Eileen Graham’s research while affiliated with Northwestern University and other places

Objectives Adverse Childhood Experiences (ACEs) are associated with increased mortality risk. Individuals with a history of certain adversity during childhood tend to report higher levels of loneliness in later life. In our pre-registered study, we examined whether loneliness mediates the ACEs to mortality risk relation. Methods Participants were from the Midlife in the United States Survey (N = 4963; M [SD] 46.44 [12.52] years, 53.3% female). Follow-up period spanned 26 years. A comprehensive measure of ACEs was employed consisting of 20 ACEs from five categories: physical abuse, emotional abuse, socioeconomic disadvantage, adverse family structure, and poor health at age 16 years. Results ACEs was a significant predictor of mortality risk. Loneliness mediated the ACEs-mortality risk relation. In other words, loneliness in adulthood accounted for the relation between ACEs and future death. These effects withstood a range of sensitivity checks and adjustments for important factors, such as social isolation. Discussion Loneliness appears to be a central mechanism in the long-term impact of ACEs on longevity, such that, for adversity during childhood, loneliness experienced during adulthood may be a toxic pathway to future death.

People who are higher in conscientiousness, extraversion, and agreeableness and lower in neuroticism tend to live longer. The present research tested the hypothesis that personality trait change in middle and older adulthood would also be associated with mortality risk, above and beyond personality trait level. Personality trait change may causally influence mortality risk through corresponding changes in health behaviors, social processes, and stress experience. Alternatively, personality trait change may be a marker of successful or unsuccessful adaptation to life circumstances, which in turn influences mortality risk, or shared risk factors may impact personality trait change and mortality risk. In the latter case, personality trait change may serve as a “psychosocial vital sign” pointing toward increased risk. In 11 samples of middle-aged and older adults (combined N = 32,348), we used multilevel growth curve models to estimate personality trait level and personality trait change across three to 11 measurement occasions spanning 6–43 years. Next, we used Cox proportional hazards models to test whether personality trait level and personality trait change were associated with mortality risk. Higher conscientiousness (hazard ratio [HR] = 0.83), extraversion (HR = 0.93), and agreeableness (HR = 0.88) were associated with longer survival while higher neuroticism was associated with shorter survival (HR = 1.22). In contrast to personality trait level, we found limited evidence for associations between personality trait change and mortality risk. We discuss conceptual and methodological implications of the present findings that may guide future research on associations between personality trait change, health, and mortality.

Introduction Patient‐reported outcomes are increasingly being utilized in clinical settings to identify psychological symptoms and track fluctuations over time. Some clinicians and researchers have expressed concerns about the validity of symptom questionnaires cognitively impaired populations. We sought to determine if differential item functioning (DIF) is present based on cognitive impairment using the patient reported outcomes measurement information system (PROMIS) inventories of anxiety and depression. Methods Data from a longitudinal aging cohort study in Chicago, IL. Mild impairment was defined as a z‐score between ‐1 and ‐1.49 and moderate/severe impairment was a z‐score of ≤‐1.5 on 2+ tests in one of 5 cognitive domains. PROMIS Anxiety and Depression measures were utsed. Due to sample size and response variation, the anxiety model examined DIF with participants categorized using three categories (none, mild, or moderate/severe cognitive impairment). The depression model used 2 categories (none/mild vs moderate/severe impairment). Analyses were conducted in R (lordif package with Monte Carlo simulations). Effect sizes for DIF were calculated by test item. Test characteristic curves (TCCs) examined measure performance by cognitive impairment. Results Data was available for 396 participants, who were 71 years old, mostly female (n = 285) and about half were non‐Hispanic White (n = 214). Approximately 72% had no cognitive impairment, 14% had mild cognitive impairment and 13% had moderate/severe impairment. There was evidence of DIF in 4 of 7 anxiety items and 3 of 8 depression items. The direction of DIF varied, suggesting the response patterns of moderately/severely impaired participants resulted in psychometric overestimation of anxiety and underestimation of depression. Effect sizes of DIF were relatively small for anxiety (0.016‐0.034) and depression (0.014‐0.043) and TCC curves were similar by cognitive impairment category for both measures. Conclusion There is psychometric measurement bias on PROMIS anxiety and depression measures by severity of cognitive impairment. Researchers and clinicians should be aware that cognitive impairment may impact responses when using assessments, however, this study showed this bias is small and results in negligible change when examining the measures overall. These findings support the conclusion that PROMIS Anxiety and Depression are reliable assessments when used among individuals with different levels of cognitive impairment.

Background Impaired functional status is a central diagnostic feature of Alzheimer’s disease and related dementias (ADRD). Informant reporting is often relied upon, given concerns surrounding the ability of persons with ADRD to validly self‐report symptoms. We sought to investigate how cognitive impairment severity impacts psychometric properties of the Patient‐Reported Outcomes Measurement Information System Physical Function (PROMIS‐PF) scale. Methods Data from 396 older adults participating in a longitudinal cohort study on aging was used for this analysis. Cognitive function was assessed using an extensive neuropsychological battery consisting of 13 tests across 5 cognitive domains; z‐scores were calculated for each test and used to determine cognitive impairment severity. Physical function was measured using the PROMIS‐PF short‐form 10a. Differential item functioning (DIF) analyses were conducted using the lordif package in R with Monte Carlo simulations. Results Participants were on average 71.2 years old (SD 5.3), primarily female (72.0%), white (54.3%), and well‐educated (60.0% at least some college). Nearly two‐thirds were cognitively normal (72.5%), while 14.4% and 13.4% had a mild and moderate/severe cognitive impairment, respectively. Seven out of 10 PROMIS‐SF items were flagged for DIF. Direction of DIF varied; participants with mild impairment exhibited a tendency to underestimate their physical function, while those with moderate/severe impairment were characterized by a tendency to overestimate. McFaddens pseudo R² effect sizes were all below published thresholds of 0.13 (range: 0.01‐0.05). Test characteristic curves (TCC) were overlapping between the 3 groups. Conclusions Findings indicate the presence of DIF by cognitive impairment severity on the PROMIS‐PF measure, although the magnitude of effect and overlapping TCC curves suggest the observed DIF was negligible. Nevertheless, researchers and clinicians should be aware that individuals with a cognitive impairment may respond differently to certain items on this measure and further study is needed to evaluate measurement bias among patients with cognitive impairment, including ADRD. However, overall, our evidence suggests patients with varying levels of cognitive impairment can provide reliable estimates of physical function using this measure. This has implications for researchers and clinicians seeking to assess functional status among older adults with ADRD, particularly if supplemental reports of functional status are not feasible or unavailable.

Multimorbidity is an increasing concern among clinicians, as older adults accumulate chronic conditions that require complex coordination of care. There is an increasing interest among research and clinical communities to understand the impacts of multimorbidity on individuals and how quality of care can be optimized to improve quality of life in the aging population. The primary objective of the Claude D. Pepper Older Americans Independence Centers national network is to advance scientific knowledge that will help older adults maintain function, independence, and quality of life. The current symposium highlights recent work put forth by three of the 15 National ‘Pepper’ Centers (Northwestern, Mt. Sinai, UCSF). Graham will present recent findings on how multimorbidity is longitudinally related to physical and mental health-related quality of life, and their cognitive, psychological, social, and health-related determinants. The next three presentations explore the role of caregivers in supporting the health and quality of life of older adults with multimorbidity. O’Conor will discuss complexity of family care dyads management of MCC, by reviewing patient and caregiver perspectives on challenges of successful management at home. Allison will discuss the interplay of care support structures and dementia care needs in the home. Reckrey will discuss results from a qualitative analysis exploring perspectives of geriatricians on the importance of paid caregivers of older adults with MCC. In sum, this symposium will present novel evidence showing the importance of understanding of how multimorbidity and dementia impact caregiving considerations, primary care decisions, and health related quality of life for older adults.

Psychosocial researchers have long sought to identify individual differences in early and middle adulthood that robustly predict health in later life. Our current pre-registered study builds on these efforts by assessing whether an individual’s subjective perception of their health is related to their cognitive health, both concurrently and later in life. Using the coordinated data analysis (CDA) methodological framework, we have identified 17 longitudinal studies (including 7 from the Gateway to Global Aging network) that had the requisite data to test our models. We conceptually harmonized all study variables and tested identical linear and logistic regression models across all study datasets. Our preliminary results suggest that individuals who have better subjective beliefs about their health have better cognitive function both concurrently and at longitudinal follow-up (ranging from 4 to 27 years after baseline). These associations were moderated by participant gender and age at baseline, and the results were somewhat robust to the inclusion of objective health indicators. We will discuss our results within the context of Lifespan Developmental Theory and the Health Behavior Model. Overall, our study serves to elucidate how subjective perceptions of health may manifest into later cognitive health outcomes, and if subjective perceptions of health could be used as an early risk factor for identifying individuals who may be vulnerable to cognitive senescence in later life.

Personality has been found to contribute to individual differences in health and aging. In search of the underlying mechanisms, health behavior and stress have been shown to play roles in the personality-health association. However, given the evidence that personality traits are dynamic within individuals, it remains unclear how the dynamic intraindividual connections between personality and health behavior engagement and stress experiences contribute to interindividual differences in health and health change over time. Using 10-wave data from the Longitudinal Internet Studies for the Social Sciences (N = 3,421), the current study examined the effects of the intraindividual interplays between changes in personality and changes in health behaviors and stress on individual differences in different health outcomes and changes in health outcomes. Adopting multilevel structural equation modeling, the results suggested that after controlling personality at the interindividual level, the dynamic intraindividual interplays between extraversion, openness and smoking at each time point still displayed significant effects on interindividual differences in smoking, which in turn significantly contributed to interindividual differences in self-rated health and symptom complains in the last wave and changes in general disease level and cardiovascular diseases over time. Meanwhile, the dynamic intraindividual interplays between extraversion, conscientiousness and perceived stress were significantly related interindividual differences in perceived stress, which were significantly linked to interindividual differences in the levels of and changes in all health outcomes. The current findings enhance our understanding of how the dynamic intraindividual connections between personality and people’s behaviors and experiences may inform interindividual differences in health during aging.

Early and midlife socioeconomic status (SES) are linked to all-cause mortality, but few studies have examined their independent and joint associations using multidimensional, prospective SES measures and track mortality over extended periods. As described in the Dorame et al. presentation in this symposium, the Boston Early Adversity and Mortality Study (BEAMS) acquired prospective data on the early-life circumstances of three well-characterized male cohorts via linkage to the 1900-1940 US Federal Censuses. This study leverages BEAMS data to test the independent and joint associations of early and midlife SES with all-cause mortality in 1,303 men from a BEAMS cohort, the Normative Aging Study. Early SES, assessed at baseline (earliest available Census before age 18), was a composite combining paternal literacy (1=yes, 0=no) and z-standardized paternal occupational income. Midlife SES was a z-score composite of men’s occupational standing, education, and income at study entry (mean age=42). Mortality status was followed through December 2020 (mean follow-up since baseline: 77 years). Early SES was weakly and positively correlated with midlife SES (r=.17, p<.0001). In Cox regression, each 1-SD higher early SES was associated with 8% lower mortality risk (95%CI: 0.84, 1.00). Each 1-SD higher midlife SES was associated with 3% lower mortality risk (95%CI: 0.95,0.99), adjusted for early SES. The main effect of midlife SES weakened the association of early SES with mortality by only 2%. Early SES did not moderate the association of midlife SES with mortality. Our findings support largely independent contributions of early and midlife SES to all-cause mortality risk.

The global prevalence of multimorbidity is increasing as the population ages. Older adults are very likely to develop multiple chronic conditions, and nearly two-thirds of the older adult population in the U.S. are estimated to be experiencing two or more chronic conditions. The current study examined whether multimorbidity was associated with longitudinal patterns in health-related quality of life (i.e., anxiety, depression, physical function) and whether these associations were moderated by sociodemographic factors (i.e., sex, race, marital status, income, insurance, education). The present pre-registered study used data from the Health Literacy and Cognitive Function Among Older Adults Longitudinal Study (LitCog), a prospective cohort study of English-speaking older adults (N=900). At each measurement occasion, participants reported anxiety, depression, and physical function using the Patient Reported Outcomes Information System (PROMIS), current chronic conditions, and socio-demographic characteristics. We employed multilevel growth models to estimate changes in health-related quality of life, with multimorbidities as a predictor and socio-demographics as covariates. Results indicated that individuals with multiple chronic conditions reported persistently high levels of anxiety and depression, and worse physical function. We found evidence for racial health disparities, such that individuals who identified as non-white experienced worsening health-related quality of life as multimorbidities increased, relative to white participants. These results contribute to the current conversation about the long term impacts of structural and systemic barriers experienced by minoritized groups. We further discuss the public health implications of multimorbidity in older adulthood.