influences of early-life factors on latelife health and wellbeing outcomes and differences in the sequence and timing of events in life histories. Although life course theories formalize propositions about pathways from early to later life, research has been limited to date because few existing longitudinal studies cover the entire life of individuals or include comprehensive coverage of multiple life domains. Recent advances in the collection of retrospective life history information have opened ways to supplement the prospective life histories obtained in large panel studies of aging. Speakers in this session illustrate some of these new life course research opportunities. Wahrendorf and colleagues used data from the Survey of Health, Ageing and Retirement in Europe (SHARE) to identify sequences of residential housing characteristics. They report associations between early-life housing histories and late-life health outcomes. Where people live can afford different early-life educational opportunities and exposure to stressful life events. Two papers use new retrospective life history data collected in the Health and Retirement Study (HRS) to examine questions about these factors and late-life cognitive aging: Lee and Smith report associations with childhood educational contexts and experiences and Helppie-McFall and colleagues examine the cumulative role of stressful life events. Analyses of life calendars from the Swiss Vivre Survey by Spini and colleagues highlight the impact of historical events on life periods participants describe as happy or vulnerable. Dannefer concludes with an integrative discussion of current efforts to advance life course research on aging. HOUSING HISTORIES OVER THE LIFE COURSE AND HEALTH AT OLDER AGES: RESULTS FROM SHARE M. Wahrendorf1, B. Vanhoutte, PhD2, E. Courtin, PhD3, 1. Heinrich-Heine-University Düsseldorf, Düsseldorf, GSA 2018 Annual Scientific Meeting Downloaded from https://academic.oup.com/innovateage/article/2/suppl_1/791/5170632 by guest on 09 January 2024 Edith Nourse Rogers Memorial Veterans Hospital & Boston University, 2. Institute for Aging Research, Hebrew SeniorLife, 3. Edith Nourse Rogers Memorial Veterans Hospital & UMass Medical School, 4. The University of Alabama, 5. Center for Healthcare Organization and Implementation Research, Edith Nourse Rogers Memorial Veterans Hospital, 6. Edith Nourse Rogers Memorial Veterans Hospital, 7. Tuscaloosa VA Medical Center & University of Alabama For nursing home residents, positive interactions with staff and engagement in activity contribute meaningfully to quality of life. We sought to improve interactions between staff and residents, with the ultimate aim of enhancing resident engagement. We used an opportunistic, snowball sample of 6 Veterans Health Administration nursing homes (e.g., Community Living Centers—CLCs). Our staff-based behavior-change intervention involved staff using a set of evidence-based practices for implementing quality improvement: frontline huddles, rapid-cycle structured observations, a strengths-based approach, and the importance of brevity coupled with regularity. CLC staff facilitated the intervention, with some assistance from researchers. Research data comprised validated resident and staff surveys, researcherconducted structured pre- and post-intervention observations, and semi-structured post-intervention staff interviews. 62 CLC residents and 308 staff members responded to the surveys. Researchers conducted 1,490 discrete observations and 66 interviews. In generalized linear models using observation and survey data, intervention implementation was associated with increased staff-resident communication during the provision of direct care activities (β = 0.083, 95% CI: 0.04–0.126, adjusted P=.0012) and decreased negative staff interactions with residents (β = -0.035, 95% CI: -0.062 – -0.009, adjusted P=0.0288). In interviews, staff consistently credited the intervention with helping them (a) develop awareness of the importance of identifying opportunities for engagement and (b) improve the quality of their interactions with residents. Overall, the intervention proved feasible. The combination of rapid-cycle observations, huddles, and a focus on identified strengths influenced staff to make simple enhancements to their behaviors that improved staff-resident interactions and staff-assessed resident engagement. Innovation in Aging, 2018, Vol. 2, No. S1