PATTERN OF HEALTH-RELATED BEHAVIORS AND RESOURCES AS PREDICTOR OF MEDICAL CARE AND MORTALITY RISK AMONG OLDER ADULTS

Abstract This study investigates the patterns and consequences of a wide range of health-related behaviors and resources that include health-compromising behaviors, health-promoting behaviors, preventive health behaviors, and health risks coping resources. We aim to identify the empirically-derived subgroups of individuals with unique profiles of health behaviors and resources to determine how subgroup membership predicts health outcomes and medical care utilization four years later. Data derived from 5,067 respondents in the 2010 and 2014 waves of the Health and Retirement Study. Latent class analysis was used to define classes based on 13 indicators in the 2010 wave, which also provided sociodemographic and health-related covariates. Outcomes were measured over 4 years. Six latent subgroups were identified: “Best Behavior/Resources”, “Low Social Support “, “Low Physical Activity”, “High Substance Abuse”, “Low Preventive Tests”, and “Low Governmental Health Insurance”. Compared with the “Best” group, older adults identified as “Low Physical Activity” and “High Substance Abuse” were found to have higher mortality risks and a lower likelihood of seeing doctors and less nursing home nights; older adults with the lowest level of receiving flu shots, cholesterol and cancer screen test (“Low Preventive Tests”) reported a less likelihood of seeing doctors; respondents in “Low Governmental Health Insurance” subgroup were associated with a lower likelihood of hospital stay and more nursing home nights. Results suggest that distinct groups of older individuals characterized by their health behaviors and resources provide a basis for identifying the high-risk segment of the older population for intervention.

older adults. To better understand the role of such tools in later life, this study examined the use of internet-based technology in relation to health care service utilization. A nationally representative samples of adults age 55 years and older (n = 1,157) were obtained from the health behaviors module of the 2014 Health and Retirement Study (HRS) in the U.S. We analyzed the use of health management websites and apps and its association with health care service utilization measured by the number of doctor visits during the last two years before the interviews. Negative binomial regression was used to model the number of doctor visits as a function of the health management technology use and other covariates. Results revealed that the level of technology use (b = 0.41, p < 0.05) was significantly associated with a 50% greater number of doctor visits. Such technology use was significantly more common among adults with good health than those with poorer health outcomes (17.32% & 9.60%, respectively, p < 0.05). The findings suggest that healthy adults take advantage of available internet-based self-management tools and proactively utilize health care services more often than individuals who can be categorized as "unhealthy". There is a continuing need for research on introducing strategies that can ultimately reduce health disparities by decreasing the digital divide.

HEALTH RISK BEHAVIORS ADVERSE CHILDHOOD EXPERIENCES AND ALCOHOL-IMPAIRED DRIVING IN LATER LIFE
Shalini Sahoo, 1 and Paul Sacco 1 , 1. University of Maryland-Baltimore, Baltimore, Maryland, United States Adverse Childhood Experiences (ACEs) are associated with a range of negative behavioral health outcomes in adulthood. Those with ACEs may use alcohol as a component of long-term coping, increasing risk of alcohol-impaired driving. Associations between ACEs and alcohol-impaired driving are relatively understudied. Using the 2012/2013 data from the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC), logistic regression models examined the relationship between five types of ACEs (e.g. child abuse and neglect) and lifetime alcohol-impaired driving among a representative sample of American adults aged between 18 to 90 years (N = 36,309). ACEs were positively associated with lifetime alcohol-impaired driving for adults under age 50 (witnessed intimate partner violence (OR = 1.624, p < 0.001), physical abuse (OR = 1.723, p < 0.001), sexual abuse (OR = 1.651, p < 0.001), physical neglect (OR = 1.571, p < 0.001), and emotional neglect, (P > 0.05). We found similar positive associations between ACEs and impaired driving among adults aged 50 and over (witnessed intimate partner violence (OR = 1.398, p < 0.05), physical abuse (OR = 1.751, p < 0.001), sexual abuse (OR = 1.690, p < 0.001), physical neglect (OR = 1.455, p < 0.001), and emotional neglect, p > 0.05). Among adults under age 50, ACEs were associated with past-year alcohol-impaired driving, but this relationship was not seen in adults aged 50 and over. Findings suggest that the effect of ACEs on alcohol-impaired driving is in younger adulthood when alcohol use and driving occur most.

SUICIDE TRENDS AND DISPARITIES AMONG OLDER ADULTS IN THE UNITED STATES, 2008-2017
Sanae El Ibrahimi, 1 Yunyu Xiao, 2 and Matthew L. Smith 3 , 1. University of Nevada,Las Vegas,Las Vegas,Nevada,United States,2. New York University,Silver School of Social Work,New York,New York,United States,3

. Center for Population Health and Aging, Texas A&M University, College Station, Texas, United States
Background: Suicide ranks within the top fifteen causes of death among adults 55 and older in the United States and is a growing concern in the face of social isolation and other end-of-life issues. This study examined differences and trends in suicide rates and methods among older adults in the U.S. Methods: Suicide mortality rates from 2008-2017 were derived from the Multiple Cause of Death files in the CDC's WONDER database. Suicide deaths were identified from the underlying causes of death using ICD-10 codes. Age-adjusted death rates (per 100,000) were calculated. Older adults were grouped into four age categories: 55-64, 65-74, 75-84, and 85+ years. Percent change in suicide rates between 2008-2017 were examined, which were then stratified by gender and top suicide methods. Results: Suicide rates increased by 16% among adults 55 years of age and older from 2008 to 2017 (15.4 vs 17.8 per 100,000 respectively). In 2017, the suicide rate among older adults was 27% higher than the general population (14.0 per 100,000). Suicide rates were significantly higher among men relative to women for those ages 85+ (14:1 ratio of males-to-females). However, females in the 65-74 age group experienced the highest increase of suicide rate (41%) compared to other females or males across age groups. The most common method of suicide was firearms, followed by poisoning and suffocation. Suffocation had the highest increase over time (37%). Conclusion: Rising suicide rates among older adults suggest the need for tailored intervention strategies that address upstream suicide-related risk factors. This study investigates the patterns and consequences of a wide range of health-related behaviors and resources that include health-compromising behaviors, health-promoting behaviors, preventive health behaviors, and health risks coping resources. We aim to identify the empirically-derived subgroups of individuals with unique profiles of health behaviors and resources to determine how subgroup membership predicts health outcomes and medical care utilization four years later. Data derived from 5,067 respondents in the 2010 and 2014 waves of the Health and Retirement Study. Latent class analysis was used to define classes based on 13 indicators in the 2010 wave, which also provided sociodemographic and health-related covariates. Outcomes were measured over 4 years. Six latent subgroups were identified: "Best Behavior/ Resources", "Low Social Support ", "Low Physical Activity", "High Substance Abuse", "Low Preventive Tests", and "Low Governmental Health Insurance". Compared with the "Best" group, older adults identified as "Low Physical Activity" and "High Substance Abuse" were found to have higher mortality risks and a lower likelihood of seeing doctors and less nursing home nights; older adults with the lowest level of receiving flu shots, cholesterol and cancer screen test ("Low Preventive Tests") reported a less likelihood of seeing doctors; respondents in "Low Governmental Health Insurance" subgroup were associated with a lower likelihood of hospital stay and more nursing home nights. Results suggest that distinct groups of older individuals characterized by their health behaviors and resources provide a basis for identifying the high-risk segment of the older population for intervention. Polysubstance use in the U.S. has been a public health concern. The prevalence of substance use among middle-aged and older adults is predicted to increase by 50% by 2020. Previous studies revealed retirement could serve as a risk factor for substance use as this population was known to be more susceptible to mental health issues. However, patterns and determinants of polysubstance use among retired population remained understudied. The purpose of this study was to investigate the patterns and determinants of polysubstance use among retired adults aged 50 and older. This study extracted 3,019 retired participants from the 2017 National Survey on Drug Use and Health study. Polysubstance use was defined as the use of two or more substances, including alcohol, tobacco, marijuana, and painkiller misuse, in the past month. Weighted multinomial logistic regression model was conducted to examine the associations between utilization of mental health treatment and poly-use of substances. The findings suggested 52.0% of retired adults used at least one substance in the past month, where 17.0% used substances concurrently; 15% co-used alcohol, tobacco, and marijuana; 8.6% co-used tobacco and marijuana. Females were less likely to poly-use substances (RRR=0.37, p<.001) than males. Those who had received mental health treatment in the past year were more likely to co-use more than two substances in relative to substance non-users in the past month (RRR=1.71, p<.05). Retirement plan incorporating behavioral intervention and early detection of mental health issues are warranted to reduce polysubstance use among the retired population in the U.S.

SOCIODEMOGRAPHIC DISADVANTAGE, LIVING WITH A SMOKER, AND OBESITY IN MIDDLE-AGED AND OLDER WOMEN
Carole K. Holahan, 1 Charles J. Holahan, 1 Sangdon Lim, 1 Yen T. Chen, 1 and Daniel A. Powers 1 , 1. University of Texas at Austin, Austin, Texas, United States Although sociodemographic disadvantage is a recognized risk factor for obesity, the potential role of living with a smoker in this relationship has been unexamined. This study investigated: (a) the association between sociodemographic disadvantage and living with a smoker, and (b) the role of living with a smoker in partially explaining the link between sociodemographic disadvantage and obesity. The study used limited access data from the Women's Health Initiative Observational Study obtained from NHLBI. Participants were 91,888 women ranging in age from 50 to 79; 6,527 participants reported living with a smoker. Analyses were cross-sectional. Logistic regression analyses examined paths in the proposed model; bias-corrected bootstrapped confidence intervals tested indirect effects. All analyses controlled for age, marital status, and participants' current smoking status. Results demonstrated a significant association (p < .001) between sociodemographic disadvantage and living with a smoker across three measures of disadvantage (for low education, low income, and Black ethnicity, ORs were 1.95, 2.10, and 2.63, respectively), as well as between living with and smoker and obesity (OR = 1.71 The creation and maintenance of sustainable data archives can be challenging but it offers clear advantage. Properly curated data can be used by multiple researchers, testing a variety of hypotheses and increasing the return on investment to the expensive process of data collection. Having an internally managed archival system also provides greater control and autonomy in the equitable distribution of data resources, insuring all researchers have full use of the data for original research, teaching and new directions once the data leaves the control of the local investigator's control. This poster reviews the advantages of having a local strategy geared toward the preservation and sharing of gerontological research data. Using the National Archive of Computerized Data on Aging (NACDA) as a working example, the poster offers an overview of collections at NACDA. Using our metadata tools and variable search database, NACDA can identify studies in its collections that examine aspects of aging and health among adults during their lifecourse. Many of the studies are longitudinal or repeat measure cross-sectional studies. We are also able to identify studies which focus on aging that are not maintained by NACDA but which are available to interested researchers. Using a strategy of archival preservation combined with a strong focus on productive research