502 CHILDHOOD PHYSICAL ABUSE INCREASES THE RISK OF SUBJECTIVE MEMORY IMPAIRMENT Lauren Roach1, Tyler Bell2, and Caitlin Pope3, 1. University of Kentucky, Holland, Michigan, United States, 2. University of California, San Diego, La Jolla, California, United States, 3. University of Kentucky, Lexington, Kentucky, United States Subjective memory impairment, defined as self-reported difficulties in recall and learning, doubles the risk of Alzheimer’s Disease and related dementia, despite being weakly related to objective memory decline. Because of its strong stability over time, it may be possible that subjective memory impairment reflects earlier life risk factors for dementia such as adverse childhood experiences. It is reported that over a fifth of older adults worldwide experienced physical abuse during childhood. Previous cross-sectional studies suggest physical abuse is associated with later cognitive impairment. Still unclear, are the longitudinal associations between childhood abuse and subjective memory impairment in later life. Using a sample of adults drawn from the Health and Retirement Study (n = 19,185, Mage = 67.05, SD = 11.33) we assessed associations between reported physical abuse by a parent before the age of 18 and subjective memory impairment (current memory problems and perceived memory decline) over periods of up to 18 years. Generalized linear mixed models examined longitudinal associations between childhood physical abuse and subjective memory impairment while controlling for depressive symptoms and other empirically relevant covariates. Experiencing childhood physical abuse was associated with increased likelihood of reporting more current memory problems (OR = 1.17, 95% CI 1.04, 1.33) and perceived memory decline in later life (OR = 1.27, 95% CI 1.13, 1.43). Findings suggest childhood physical abuse is associated with subjective memory impairment, a strong predictor of dementia. Understanding early life conditions, including adverse childhood experiences may help explain associations between subjective memory impairment and dementia risk. IS THE HEALTH OF OLDER AMERICANS WITH A GED EQUIVALENT TO THEIR PEERS WITH A HIGH SCHOOL DIPLOMA? Esme Fuller-Thomson, Robin Grossman, and Andie MacNeil, University of Toronto, Toronto, Ontario, Canada Previous research has found higher levels of educational attainment to be strongly associated with better health outcomes in later life, such as better cognitive functioning and fewer functional and sensory impairments. However, most studies have grouped General Educational Development (GED) recipients with high school graduates, neglecting potential differences in socioeconomic status, health behaviours, and health outcomes among these two groups. The aim of the current study is to identify differences in the age-sex-racepoverty adjusted prevalence and odds of cognitive impairment, hearing impairment, vision impairment, limitations in activities of daily living (ADLs), and ambulation limitations among three groups of older American adults: high school dropouts, GED recipients, and high school graduates with no post-secondary education. The present study uses secondary analysis of the 2017 American Community Survey, a nationally representative survey of community-dwelling and institutionalized older adults aged 65 years and older, of whom 20,489 were GED recipients, 154,892 had a high school diploma and 49,912 had finished grade 8 but had not completed high school. Our findings indicate that there is a gradient in health outcomes among Americans aged 65-84, with the highest prevalence and odds of cognitive impairment, hearing impairment, vision impairment, ADL limitations, and ambulation limitations occurring among high school dropouts, followed by GED recipients, and the lowest prevalence among high school graduates. These findings suggestion that although GED recipients have better health outcomes than high school dropouts, there is still a significant disparity in health status between GED recipients and high school graduates. HISTORY OF JOB STRAIN AND RISK OF LATE-LIFE DEPENDENCY: A NATIONWIDE SWEDISH REGISTERBASED STUDY Ingemar Kåreholt1, Charlotta Nilsen1, Janne Agerholm2, Susanne Kelfve3, Jonas Wastesson2, Kirsten Nabe-Nielsen4, and Bettina Meinow5, 1. Jönköping University, Jönköping, Jonkopings Lan, Sweden, 2. Karolinska Institutet, Stockholm, Stockholms Lan, Sweden, 3. Linköping University, Linköping, Ostergotlands Lan, Sweden, 4. University of Copenhagen, Copenhagen, Hovedstaden, Denmark, 5. Stockholm Gerontology Research Center, Stockholm, Stockholms Lan, Sweden There is substantial evidence that work plays a significant role in post-retirement health. Yet little is known about its role in when late-life dependency may occur. We examined associations between job strain and the risk of entering late-life dependency. Individually linked nationwide Swedish registers were used to identify people 70+ alive in January 2014, and who did not experience the outcome (late-life dependency) during two months prior to the start of the follow-up. Late-life Downloaded from https://academic.oup.com/innovateage/article/6/Supplement_1/502/6938329 by guest on 23 December 2022 of cognitive aging, it does not account for the impact of micro-level measures of neighborhood stressors on cognitive function. Therefore, the purpose of the current study was to determine whether violent crime incident proximity (VCIP) contributes to later-life cognitive function above and beyond ADI in older African Americans. Participants (N=147; M= 68.34) from an ongoing study, Pathways to Healthy Aging in African Americans—a Rutgers University-Newark community partnership fostered over 16 years of community engagement, health education, and public service—responded to measures of cognitive ability, SDOH, and demographic details. The results show that VCIP is a trending predictor of cognitive performance, when adjusting for age, gender, education, depression, and ADI. The result aligns with our hypothesis that individuals living in areas with greater VCIP will have poorer performance on cognitive tasks. Our findings suggest that for African Americans in an urban setting, hyper-local VCIP appears to be more useful at capturing the impact of neighborhood disadvantage on cognitive decline and Alzheimer’s disease risk. Therefore, for later-life cognitive health in African Americans, it is important to consider micro-level measures of neighborhood stressors such as VCIP. Innovation in Aging, 2022, Vol. 6, No. S1 Innovation in Aging, 2022, Vol. 6, No. S1 SEQUENCING OF PLANNED AND UNPLANNED BIRTHS AND IMPLICATIONS FOR MID- AND LATERLIFE HEALTH AMONG NLSY79 WOMEN Mieke Thomeer1, Clifford Ross1, Rin Reczek2, and Christina Bijou2, 1. University of Alabama at Birmingham, Birmingham, Alabama, United States, 2. Ohio State University, Columbus, Ohio, United States Existing studies demonstrate that unplanned births (e.g., unwanted, mistimed) are associated with worse health for mothers in the short-term and—according to some preliminary evidence—in mid- and later-life. Yet as life course and reproductive career frameworks highlight, childbearing experiences often unfold over a number of years, with a considerable amount of diversity in pregnancy and birth experiences even for the same individual. For example, a person may have an unplanned birth in late adolescence followed by only planned births in early adulthood. In order to provide a more holistic understanding of how birthing experiences births are associated with midlife health, we use Sequence Analysis (SA) on the 1979 National Longitudinal Survey of Youth (NLSY79; N=3,992) to examine how patterning of planned and unplanned births is associated with physical and mental health at ages 50 and 60 (SF-12). Preliminary analysis indicates that compared to respondents with only planned births, respondents with unplanned birth(s) followed by planned birth(s) have worse physical and mental health at midlife, but there is no difference in health for respondents with only planned births, only unplanned births, and planned birth(s) followed by unplanned birth(s). Future analysis with SA will consider how more detailed sequences (e.g., timing, number and type, ordering, spacing) are associated with these midand later-life health outcomes, taking into account selection factors such as childhood SES and educational attainment. This project demonstrates the need for life course perspectives on the long-term health implications of unplanned births, recognizing diversity within and between individuals. COGNITIVE AGING IN THE SHADOW OF THE CLOSET: THE EFFECTS OF IDENTITY CONCEALMENT AND DISCRIMINATION Douglas Hanes, and Sean Clousten, Stony Brook University, Stony Brook, New York, United States Lesbian, gay, and bisexual (LGB) people tend to experience poorer cognitive health as they age; discrimination and identity concealment are common LGB experiences that carry emotional, health, and cognitive tolls. Previous research utilized data on research participants in same-sex relationships (SSR) to successfully identify a subset of LGB people and analyze their aging experiences. The present study relied on one situation in which identity concealment was legally mandated: military service. Until 2011, LGB people were banned from participating in military service in the U.S., leading to either concealment or institutionalized discrimination (i.e., discharge) for most LGB servicepeople. Using Health and Retirement Study (HRS; 1998–2016) data, this project analyzed the combined effects of veteran status and SSR on cognitive performance. Using multilevel longitudinal modelling, we found that while non-SSR veterans had higher mean scores than non-SSR civilian participants (the reference group), SSR civilians and veterans both had mean scores lower than the reference (SSR: β=-0.514, p=0.031; SSR+veteran (β=-1.065 p=0.081). These results may be limited due to the low number of SSR+veteran participants (n = 28). Nevertheless, these results suggest the potential for using veteran status among LGB people to study the health effects of identity concealment and discrimination on aging. EARLY-LIFE EXPOSURE TO THE CHINESE FAMINE OF 1959–1961 AND LATER-LIFE HEALTH: EARLY LIFE AS A CRITICAL PERIOD Mengling Cheng, Nicolas Sommet, Daniela Jopp, and Dario Spini, University of Lausanne, Lausanne, Vaud, Switzerland Barker’s fetal origins hypothesis and the critical period theory suggest that early life events have long-term health effects. However, evidence of the famine exposure in early life and its effects on health in later life is scarce and inconsistent. To explore the effects of early-life exposure to the Chinese famine of 1959-1961 on later-life multimorbidity, we performed Poisson growth curve models using CHARLS Life History 2014 and CHARLS 2011-2018 (42,775 observations from 12,060 respondents). Our analyses revealed two findings. First, there was an overall detrimental effect of the early-life famine exposure on multimorbidity, although there was no effect of severity of famine exposure. Second, there was no overall interaction between famine exposure and life stages, although a more parsimonious model suggested that the detrimental effect of famine exposure was more pronounced in earlier life stages than in later life stages. Findings suggest that early life is a critical period in the life course and provides developmental origins of health and disease in later life. LIFE-COURSE SOCIOECONOMIC STATUS AND MORBIDITY AND MORTALITY IN THE MIDUS NATIONAL SAMPLE Yanping Jiang1, Li Chen2, and Samuele Zilioli3, 1. Rutgers, The State University of New Jersey, New Brunswick, New Jersey, United States, 2. Augusta University, Augusta, Georgia, United States, 3. Wayne State University, Detroit, Michigan, United States Socioeconomic status (SES) is a well-established social determinant of health shaping the distribution of the burden of morbidity and mortality. In this area, a less understood topic Downloaded from https://academic.oup.com/innovateage/article/6/Supplement_1/502/6938329 by guest on 23 December 2022 dependency was operationalized as use of long-term care. Information about job strain was obtained via a job exposure matrice and matched with job titles. Cox regression models with age as time-scale (adjusted for living situation, educational attainment, country of birth, and sex) were conducted to estimate hazard ratios (HR) for entering late-life dependency during the 24 months of follow-up (n=993,595). Having an initial high starting point of job strain followed by an increasing trajectory throughout working life implied a 23% higher risk of entering late-life dependency at a younger age, compared with the reference group (low starting point with a decreasing trajectory). High initial starting point followed by a stable trajectory implied a 12% higher risk of entering latelife dependency at a younger age. High initial starting point followed by a decreasing trajectory implied a 10% risk reduction, and a low starting point with a stable trajectory implied a 22% risk reduction, of entering late-life dependency at a younger age. Reducing stressful jobs across working life may contribute to postponing late-life dependency. 503