Avoiding the Major Causes of Death: Does Childhood Misfortune Reduce the Likelihood of Being Disease Free in Later Life?

Avoiding the Major Causes of Death: Does Childhood Misfortune Reduce the Likelihood of Being... Abstract Objectives Although previous research reveals the detrimental effects of early misfortune on the development of chronic diseases in later life, few studies have investigated its effects on remaining disease free. This study draws on cumulative inequality theory to investigate whether experiencing childhood misfortune reduces the likelihood of remaining disease free over time. Method This study utilizes five waves of data from the Health and Retirement Study to test whether five domains of childhood misfortune predict being disease free at baseline (2004) and developing disease over time (2004–2012). Results Respondents reporting risky parental behaviors during childhood were less likely to be disease free at baseline and had an increased risk of disease onset over time, the latter driven by having a guardian who smoked in combination with more pack-years smoked in adulthood. Furthermore, we find that adult resources, that is wealth, help to mitigate the noxious effects of other misfortunes, notably poor socioeconomic conditions. Discussion Consistent with cumulative inequality theory, these findings reveal that experiencing multiple types of misfortune during childhood decreases the likelihood of remaining disease free in later life, but engaging in health behaviors, such as physical activity, can help to ameliorate some of the noxious effects of early misfortune. Childhood disadvantage, Cumulative inequality theory, Successful aging Nearly 86% of Americans aged 65 and older live with at least one chronic disease, whereas more than 60% have two or more diseases (Ward, Schiller, & Goodman, 2014). Many chronic diseases are also the leading causes of death for older Americans: Two thirds of their deaths are due to heart disease, cancer, chronic lower respiratory disease, stroke, Alzheimer’s disease, and diabetes (Heron, 2016; Xu, Murphy, Kochanek, & Arias, 2016). Although chronic disease in later life is widespread and studied extensively, relatively few investigators examine how people are able to avoid any of the chronic “killer” diseases during later life. Most gerontologists do not consider being disease free a sine qua non for “successful aging,” but avoiding the leading causes of death greatly improves one’s chances for exceptional longevity (Evert, Lawler, Bogan, & Perls, 2003; Rowe & Kahn, 1987). Studying the antecedents of remaining free of chronic diseases is a logical way to identify risks that are common to multiple diseases and fundamental factors that promote health and well-being. Epidemiology’s search for risk factors common across life-threatening diseases also has led to greater consideration of experiences from childhood. Although hundreds of studies have identified adult risk factors of various chronic diseases, a now well-established body of research has unveiled the influence of early-life risk factors on adult diseases, ranging from lung cancer (Brown et al., 2010) to sarcopenia (Sayer, Syddall, Gilbody, Dennison, & Cooper, 2004). Kuh and Ben-Shlomo (2016) describe the shift in epidemiology from an adult life style model to one that also accounts for life course influences. In this article, we examine life course influences on the likelihood of being disease free in later life. We ask whether childhood misfortune reduces the likelihood that older adults will be able to avoid major, life-threatening diseases. Although reaching later life free of chronic disease is both rare and remarkable, to our knowledge, there is only one study that systematically examines early-life influences on this outcome (Schafer & Ferraro, 2012). The present study examines the enduring effects of childhood misfortune on health in later life and is distinct from parallel studies in several ways outlined in the following section. Childhood Misfortune and Avoiding Disease Dozens of studies during the past two decades reveal how early-life insults elevate the risk of specific diseases during adulthood. There is compelling evidence that childhood misfortune raises the risk of heart disease (Morton, Mustillo, & Ferraro, 2014; O’Rand & Hamil-Luker, 2005), diabetes (Thomas, Hyppönen, & Power, 2008), chronic lung disease (Svanes et al., 2010), and cancer (Brown et al., 2010; Kemp, Ferraro, Morton, & Mustillo, 2018) directly, through stress processes, and indirectly, through poor environments and health behaviors. These findings have been valuable for identifying life course influences on conditions often identified as “age-related diseases.” Although these diseases are prevalent among older people, the etiology of these and other conditions begins much earlier. Investigating individual diseases is fruitful for advancing life course epidemiology, but it is plausible that some risk factors are pluripotent. For example, Blackwell, Hayward, and Crimmins (2001) found that poor childhood health is associated with cancer, lung disease, cardiovascular conditions, and arthritis, even after adjusting for life course socioeconomic status. Furthermore, research examining disease accumulation reveals that early-life misfortune raises the risk of comorbidity in later life (Ferraro, Schafer, & Wilkinson, 2016; Pavela & Latham, 2016). Failure to account for related diseases in studies of single diseases, however, may lead to misdirected conclusions. For example, a person may be cancer-free giving the appearance of relative health, yet suffer from other, equally life-threatening conditions. Although studying the early origins of disease-free aging is a promising line of inquiry to identify factors that promote longevity and overall health, it is surprising that so few life course studies have examined being disease free as an outcome. The study that directly motivated the present investigation used data from the Midlife Development in the United States Study (MIDUS). Schafer and Ferraro (2012) found that a simple sum of childhood misfortune indicators, ranging from physical and mental abuse to financial strain, decreased the likelihood of being disease free in adulthood, both at baseline and 10 years later (1995 and 2015). Although being disease free at the follow-up survey was infrequent (<7%), the influence of early misfortune did not wane. The findings from this study are important but limited in several ways. First, given its focus on midlife, the MIDUS sample has a baseline age ceiling of 74 years, whereas the Health and Retirement Study (HRS) does not, enabling us to examine the highly selective situation where septuagenarians and octogenarians are disease free. Second, although Schafer and Ferraro (2012) examined disease prevalence and incidence across two waves of data, separated by 8 years, we use five waves of HRS data collected at 2-year intervals to permit a more fine-grained analysis of transitions involving disease incidence. This longitudinal approach enables us to examine trends and the risk of getting a major disease over time. Third, the authors incorporated an extensive list of conditions in their conceptualization of disease free, but some of these conditions are symptoms and often temporary (e.g., teeth problems, constipation). Instead, we focus on leading causes of death. Fourth, rather than adding all types of misfortune into a single variable, we examine specific domains of childhood misfortune to identify which types of misfortune are most detrimental in a parsimonious way. Finally, an oversampling of minority respondents in the HRS allows for in-depth investigation into whether the relationship between childhood misfortune and remaining disease free in later life varies by race and ethnicity. Black adults have higher prevalence of several diseases considered to be leading causes of death and higher mortality rates than White adults (National Center for Health Statistics, 2016). Evidence regarding health disparities between Hispanic and non-Hispanic Americans is less consistent but merits study nonetheless. Theoretical Background Cumulative inequality (CI) theory informs this study of the early origins of adult health (Schafer, Ferraro, & Mustillo, 2011), integrating elements of the life course perspective (Elder, 1998), cumulative disadvantage theory (Dannefer, 2003), and stress process theories (Pearlin, 1989). Instead of focusing on proximal factors such as adult socioeconomic status and health behaviors, the theory emphasizes how social stratification manifests over the life course, especially how early disadvantage shapes risks and resources over time (Ferraro & Shippee, 2009; Ferraro et al., 2016). Early disadvantage may alter health trajectories in several ways, and four elements of the theory are germane for the present study. First, recognizing the role of family lineage, CI theory draws on the life course perspective and “linked lives” to argue that childhood conditions, including inequality across families of origin, are integral to stratification processes (Ferraro, Shippee, & Schafer, 2009). Second, the theory emphasizes how inequality diffuses across life domains—“disadvantage in one domain (e.g., family life) may spill over to other domains (e.g., mental and physical health)” (Ferraro & Shippee, 2009, p. 339), resulting in stress proliferation (Pearlin, 1989). Rather than add all indicators of misfortune into one variable, a formerly common practice, the theory calls for examining multiple domains of misfortune, when such measures are available. Third, a novel contribution of CI theory is that it emphasizes the plasticity of human development and the ways in which disadvantage does not accumulate. Although experiencing misfortune in childhood can be a strong predictor of health problems in later life by influencing continued exposure to stressors and behaviors unconducive to adult health (e.g., smoking), human agency and resource activation can mitigate the unfavorable consequences of early misfortune (Ferraro et al., 2009, 2016). The theory cautions about “runaway cumulative disadvantage” and calls for identifying factors that can avert the presumed health consequences. Finally, cumulative inequality often leads to premature mortality among those most exposed to adversity, creating compositional change in a population that can masquerade as decreasing inequality (Ferraro & Shippee, 2009). It may be more difficult to detect the potential influence of early misfortune on persons belonging to racial and ethnic minorities because of selective mortality. Thus, we systematically examine racial and ethnic differences and mortality selection in our analyses. Research Questions The present study is innovative in several ways and addresses the following research questions: Does childhood misfortune decrease the likelihood of older adults being disease free? We address this question at the initial survey and across 8 years of longitudinal follow-ups, anticipating that childhood misfortune is a threat to disease-free aging. Do adult resources and lifestyle factors moderate the relationship between childhood misfortune and being disease free in later life? We anticipate that the relationship between childhood misfortune and being disease free will be moderated by adult resources (e.g., education), and lifestyle factors (e.g., smoking). Method Sample Data were obtained from the HRS—a multistage, probability study of adults aged 51 years and older, with an oversampling of Black and Hispanic adults and Florida residents. Waves 7–11 (2004–2012) were used for this research because a detailed battery of childhood misfortune indicators were first measured on the full sample in 2004. For this study, respondents were excluded if they were age ineligible, deceased as of 2004, or entered the study in 2010. In addition, those who had proxy responses or fell 2 SD below the average cognition score were excluded to preserve the validity of the responses, particularly for the retrospective childhood misfortune variables. These restrictions resulted in a final, analytic sample of 13,921 adults. Disease Free There are different ways to define being disease free. One could give particular attention to chronic diseases and life-threatening conditions or could include every potential health condition experienced throughout life. However, the majority of adults aged 50 and older experience an array of minor health symptoms and conditions, yet continue to live long and fairly healthy lives until faced with one of the major causes of death. We focus on those diseases that are both chronic and life threatening. As of 2014, the top chronic disease-related causes of death for those aged 50 and older included heart disease, cancer, chronic lower respiratory disease, cerebrovascular disease, diabetes, Alzheimer’s disease, kidney disease, and liver disease (National Center for Health Statistics, 2015); we include the aforementioned diseases to create our indicators of being disease free at 2004 (prevalence) and developing disease (incidence) from 2004 to 2012. Respondents were asked if a doctor had ever told them that they have diabetes, lung disease, stroke, heart problems, hypertension, or cancer (excluding skin). Respondents also were asked if they had ever been diagnosed with a “memory-related disease.” In 2010 and 2012, this question was dropped in favor of two questions addressing Alzheimer’s disease and dementia separately. Respondents also were asked if they had other diseases or medical conditions not addressed. We used these responses to gather information about chronic lower respiratory disease, liver disease, and kidney disease. We also included decedents from any of the nine aforementioned diseases using the National Death Index linkage. As specified in axiom 5 of CI theory, failure to account for selective mortality may underestimate disease incidence for the most virulent conditions (i.e., diagnosed with a disease and died between interviews). For the baseline analyses of prevalence, a binary variable was created: one or more diseases (=0) and disease free (=1). Respondents who reported a memory disease prior to 2010 or Alzheimer’s disease or dementia during or after 2010 were coded as 0. Respondents reporting one or more diseases were coded as 0, regardless of missingness on other disease variables. However, respondents reporting no diseases were coded as 1 only if they had valid answers for at least half (four of seven), physician-diagnosed diseases. For the longitudinal analyses of incidence, we inverted the coding to be consistent with the widely used concept of disease incidence (0 = disease free; 1 = one or more diseases between 2004 and 2012). Childhood Misfortune Childhood misfortune was operationalized differently in the literature. Instead of adding indicators of misfortune into an overall construct as others have done, we draw from CI theory to identify which domains of childhood misfortune are consequential to remaining disease free (proposition 2.b: Ferraro & Shippee, 2009). In preliminary inquiry, we used both exploratory and confirmatory factor analyses to finalize the indicators used to specify the domains, eliminating two potential indicators. We categorized misfortunes experienced before the age of 18 into five domains: socioeconomic status (SES), risky parental behaviors, infectious diseases, impairments, and risky adolescent behaviors, using 20 retrospective questions collected between 2004 and 2008. Socioeconomic status (SES) misfortune consisted of five indicators: reporting poor or fair finances in relation to others, moving due to finances, having a father with an unskilled labor occupation, mother having less than a high school education, and/or father having less than a high school education. Risky parental behavior was assessed by having a parent who smoked, abused substances, and/or was physically abusive. Infectious disease included measles, mumps, and/or chicken pox. The presence of a childhood impairment included head injury, disability for more than 6 months, learning problems, visual impairment, and/or speech impairment. Risky adolescent behavior was assessed by having trouble with police, substance abuse, depressive symptoms, and/or other psychological issues. Each indicator was coded dichotomously and summed to create a count of misfortune within each domain, top coded at 2 (0, 1, and 2+) to examine a threshold effect of accumulating two or more misfortunes as compared to none. Several domain coding schemes were examined (e.g., dichotomous, top coded at 3+); however, differentiating 0, 1, and 2+ indicators in each domain yielded the most useful information. Multiple imputation by Chained Equations (MICE) was used to handle missing data among childhood misfortune domains with missing exceeding 5% (SES = 30.19%; risky parental behavior = 20.69%; risky adolescent behavior = 20.21%); observations on other variables that initially had missing were excluded prior to analysis (Von Hippel, 2007). All variables included in the analysis were used to impute 20 datasets using the “mi impute chained” command in Stata with the appropriate link functions for each variable. Adult Resources and Lifestyle Factors Although a challenging childhood can predict health problems for older adults, not everyone who experiences childhood misfortune suffers from poor health in later life. Childhood circumstances can have lasting impacts on adult health by influencing continued exposure to stressors and behaviors unconducive to adult health. Yet, resilience through human agency and resource activation can help to protect against some of these health risks (Ferraro et al., 2009). For example, smoking cessation has been associated with substantial reduction in risk of all-cause mortality among those with coronary heart disease (Critchley & Capewell, 2003). Thus, we include both adult resources and lifestyle factors that may modify health risks. We examined four indicators of adult resources: marital status, education, wealth, and health insurance status. Marital status was coded as never married, married (reference), widowed, and divorced/separated. Education was measured in years and top coded at 17+. Wealth is presented in tens of thousands of dollars and cube rooted to correct for skewness. A four-category indicator of health insurance status was created: no insurance (reference), Medicaid only, Medicare only, and private insurance (private only and/or Medigap plans). Five indicators of adult health characteristics were included in the models: depressive symptoms, smoking status, heavy drinking, body mass index (BMI), and physical activity. Depressive symptoms were measured using an eight-item Center for Epidemiological Studies–Depression (CES-D) scale. Smoking was assessed by a pack-years variable of cigarette smoking (one pack-year is equivalent to smoking 20 cigarettes daily for 1 year) where never smokers were coded as 0. A heavy drinker was dichotomously coded as having an average of five or more drinks per day for men and four or more drinks per day for women (Dawson, 2011). BMI was based on self-reports and categorized into underweight or normal weight (BMI < 25), overweight (25 ≥ BMI < 30), and obese (BMI ≥ 30). A self-reported physical activity scale accounts for both frequency and intensity. Respondents were asked the frequency in which they participated in mild, moderate, and vigorous activity; possible answers included never/rarely (=0), one to three times per month (=1), once per week (=2), two or more times per week (=3), and every day (=4). Drawing from other researchers, the physical activity scale was created by weighting the type of physical activity by intensity (mild = 1.2, moderate = 1.4, and vigorous = 1.8) based on metabolic equivalent recommendations (Latham & Williams, 2015; Umstattd Meyer, Janke, & Beaujean, 2014). Possible scores ranged from 0 (no physical activity) to 17.6 (mild, moderate, and vigorous physical activity daily). Demographics Finally, models adjusted for demographic characteristics reported at baseline including age (in years), gender (female = 1), and race/ethnicity (non-Hispanic white = 0, non-Hispanic black, and Hispanic). Other races were excluded as the number of cases was insufficient for meaningful comparisons (n = 303). Analysis The analysis was conducted in three parts. First, logistic regression was used to estimate the prevalence of being disease free in 2004 in the full sample. Second, Cox proportional hazards models were estimated, using time-on-study with wave (every 2 years) as our time scale, to investigate disease incidence (reporting one or more diseases) between 2004 and 2012 among a more selective, healthy sample—those who were disease free in the prevalence model (2004). The proportional hazards assumption was tested for potential time-varying covariates with no violation encountered, and ties were handled with the Breslow’s (1974) method. Finally, a series of theoretically informed interactions were tested to assess the potential moderating effects of adult risks and resources. Models utilized 2004 HRS weights and accounted for the survey design. Analyses were conducted using Stata ST/SE 14.1. Additional information is presented in Supplementary Material. Results Sample Characteristics The weighted descriptive statistics for focal variables are presented in Table 1 (see Supplementary Table S1 for descriptive statistics for all variables in the final analyses). At baseline (2004), approximately 35.30% of the sample was disease free. By 2012, only 15.92% of the sample remained disease free. For the childhood misfortune variables, respondents were most likely to report infectious disease (1.74; SD = 0.57) and socioeconomic misfortunes (1.45; SD = 0.78). Table 1. Select Descriptive Statistics for the Analytical Sample (N = 13,921)a Variable Range Percent Mean (SE) Disease free (prevalence)b  2004 0, 1 35.30  2006 0, 1 28.02  2008 0, 1 23.02  2010 0, 1 18.81  2012 0, 1 15.92 Childhood misfortune  Socioeconomic 0–2 1.45 (0.78)  Risky parental behavior 0–2 0.90 (0.67)  Infectious disease 0–2 1.74 (0.57)  Impairments 0–2 0.22 (0.48)  Risky adolescent behavior 0–2 0.29 (0.54) Variable Range Percent Mean (SE) Disease free (prevalence)b  2004 0, 1 35.30  2006 0, 1 28.02  2008 0, 1 23.02  2010 0, 1 18.81  2012 0, 1 15.92 Childhood misfortune  Socioeconomic 0–2 1.45 (0.78)  Risky parental behavior 0–2 0.90 (0.67)  Infectious disease 0–2 1.74 (0.57)  Impairments 0–2 0.22 (0.48)  Risky adolescent behavior 0–2 0.29 (0.54) Note:aWeighted values presented. bDiseases include cancer, heart disease, hypertension, chronic lower respiratory disease, cerebrovascular disease, diabetes, liver disease, memory disease, and kidney disease. View Large Table 1. Select Descriptive Statistics for the Analytical Sample (N = 13,921)a Variable Range Percent Mean (SE) Disease free (prevalence)b  2004 0, 1 35.30  2006 0, 1 28.02  2008 0, 1 23.02  2010 0, 1 18.81  2012 0, 1 15.92 Childhood misfortune  Socioeconomic 0–2 1.45 (0.78)  Risky parental behavior 0–2 0.90 (0.67)  Infectious disease 0–2 1.74 (0.57)  Impairments 0–2 0.22 (0.48)  Risky adolescent behavior 0–2 0.29 (0.54) Variable Range Percent Mean (SE) Disease free (prevalence)b  2004 0, 1 35.30  2006 0, 1 28.02  2008 0, 1 23.02  2010 0, 1 18.81  2012 0, 1 15.92 Childhood misfortune  Socioeconomic 0–2 1.45 (0.78)  Risky parental behavior 0–2 0.90 (0.67)  Infectious disease 0–2 1.74 (0.57)  Impairments 0–2 0.22 (0.48)  Risky adolescent behavior 0–2 0.29 (0.54) Note:aWeighted values presented. bDiseases include cancer, heart disease, hypertension, chronic lower respiratory disease, cerebrovascular disease, diabetes, liver disease, memory disease, and kidney disease. View Large Disease-Free Prevalence Table 2 presents the logistic regression models predicting being disease free in 2004. When adjusting for age, sex, and race in Model 1, respondents who experienced 2+ socioeconomic misfortunes during childhood had 31.2% lower odds of being disease free in 2004, relative to those who experienced no misfortune. Furthermore, respondents who experienced one or 2+ risky parental behaviors (odds ratios [ORs] = 0.882, 0.737, respectively), impairments (ORs = 0.863, 0.712), or risky adolescent behaviors (ORs = 0.815, 0.706) also had lower odds of being disease free in 2004 than those who experienced none. As expected, the odds of being disease free decreased with age, and Black adults had about half the odds of being disease free as White adults. We present Model 1 to highlight the relationships between types of childhood misfortune and disease-free status without adjusting for adult resources but urge caution when interpreting the results. Whereas Model 1 omits variables related to being disease free, and omitted variables bias the ORs for the variables that are included in the model (Mood, 2010), we interpret the ORs recognizing that the estimates are probably biased downward. Estimates from Model 2 will be closer to the true effects. Table 2. Logistic Regression Predicting Prevalence of Being Disease Free at Baseline (2004)a Model 1 Model 2 OR 95% CI OR 95% CI Childhood misfortuneb  Socioeconomic (=1) 0.865 0.717–1.043 0.974 0.784–1.209  Socioeconomic (=2+) 0.688† 0.577–0.820 0.870 0.707–1.069  Risky parent behavior (=1) 0.882* 0.794–0.979 0.908 0.805–1.023  Risky parent behavior (=2+) 0.737† 0.635–0.855 0.791‡ 0.667–0.937  Infectious disease (=1) 1.111 0.866–1.427 0.950 0.714–1.264  Infectious disease (=2+) 0.923 0.752–1.132 0.806 0.644–1.009  Impairment (=1) 0.863* 0.772–0.965 0.904 0.787–1.039  Impairment (=2+) 0.712‡ 0.564–0.898 0.843 0.637–1.115  Risky adolescent behavior (=1) 0.815‡ 0.704–0.943 0.951 0.804–1.127  Risky adolescent behavior (=2+) 0.706* 0.528–0.944 0.935 0.681–1.282 Demographics  Age (years at baseline) 0.945† 0.940–0.950 0.936† 0.930–0.942  Femalec 1.053 0.957–1.159 1.085 0.965–1.219  Blackd 0.504† 0.418–0.608 0.682† 0.558–0.833  Hispanicd 0.915 0.780–1.072 1.120 0.901–1.394 Adult resources and lifestyle factors  Divorcede 1.091 0.930–1.280  Widowede 1.229‡ 1.077–1.403  Never marriede 1.165 0.888–1.529  Education (years) 1.001 0.979–1.022  Wealth (∛) 1.004† 1.002–1.005  Medicaidf 0.424† 0.309–0.581  Medicare/VAf 0.757 0.571–1.003  Private/Medigapf 0.739‡ 0.578–0.944  Depressive symptoms 0.884† 0.851–0.918  Smoking (pack-years) 0.994† 0.991–0.996  Heavy drinkerg 1.250 0.914–1.709  25 ≤ BMI < 30h 0.586† 0.523–0.658  BMI ≥ 30h 0.335† 0.289–0.389  Physical activity 1.046† 1.032–1.061 Constant 32.389† 20.877–50.249 74.139† 40.656–135.198 F 51.51† 67.98† N 13,824 12,023 Model 1 Model 2 OR 95% CI OR 95% CI Childhood misfortuneb  Socioeconomic (=1) 0.865 0.717–1.043 0.974 0.784–1.209  Socioeconomic (=2+) 0.688† 0.577–0.820 0.870 0.707–1.069  Risky parent behavior (=1) 0.882* 0.794–0.979 0.908 0.805–1.023  Risky parent behavior (=2+) 0.737† 0.635–0.855 0.791‡ 0.667–0.937  Infectious disease (=1) 1.111 0.866–1.427 0.950 0.714–1.264  Infectious disease (=2+) 0.923 0.752–1.132 0.806 0.644–1.009  Impairment (=1) 0.863* 0.772–0.965 0.904 0.787–1.039  Impairment (=2+) 0.712‡ 0.564–0.898 0.843 0.637–1.115  Risky adolescent behavior (=1) 0.815‡ 0.704–0.943 0.951 0.804–1.127  Risky adolescent behavior (=2+) 0.706* 0.528–0.944 0.935 0.681–1.282 Demographics  Age (years at baseline) 0.945† 0.940–0.950 0.936† 0.930–0.942  Femalec 1.053 0.957–1.159 1.085 0.965–1.219  Blackd 0.504† 0.418–0.608 0.682† 0.558–0.833  Hispanicd 0.915 0.780–1.072 1.120 0.901–1.394 Adult resources and lifestyle factors  Divorcede 1.091 0.930–1.280  Widowede 1.229‡ 1.077–1.403  Never marriede 1.165 0.888–1.529  Education (years) 1.001 0.979–1.022  Wealth (∛) 1.004† 1.002–1.005  Medicaidf 0.424† 0.309–0.581  Medicare/VAf 0.757 0.571–1.003  Private/Medigapf 0.739‡ 0.578–0.944  Depressive symptoms 0.884† 0.851–0.918  Smoking (pack-years) 0.994† 0.991–0.996  Heavy drinkerg 1.250 0.914–1.709  25 ≤ BMI < 30h 0.586† 0.523–0.658  BMI ≥ 30h 0.335† 0.289–0.389  Physical activity 1.046† 1.032–1.061 Constant 32.389† 20.877–50.249 74.139† 40.656–135.198 F 51.51† 67.98† N 13,824 12,023 Note: BMI = body mass index; CI = confidence interval; OR = odds ratio. aAnalyses weighted to adjust for clustering. bReference group for each of the five domains is no misfortune. cReference group is male. dReference group is non-Hispanic White. eReference group is married. fReference group is no insurance. gReference group is non-heavy drinker. hReference group is BMI < 25. *p < .05; ‡p < .01, †p < .001. View Large Table 2. Logistic Regression Predicting Prevalence of Being Disease Free at Baseline (2004)a Model 1 Model 2 OR 95% CI OR 95% CI Childhood misfortuneb  Socioeconomic (=1) 0.865 0.717–1.043 0.974 0.784–1.209  Socioeconomic (=2+) 0.688† 0.577–0.820 0.870 0.707–1.069  Risky parent behavior (=1) 0.882* 0.794–0.979 0.908 0.805–1.023  Risky parent behavior (=2+) 0.737† 0.635–0.855 0.791‡ 0.667–0.937  Infectious disease (=1) 1.111 0.866–1.427 0.950 0.714–1.264  Infectious disease (=2+) 0.923 0.752–1.132 0.806 0.644–1.009  Impairment (=1) 0.863* 0.772–0.965 0.904 0.787–1.039  Impairment (=2+) 0.712‡ 0.564–0.898 0.843 0.637–1.115  Risky adolescent behavior (=1) 0.815‡ 0.704–0.943 0.951 0.804–1.127  Risky adolescent behavior (=2+) 0.706* 0.528–0.944 0.935 0.681–1.282 Demographics  Age (years at baseline) 0.945† 0.940–0.950 0.936† 0.930–0.942  Femalec 1.053 0.957–1.159 1.085 0.965–1.219  Blackd 0.504† 0.418–0.608 0.682† 0.558–0.833  Hispanicd 0.915 0.780–1.072 1.120 0.901–1.394 Adult resources and lifestyle factors  Divorcede 1.091 0.930–1.280  Widowede 1.229‡ 1.077–1.403  Never marriede 1.165 0.888–1.529  Education (years) 1.001 0.979–1.022  Wealth (∛) 1.004† 1.002–1.005  Medicaidf 0.424† 0.309–0.581  Medicare/VAf 0.757 0.571–1.003  Private/Medigapf 0.739‡ 0.578–0.944  Depressive symptoms 0.884† 0.851–0.918  Smoking (pack-years) 0.994† 0.991–0.996  Heavy drinkerg 1.250 0.914–1.709  25 ≤ BMI < 30h 0.586† 0.523–0.658  BMI ≥ 30h 0.335† 0.289–0.389  Physical activity 1.046† 1.032–1.061 Constant 32.389† 20.877–50.249 74.139† 40.656–135.198 F 51.51† 67.98† N 13,824 12,023 Model 1 Model 2 OR 95% CI OR 95% CI Childhood misfortuneb  Socioeconomic (=1) 0.865 0.717–1.043 0.974 0.784–1.209  Socioeconomic (=2+) 0.688† 0.577–0.820 0.870 0.707–1.069  Risky parent behavior (=1) 0.882* 0.794–0.979 0.908 0.805–1.023  Risky parent behavior (=2+) 0.737† 0.635–0.855 0.791‡ 0.667–0.937  Infectious disease (=1) 1.111 0.866–1.427 0.950 0.714–1.264  Infectious disease (=2+) 0.923 0.752–1.132 0.806 0.644–1.009  Impairment (=1) 0.863* 0.772–0.965 0.904 0.787–1.039  Impairment (=2+) 0.712‡ 0.564–0.898 0.843 0.637–1.115  Risky adolescent behavior (=1) 0.815‡ 0.704–0.943 0.951 0.804–1.127  Risky adolescent behavior (=2+) 0.706* 0.528–0.944 0.935 0.681–1.282 Demographics  Age (years at baseline) 0.945† 0.940–0.950 0.936† 0.930–0.942  Femalec 1.053 0.957–1.159 1.085 0.965–1.219  Blackd 0.504† 0.418–0.608 0.682† 0.558–0.833  Hispanicd 0.915 0.780–1.072 1.120 0.901–1.394 Adult resources and lifestyle factors  Divorcede 1.091 0.930–1.280  Widowede 1.229‡ 1.077–1.403  Never marriede 1.165 0.888–1.529  Education (years) 1.001 0.979–1.022  Wealth (∛) 1.004† 1.002–1.005  Medicaidf 0.424† 0.309–0.581  Medicare/VAf 0.757 0.571–1.003  Private/Medigapf 0.739‡ 0.578–0.944  Depressive symptoms 0.884† 0.851–0.918  Smoking (pack-years) 0.994† 0.991–0.996  Heavy drinkerg 1.250 0.914–1.709  25 ≤ BMI < 30h 0.586† 0.523–0.658  BMI ≥ 30h 0.335† 0.289–0.389  Physical activity 1.046† 1.032–1.061 Constant 32.389† 20.877–50.249 74.139† 40.656–135.198 F 51.51† 67.98† N 13,824 12,023 Note: BMI = body mass index; CI = confidence interval; OR = odds ratio. aAnalyses weighted to adjust for clustering. bReference group for each of the five domains is no misfortune. cReference group is male. dReference group is non-Hispanic White. eReference group is married. fReference group is no insurance. gReference group is non-heavy drinker. hReference group is BMI < 25. *p < .05; ‡p < .01, †p < .001. View Large When further adjusting for adult resources and lifestyle factors in Model 2, the magnitude of several relationships between childhood misfortune and being disease free is reduced; the ORs for socioeconomic, impairments, and risky adolescent behavior misfortunes are no longer significant. However, experiencing 2+ risky parental behaviors in childhood was still associated with being disease free in 2004 (OR = 0.791). Age (OR = 0.936) and race (Black; OR = 0.682) were associated with lower odds of being disease free. We used linear probability models to examine potential interactions between race/ethnicity and each of the childhood misfortune domains (not shown) but did not find any evidence that childhood misfortune differentially affects the odds of being disease free by race. Among adult resources, wealth (OR = 1.004) and being widowed as opposed to married (OR = 1.229) increased one’s odds of being disease free. Conversely, older adults with Medicaid exclusively (OR = 0.424) and private and/or Medigap insurance plans (OR = 0.739) were less likely to report being disease free compared with those with no insurance. Depressive symptoms (OR = 0.884), smoking (OR = 0.994), and being overweight (OR = 0.586) or obese (OR = 0.335) decrease the odds, whereas physical activity increases those odds of being disease free (OR = 1.046). Disease Incidence Table 3 presents the longitudinal results for disease incidence among those who were disease free in 2004. Cox proportional hazard models are used to predict disease onset by 2012. When adjusting for demographic factors only (Model 1), respondents who experienced 2+ SES or risky adolescent misfortunes in childhood, had 17.1% and 34.1%, respectively, higher hazard of developing disease between 2004 and 2012 compared with those who experienced no misfortune within those domains. In addition, age (hazard ratio [HR] = 1.014) increased the hazard of getting one or more diseases, whereas being female lowered the hazard by 15.6%. Finally, Black and Hispanic adults had 26.3% and 25.4% greater hazard of developing disease than White adults. Table 3. Cox Proportional Hazards Predicting Disease Incidence (2004–2012)a Model 1 Model 2 Model 3 HR 95% CI HR 95% CI HR 95% CI Childhood misfortuneb  Socioeconomic (=1) 1.105 0.946–1.292 1.118 0.946–1.323 1.114 0.940–1.322  Socioeconomic (=2+) 1.171* 1.017–1.349 1.110 0.932–1.322 1.107 0.928–1.322  Risky parent b. (=1) 1.046 0.935–1.171 1.057 0.939–1.191 0.977 0.852–1.121  Risky parent b. (=2+) 1.063 0.927–1.217 1.072 0.930–1.235 1.022 0.849–1.230  Infectious disease (=1) 1.081 0.894–1.306 1.164 0.936–1.447 1.162 0.936–1.443  Infectious disease (=2+) 1.048 0.881–1.247 1.075 0.885–1.306 1.077 0.888–1.307  Impairment (=1) 1.015 0.896–1.151 0.962 0.840–1.101 0.968 0.846–1.107  Impairment (=2+) 1.084 0.865–1.358 1.134 0.880–1.462 1.147 0.888–1.480  Risky adolescent b. (=1) 1.058 0.931–1.204 1.049 0.910–1.209 1.049 0.912–1.206  Risky adolescent b. (=2+) 1.341* 1.033–1.740 1.265 0.948–1.687 1.239 0.924–1.663 Demographics  Age (years at baseline) 1.014† 1.009–1.019 1.017† 1.011–1.022 1.017† 1.011–1.022  Female 0.844‡ 0.761–0.935 0.873* 0.785–0.969 0.873* 0.786–0.971  Race/ethnicity (White = rf)   Black 1.263‡ 1.107–1.440 1.137 0.981–1.318 1.132 0.975–1.314   Hispanic 1.254‡ 1.089–1.443 1.166 0.990–1.375 1.179* 1.005–1.384 Adult resources and lifestyle  Marital status (married = rf)   Divorced 0.925 0.812–1.054 0.929 0.815–1.060   Widowed 1.011 0.873–1.170 1.001 0.866–1.158   Never married 1.097 0.906–1.328 1.102 0.909–1.337  Education (years) 1.012 0.992–1.033 1.014 0.994–1.034  Wealth (∛) 0.998* 0.997–1.000 0.998* 0.997–1.000  Insurance (none = rf)   Medicaid 1.152 0.839–1.581 1.203 0.891–1.623   Medicare/VA 1.096 0.876–1.371 1.093 0.874–1.366   Private/Medigap 1.022 0.858–1.218 1.023 0.857–1.222  Depressive symptoms 1.059† 1.034–1.084 1.059 1.034–1.084  Smoking (pack-years) 1.000 0.998–1.002 0.996* 0.992–1.000  Heavy drinker 1.365* 1.078–1.729 1.371‡ 1.089–1.727 BMI (<25 = rf)   25 ≤ BMI < 30 1.205† 1.096–1.324 1.206 1.097–1.325   BMI ≥ 30 1.611† 1.441–1.801 1.619 1.448–1.810  Physical activity 0.982* 0.966–0.998 0.982* 0.966–0.998 1 Risky parent × smoking 1.007‡ 1.002–1.012 2+ Risky parent × smoking 1.005 0.998–1.012 F 6.96† 16.36† 18.84† N 4,101 3,594 3,594 Model 1 Model 2 Model 3 HR 95% CI HR 95% CI HR 95% CI Childhood misfortuneb  Socioeconomic (=1) 1.105 0.946–1.292 1.118 0.946–1.323 1.114 0.940–1.322  Socioeconomic (=2+) 1.171* 1.017–1.349 1.110 0.932–1.322 1.107 0.928–1.322  Risky parent b. (=1) 1.046 0.935–1.171 1.057 0.939–1.191 0.977 0.852–1.121  Risky parent b. (=2+) 1.063 0.927–1.217 1.072 0.930–1.235 1.022 0.849–1.230  Infectious disease (=1) 1.081 0.894–1.306 1.164 0.936–1.447 1.162 0.936–1.443  Infectious disease (=2+) 1.048 0.881–1.247 1.075 0.885–1.306 1.077 0.888–1.307  Impairment (=1) 1.015 0.896–1.151 0.962 0.840–1.101 0.968 0.846–1.107  Impairment (=2+) 1.084 0.865–1.358 1.134 0.880–1.462 1.147 0.888–1.480  Risky adolescent b. (=1) 1.058 0.931–1.204 1.049 0.910–1.209 1.049 0.912–1.206  Risky adolescent b. (=2+) 1.341* 1.033–1.740 1.265 0.948–1.687 1.239 0.924–1.663 Demographics  Age (years at baseline) 1.014† 1.009–1.019 1.017† 1.011–1.022 1.017† 1.011–1.022  Female 0.844‡ 0.761–0.935 0.873* 0.785–0.969 0.873* 0.786–0.971  Race/ethnicity (White = rf)   Black 1.263‡ 1.107–1.440 1.137 0.981–1.318 1.132 0.975–1.314   Hispanic 1.254‡ 1.089–1.443 1.166 0.990–1.375 1.179* 1.005–1.384 Adult resources and lifestyle  Marital status (married = rf)   Divorced 0.925 0.812–1.054 0.929 0.815–1.060   Widowed 1.011 0.873–1.170 1.001 0.866–1.158   Never married 1.097 0.906–1.328 1.102 0.909–1.337  Education (years) 1.012 0.992–1.033 1.014 0.994–1.034  Wealth (∛) 0.998* 0.997–1.000 0.998* 0.997–1.000  Insurance (none = rf)   Medicaid 1.152 0.839–1.581 1.203 0.891–1.623   Medicare/VA 1.096 0.876–1.371 1.093 0.874–1.366   Private/Medigap 1.022 0.858–1.218 1.023 0.857–1.222  Depressive symptoms 1.059† 1.034–1.084 1.059 1.034–1.084  Smoking (pack-years) 1.000 0.998–1.002 0.996* 0.992–1.000  Heavy drinker 1.365* 1.078–1.729 1.371‡ 1.089–1.727 BMI (<25 = rf)   25 ≤ BMI < 30 1.205† 1.096–1.324 1.206 1.097–1.325   BMI ≥ 30 1.611† 1.441–1.801 1.619 1.448–1.810  Physical activity 0.982* 0.966–0.998 0.982* 0.966–0.998 1 Risky parent × smoking 1.007‡ 1.002–1.012 2+ Risky parent × smoking 1.005 0.998–1.012 F 6.96† 16.36† 18.84† N 4,101 3,594 3,594 Note: b. = behavior; BMI = body mass index; CI = confidence interval; HR = hazard ratio; rf = reference group. aAnalyses weighted to adjust for clustering. bReference group is no misfortune. *p < .05; ‡p < .01, †p < .001. View Large Table 3. Cox Proportional Hazards Predicting Disease Incidence (2004–2012)a Model 1 Model 2 Model 3 HR 95% CI HR 95% CI HR 95% CI Childhood misfortuneb  Socioeconomic (=1) 1.105 0.946–1.292 1.118 0.946–1.323 1.114 0.940–1.322  Socioeconomic (=2+) 1.171* 1.017–1.349 1.110 0.932–1.322 1.107 0.928–1.322  Risky parent b. (=1) 1.046 0.935–1.171 1.057 0.939–1.191 0.977 0.852–1.121  Risky parent b. (=2+) 1.063 0.927–1.217 1.072 0.930–1.235 1.022 0.849–1.230  Infectious disease (=1) 1.081 0.894–1.306 1.164 0.936–1.447 1.162 0.936–1.443  Infectious disease (=2+) 1.048 0.881–1.247 1.075 0.885–1.306 1.077 0.888–1.307  Impairment (=1) 1.015 0.896–1.151 0.962 0.840–1.101 0.968 0.846–1.107  Impairment (=2+) 1.084 0.865–1.358 1.134 0.880–1.462 1.147 0.888–1.480  Risky adolescent b. (=1) 1.058 0.931–1.204 1.049 0.910–1.209 1.049 0.912–1.206  Risky adolescent b. (=2+) 1.341* 1.033–1.740 1.265 0.948–1.687 1.239 0.924–1.663 Demographics  Age (years at baseline) 1.014† 1.009–1.019 1.017† 1.011–1.022 1.017† 1.011–1.022  Female 0.844‡ 0.761–0.935 0.873* 0.785–0.969 0.873* 0.786–0.971  Race/ethnicity (White = rf)   Black 1.263‡ 1.107–1.440 1.137 0.981–1.318 1.132 0.975–1.314   Hispanic 1.254‡ 1.089–1.443 1.166 0.990–1.375 1.179* 1.005–1.384 Adult resources and lifestyle  Marital status (married = rf)   Divorced 0.925 0.812–1.054 0.929 0.815–1.060   Widowed 1.011 0.873–1.170 1.001 0.866–1.158   Never married 1.097 0.906–1.328 1.102 0.909–1.337  Education (years) 1.012 0.992–1.033 1.014 0.994–1.034  Wealth (∛) 0.998* 0.997–1.000 0.998* 0.997–1.000  Insurance (none = rf)   Medicaid 1.152 0.839–1.581 1.203 0.891–1.623   Medicare/VA 1.096 0.876–1.371 1.093 0.874–1.366   Private/Medigap 1.022 0.858–1.218 1.023 0.857–1.222  Depressive symptoms 1.059† 1.034–1.084 1.059 1.034–1.084  Smoking (pack-years) 1.000 0.998–1.002 0.996* 0.992–1.000  Heavy drinker 1.365* 1.078–1.729 1.371‡ 1.089–1.727 BMI (<25 = rf)   25 ≤ BMI < 30 1.205† 1.096–1.324 1.206 1.097–1.325   BMI ≥ 30 1.611† 1.441–1.801 1.619 1.448–1.810  Physical activity 0.982* 0.966–0.998 0.982* 0.966–0.998 1 Risky parent × smoking 1.007‡ 1.002–1.012 2+ Risky parent × smoking 1.005 0.998–1.012 F 6.96† 16.36† 18.84† N 4,101 3,594 3,594 Model 1 Model 2 Model 3 HR 95% CI HR 95% CI HR 95% CI Childhood misfortuneb  Socioeconomic (=1) 1.105 0.946–1.292 1.118 0.946–1.323 1.114 0.940–1.322  Socioeconomic (=2+) 1.171* 1.017–1.349 1.110 0.932–1.322 1.107 0.928–1.322  Risky parent b. (=1) 1.046 0.935–1.171 1.057 0.939–1.191 0.977 0.852–1.121  Risky parent b. (=2+) 1.063 0.927–1.217 1.072 0.930–1.235 1.022 0.849–1.230  Infectious disease (=1) 1.081 0.894–1.306 1.164 0.936–1.447 1.162 0.936–1.443  Infectious disease (=2+) 1.048 0.881–1.247 1.075 0.885–1.306 1.077 0.888–1.307  Impairment (=1) 1.015 0.896–1.151 0.962 0.840–1.101 0.968 0.846–1.107  Impairment (=2+) 1.084 0.865–1.358 1.134 0.880–1.462 1.147 0.888–1.480  Risky adolescent b. (=1) 1.058 0.931–1.204 1.049 0.910–1.209 1.049 0.912–1.206  Risky adolescent b. (=2+) 1.341* 1.033–1.740 1.265 0.948–1.687 1.239 0.924–1.663 Demographics  Age (years at baseline) 1.014† 1.009–1.019 1.017† 1.011–1.022 1.017† 1.011–1.022  Female 0.844‡ 0.761–0.935 0.873* 0.785–0.969 0.873* 0.786–0.971  Race/ethnicity (White = rf)   Black 1.263‡ 1.107–1.440 1.137 0.981–1.318 1.132 0.975–1.314   Hispanic 1.254‡ 1.089–1.443 1.166 0.990–1.375 1.179* 1.005–1.384 Adult resources and lifestyle  Marital status (married = rf)   Divorced 0.925 0.812–1.054 0.929 0.815–1.060   Widowed 1.011 0.873–1.170 1.001 0.866–1.158   Never married 1.097 0.906–1.328 1.102 0.909–1.337  Education (years) 1.012 0.992–1.033 1.014 0.994–1.034  Wealth (∛) 0.998* 0.997–1.000 0.998* 0.997–1.000  Insurance (none = rf)   Medicaid 1.152 0.839–1.581 1.203 0.891–1.623   Medicare/VA 1.096 0.876–1.371 1.093 0.874–1.366   Private/Medigap 1.022 0.858–1.218 1.023 0.857–1.222  Depressive symptoms 1.059† 1.034–1.084 1.059 1.034–1.084  Smoking (pack-years) 1.000 0.998–1.002 0.996* 0.992–1.000  Heavy drinker 1.365* 1.078–1.729 1.371‡ 1.089–1.727 BMI (<25 = rf)   25 ≤ BMI < 30 1.205† 1.096–1.324 1.206 1.097–1.325   BMI ≥ 30 1.611† 1.441–1.801 1.619 1.448–1.810  Physical activity 0.982* 0.966–0.998 0.982* 0.966–0.998 1 Risky parent × smoking 1.007‡ 1.002–1.012 2+ Risky parent × smoking 1.005 0.998–1.012 F 6.96† 16.36† 18.84† N 4,101 3,594 3,594 Note: b. = behavior; BMI = body mass index; CI = confidence interval; HR = hazard ratio; rf = reference group. aAnalyses weighted to adjust for clustering. bReference group is no misfortune. *p < .05; ‡p < .01, †p < .001. View Large When adjusting for adult resources and lifestyle factors in Model 2, the relationship between domains of childhood misfortune and disease incidence attenuated, such that the HRs for 2+ SES and risky adolescent misfortunes are no longer significant. As in Model 1, both age (HR = 1.017) and being female (HR = 0.873) were associated with disease incidence, but Black and Hispanic adults no longer differed from White adults. We tested interaction terms for race and childhood misfortune domains, but none were significant (not shown). Depressive symptoms (HR = 1.059), heavy drinking (HR = 1.365), and being overweight (HR = 1.205) or obese (HR = 1.611) as compared to normal weight increased the risk of incident disease; however, more wealth (HR = 0.998) and physical activity (HR = 0.982) decreased the risk of disease. Tests of Moderation To address our second research question, we estimated a series of models with product terms to assess the moderating effects of adult resources and lifestyle factors on the relationship between childhood misfortune and disease incidence. None of the product terms were significant in the fully adjusted prevalence model. The process was repeated for the incidence analyses, where we identified one significant interaction. As shown in Model 3 of Table 3, among those who were disease free in 2004, disease incidence by 2012 was higher for persons with one risky parental behavior and a high level of pack-years smoking (HR = 1.007). Most HRs remained stable from Model 2 except for the effect for Hispanic adults (HR = 1.179, p < .05). Finding that one risky parental behavior was significant, we subsequently tested product terms between each of the three indicators of risky parental behavior and pack-years. Analyses showed that parental smoking was the source of the interaction. Specifically, having a parent who smoked, in combination with the number of pack-years smoked in adulthood, increased the risk of disease onset by 2012 (HR = 1.006, p < .01; not shown). Discussion This research contributes to the literature on the early origins of adult health by identifying types of childhood misfortune associated with being free of any disease during middle and later life. Our first research question asked whether experiencing childhood misfortune decreased the odds of being disease free in later life. Results from logistic regression of baseline data revealed that experiencing childhood misfortune from several domains—socioeconomic, risky parental behavior, impairments, and/or risky adolescent behavior—reduces the odds of being disease free in later life. Even when adjusting for adult behaviors and resources, experiencing 2+ risky parental behaviors reduced the odds of being disease free. These findings add to the literature demonstrating the long-term influence of early misfortune. Growing up in a risky household, experiencing poor socioeconomic conditions, and/or exhibiting risky behavior in adolescence are the types of childhood misfortune most consequential to being disease free. This is consistent with conclusions of prior studies examining selected diseases among HRS respondents (O’Rand & Hamil-Luker, 2005) as well as those using the MIDUS (Schafer & Ferraro, 2012), Wisconsin Longitudinal Study (Springer, 2009), and Adverse Childhood Experiences (ACE) Study (Brown et al., 2010; Felitti et al., 1998). This study extends the early origins of health literature by identifying types of childhood misfortune that reduce the likelihood of being disease free. CI theory specifies pathways by which disadvantage may or may not accumulate, highlighting how the consequences of early insults are partly contingent on resource activation. Prior to including adult resources and lifestyles in our models, many domains of misfortune were negatively associated with being disease free. However, several of these findings were attenuated after including these adult factors, highlighting the importance of resources and human agency in shaping health trajectories. For instance, remaining physically active increases the likelihood of being disease free at baseline and remaining so over time, consistent with work by others (Barengo, Antikainen, Borodulin, Harald, & Jousilahti, 2017), despite experiencing considerable childhood misfortune. In addition, SES resources are central to overcoming the anticipated deleterious sequelae of early misfortune. In both the prevalence and incidence models, we observed that childhood SES misfortune was related to disease-free status but that these associations were no longer significant after adjusting for adult wealth. This suggests that some of the negative health consequences associated with financial strain or poverty during childhood can be offset by sufficient socioeconomic assets during adulthood. Our second research question asked if and to what extent adult lifestyle and resources moderate the relationship between childhood misfortune and developing disease over time. Our analyses revealed an interaction associated within the risky parental behavior domain: having a parent who smoked during one’s childhood—in combination with the number of pack-years smoked in adulthood—significantly increased the risk of developing disease over time. To our knowledge, no previous study of the early origins of adult health reports such an interaction. To illustrate the dose–response relationship when a parent smoked during one’s childhood, the risk of getting more or more disease rises 12% between a 40 pack-year smoker and a 60 pack-year smoker (i.e., 10 additional cigarettes daily for 40 years). This finding shines a light on the intergenerational and life course implications of tobacco consumption. Parental smoking may influence children’s health through environmental exposure to tobacco smoke (i.e., second-hand smoke), by modeling a lifestyle of tobacco use, and perhaps biologically by inducing symptoms of early nicotine dependence in the womb (Alberg & Korte, 2014). Any sustained exposure to tobacco smoke is harmful, but our analyses show the potency of lifelong exposure to smoking. Some mothers and fathers stop smoking during pregnancy, which is certainly beneficial, yet smoking resumption among parents after birth remains a serious threat to their progeny’s likelihood of disease-free aging. In addition to the general influence of one’s upbringing, the imprint of family lineage is clear (proposition 1.c, Ferraro & Shippee, 2009): risky parental behaviors have long-term consequences on the health of the next generation. Although future research may be able to probe more carefully how much of the family lineage effect is due to genetic and environmental influence, CI theory and the results of this study call for more attention to family lineage as a core element of life course epidemiology. In addition to our main findings, we briefly discuss those pertaining to gender and race. Gender did not predict being disease free at baseline, but predicted disease incidence with women less likely to develop disease. This finding stands in contrast to much of the literature that reports women are more likely to develop chronic diseases (Rieker & Bird, 2005). Although men and women suffer from similar diseases in later life, women live longer lives and more likely to suffer from more chronic, nonlife-threatening diseases (i.e., arthritis) that are not accounted for in our measure. Black adults had significantly lower odds of being disease free in 2004, even in the fully adjusted model. These prevalence results are consistent with much of the disparities literature showing that stark Black–White racial inequalities develop early in the life course, which may be understated in studies of older adults because of early-life mortality selection (Ferraro, Kemp, & Williams, 2017; Woolf & Braveman, 2011). By contrast, although both Black and Hispanic adults had higher risk of developing disease by 2012, these effects became nonsignificant after accounting for adult resources and lifestyle factors—confirming that these factors can ameliorate or exacerbate racial disparities, even during later life. These findings should be interpreted with limitations in mind. First, childhood misfortune was measured using retrospective data, which may involve recall bias. We attempted to reduce the effect of this bias by adjusting for depressive symptoms and socioeconomic resources as recommended by Vuolo, Ferraro, Morton, and Yang (2014). In addition, we restricted the sample to adults who fell within 2 SD of the mean cognition score. Nevertheless, results should be interpreted with care. Second, interpretation of the moderation findings regarding risky parental behaviors, particularly smoking, is limited because of the crudeness of the measure. Although it taps exposure to family tobacco use during childhood, there is no information on duration or quantity of smoking. Future studies that capture more detailed information on parental smoking may help identify the processes at work, including potential mediation. Third, self-reports of a number of variables used here can be limiting. For example, in a review of physical activity measures, Prince and colleagues (2008) found that correlations between self-report and direct measures of physical activity were only low to moderate, creating potential issues of reliability. Furthermore, self-reported, physician-diagnosed disease may underestimate prevalence and incidence due to limited access to medical care, particularly for uninsured and minority adults. Although we adjusted for insurance status and wealth, these three preceding limitations probably lead to a conservative bias, underestimating the influence of early origins on adult health. Finally, studies of disease-free aging have to grapple with the issue of which diseases should be included in one’s definition of this highly favorable health status. Although some include an exhaustive list of diseases and symptoms (see Schafer & Ferraro, 2012), we chose to focus on the leading causes of death in the United States. Empirical generalizations across studies need to consider the types of conditions examined. Conclusion Kuh and Ben-Shlomo (2016) call for more “study of common risk and protective factors” (p. 117). Knowledge about risk factors for specific diseases is vitally important; yet, the public is often confused by conflicting reports about health practices derived from “single-disease epidemiology.” Consequently, identifying risk factors associated with any of the leading causes of death is important for fundamental scientific and public understanding of health. Our findings show that risky parental behaviors, especially smoking, have long-term health consequences on their children. Perhaps more attention on the intergenerational transmission of health risks would aid smoking cessation programs. This study extends the findings of Schafer and Ferraro (2012), demonstrating the early origins of remaining disease-free in middle and later life. Family upbringing characterized by SES misfortune and risky parental behaviors place notable long-term health risks on the children. One might expect that the effect of childhood misfortune on disease onset would be limited to early and middle adulthood, but these results show elevated risk decades after the insults. We do not claim that avoiding a noxious home environment guarantees that children will be able to escape disease altogether, but avoiding or even delaying these killer diseases begins long before symptoms are identified. Our findings may seem disheartening; yet, it is valuable to remember that the trajectories set into motion by experiencing early adversity are not determined for life. Engaging in regular physical activity, abstaining from—or even reducing—tobacco consumption, and accumulating socioeconomic resources may help to interrupt or counteract the likely sequelae of childhood misfortune. Supplementary Material Supplementary data is available at The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences online. Funding This research was supported by a grant from the National Institute on Aging to K. Ferraro (AG043544). Conflict of Interest None reported. Acknowledgments M. M. Williams drafted the manuscript and performed the statistical analyses. B. R. Kemp assisted with the study’s design and analyses. K. F. Ferraro planned the study, supervised the data analysis, and wrote sections of the manuscript. S. A. Mustillo assisted with the study’s analysis and interpretation of results. All authors contributed to revising the manuscript for publication. References Alberg , A. , & Korte , J . ( 2014 ). Parental smoking as a risk factor for adult tobacco use: Can maternal smoking during pregnancy be distinguished from the social environmental influence during childhood ? American Journal of Epidemiology , 179 , 1418 – 1421 . doi: 10.1093/aje/kwu079 Google Scholar CrossRef Search ADS PubMed Barengo, N., Antikainen, R., Borodulin, K., Harald, K., & Jousilahti, P. (2017). Leisure-time physical activity reduces total and cardiovascular mortality and cardiovascular disease incidence in older adults. Journal of the American Geriatrics Society, 65, 504–510. doi:10.1111/jgs.14694 Blackwell , D. L. , Hayward , M. D. , & Crimmins , E. M . ( 2001 ). Does childhood health affect chronic morbidity in later life ? Social Science & Medicine , 52 , 1269 – 1284 . doi:10.1016/S0277-9536(00)00230-6 Google Scholar CrossRef Search ADS Breslow , N . ( 1974 ). Covariance analysis of censored survival data . Biometrics , 30 , 89 – 99 . doi:10.2307/2529620 Google Scholar CrossRef Search ADS PubMed Brown , D. W. , Anda , R. F. , Felitti , V. J. , Edwards , V. J. , Malarcher , A. M. , Croft , J. B. , & Giles , W. H . ( 2010 ). Adverse childhood experiences are associated with the risk of lung cancer: A prospective cohort study . BMC Public Health , 10 , 20 . doi: 10.1186/1471-2458-10-20 Google Scholar CrossRef Search ADS PubMed Critchley , J. A. , & Capewell , S . ( 2003 ). Mortality risk reduction associated with smoking cessation in patients with coronary heart disease: A systematic review . JAMA , 290 , 86 – 97 . doi: 10.1001/jama.290.1.86 Google Scholar CrossRef Search ADS PubMed Dannefer , D . ( 2003 ). Cumulative advantage/disadvantage and the life course: Cross-fertilizing age and social science theory . The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences , 58 , S327 – S337 . Google Scholar CrossRef Search ADS Dawson , D. A . ( 2011 ). Defining risk drinking . Alcohol Research and Health , 34 , 144 – 156 . Google Scholar PubMed Elder , G. H. , Jr . ( 1998 ). The life course as developmental theory . Child Development , 69 , 1 – 12 . Google Scholar CrossRef Search ADS PubMed Evert , J. , Lawler , E. , Bogan , H. , & Perls , T . ( 2003 ). Morbidity profiles of centenarians: Survivors, delayers, and escapers . The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences , 58 , 232 – 237 . Google Scholar CrossRef Search ADS Felitti , V. J. , Anda , R. F. , Nordenberg , D. , Williamson , D. F. , Spitz , A. M. , Edwards , V. ,… Marks , J. S . ( 1998 ). Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults. The Adverse Childhood Experiences (ACE) Study . American Journal of Preventive Medicine , 14 , 245 – 258 . Google Scholar CrossRef Search ADS PubMed Ferraro , K. F. , Kemp , B. R. , & Williams , M. M . ( 2017 ). Diverse aging and health inequality by race and ethnicity . Innovation in Aging , 1 , 1 – 11 . doi: 10.1093/geroni/igx002 Ferraro , K. F. , Schafer , M. H. , & Wilkinson , L. R . ( 2016 ). Childhood disadvantage and health problems in middle and later life: Early imprints on physical health ? American Sociological Review , 81 , 107 – 133 . doi: 10.1177/0003122415619617 Google Scholar CrossRef Search ADS PubMed Ferraro , K. F. , & Shippee , T. P . ( 2009 ). Aging and cumulative inequality: How does inequality get under the skin ? The Gerontologist , 49 , 333 – 343 . doi: 10.1093/geront/gnp034 Google Scholar CrossRef Search ADS PubMed Ferraro , K. F. , Shippee T. P. , & Schafer M. H . ( 2009 ). Cumulative inequality theory for research on aging and the life course . In V. L. Bengtson , D. G. Gans , N. M. Putney , & M. Silverstein (Eds.), Handbook of theories of aging (pp. 413 – 434 ). New York, NY : Springer . Heron , M . ( 2016 ). Deaths: Leading causes for 2014 . National Vital Statistics Report , 65 , 1 – 96 . Kemp , B. R. , Ferraro , K. F. , Morton , P. M. , & Mustillo , S. A . ( 2018 ). Early origins of adult cancer risk among men and women: Influence of childhood misfortune ? Journal of Aging and Health , 30 , 140 – 163 . doi: 10.1177/0898264316670049 Google Scholar CrossRef Search ADS Kuh , D. , & Ben-Shlomo , Y . ( 2016 ). Early life origins of adult health and ageing . In L. K. George & K. F. Ferraro (Eds.), Handbook of aging and the social sciences ( 8th ed ., pp. 101 – 122 ). San Diego, CA : Academic Press . Latham , K. , & Williams , M. M . ( 2015 ). Does neighborhood disorder predict recovery from mobility limitation? Findings from the health and retirement study . Journal of Aging and Health , 27 , 1415 – 1442 . doi: 10.1177/0898264315584328 Google Scholar CrossRef Search ADS PubMed Mood , C . ( 2010 ). Logistic regression: Why we cannot do what we think we can do, and what we can do about it . European Sociological Review , 26 , 67 – 82 . doi: 10.1093/esr/jcp006 Google Scholar CrossRef Search ADS Morton , P. M. , Mustillo , S. A. , & Ferraro , K. F . ( 2014 ). Does childhood misfortune raise the risk of acute myocardial infarction in adulthood ? Social Science & Medicine , 104 , 133 – 141 . doi: 10.1016/j.socscimed.2013.11.026 Google Scholar CrossRef Search ADS National Center for Health Statistics . ( 2015 ). 10 Leading causes of death by age group, United States, 2015 . Hyattsville, MD : Author . National Center for Health Statistics . ( 2016 ). Health, United States, 2015: With special feature on racial and ethnic health disparities . Hyattsville, MD : Author . O’Rand , A. M. , & Hamil-Luker , J . ( 2005 ). Processes of cumulative adversity: Childhood disadvantage and increased risk of heart attack across the life course . The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences , 60 , 117 – 124 . Google Scholar CrossRef Search ADS Pavela , G. , & Latham , K . ( 2016 ). Childhood conditions and multimorbidity among older adults . The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences , 71 , 889 – 901 . doi: 10.1093/geronb/gbv028 Google Scholar CrossRef Search ADS Pearlin , L. I . ( 1989 ). The sociological study of stress . Journal of Health and Social Behavior , 30 , 241 – 256 . Google Scholar CrossRef Search ADS PubMed Prince , S. A. , Adamo , K. B. , Hamel , M. E. , Hardt , J. , Connor Gorber , S. , & Tremblay , M . ( 2008 ). A comparison of direct versus self-report measures for assessing physical activity in adults: A systematic review . The International Journal of Behavioral Nutrition and Physical Activity , 5 , 56 . doi: 10.1186/1479-5868-5-56 Google Scholar CrossRef Search ADS PubMed Rieker , P. P. , & Bird , C. E . ( 2005 ). Rethinking gender differences in health: Why we need to integrate social and biological perspectives . The Journals of Gerontology. Series B: Psychological Sciences and Social Sciences , 60 , 40 – 47 . Google Scholar CrossRef Search ADS Rowe , J. W. , & Kahn , R. L . ( 1987 ). Human aging: Usual and successful . Science , 237 , 143 – 149 . Google Scholar CrossRef Search ADS PubMed Sayer , A. A. , Syddall , H. E. , Gilbody , H. J. , Dennison , E. M. , & Cooper , C . ( 2004 ). Does sarcopenia originate in early life? Findings from the Hertfordshire cohort study . The Journals of Gerontology. Series A: Biological Sciences and Medical Sciences , 59 , M930 – M934 . Google Scholar CrossRef Search ADS Schafer , M. H. , & Ferraro , K. F . ( 2012 ). Childhood misfortune as a threat to successful aging: Avoiding disease . The Gerontologist , 52 , 111 – 120 . doi: 10.1093/geront/gnr071 Google Scholar CrossRef Search ADS PubMed Schafer , M. H. , Ferraro , K. F. , & Mustillo , S. A . ( 2011 ). Children of misfortune: Early adversity and cumulative inequality in perceived life trajectories . American Journal of Sociology , 116 , 1053 – 1091 . Google Scholar CrossRef Search ADS Springer , K. W . ( 2009 ). Childhood physical abuse and midlife physical health: Testing a multi-pathway life course model . Social Science & Medicine , 69 , 138 – 146 . doi: 10.1016/j.socscimed.2009.04.011 Google Scholar CrossRef Search ADS Svanes , C. , Sunyer , J. , Plana , E. , Dharmage , S. , Heinrich , J. , Jarvis , D. ,… Antó , J. M . ( 2010 ). Early life origins of chronic obstructive pulmonary disease . Thorax , 65 , 14 – 20 . doi: 10.1136/thx.2008.112136 Google Scholar CrossRef Search ADS PubMed Thomas , C. , Hyppönen , E. , & Power , C . ( 2008 ). Obesity and type 2 diabetes risk in midadult life: The role of childhood adversity . Pediatrics , 121 , e1240 – e1249 . doi: 10.1542/peds.2007-2403 Google Scholar CrossRef Search ADS PubMed Umstattd Meyer , M. R. , Janke , M. C. , & Beaujean , A. A . ( 2014 ). Predictors of older adults’ personal and community mobility: Using a comprehensive theoretical mobility framework . The Gerontologist , 54 , 398 – 408 . doi: 10.1093/geront/gnt054 Google Scholar CrossRef Search ADS PubMed Von Hippel , P . ( 2007 ). Regression with missing Ys: An improved strategy for analyzing multiply imputed data . Sociological Methodology , 37 , 83 – 117 . doi: 10.1111/j.1467-9531.2007.00180.x Google Scholar CrossRef Search ADS Vuolo , M. , Ferraro , K. F. , Morton , P. M. , & Yang , T. Y . ( 2014 ). Why do older people change their ratings of childhood health ? Demography , 51 , 1999 – 2023 . doi: 10.1007/s13524-014-0344-3 Google Scholar CrossRef Search ADS PubMed Ward , B. W. , Schiller , J. S. , & Goodman , R. A . ( 2014 ). Multiple chronic conditions among US adults: A 2012 update . Preventing Chronic Disease , 11 , E62 . doi: 10.5888/pcd11.130389 Google Scholar PubMed Woolf , S. H. , & Braveman , P . ( 2011 ). Where health disparities begin: The role of social and economic determinants – And why current policies may make matters worse . Health Affairs (Project Hope) , 30 , 1852 – 1859 . doi: 10.1377/hlthaff.2011.0685 Google Scholar CrossRef Search ADS PubMed Xu , J. , Murphy , S. L. , Kochanek , K. D. , & Arias , E . ( 2016 ). Mortality in the United States, 2015 (NCHS Data Brief No. 267). Atlanta, GA : National Center for Health Statistics , 1 – 8 . © The Author(s) 2018. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The Journals of Gerontology Series B: Psychological Sciences and Social Sciences Oxford University Press

Avoiding the Major Causes of Death: Does Childhood Misfortune Reduce the Likelihood of Being Disease Free in Later Life?

Loading next page...
 
/lp/ou_press/avoiding-the-major-causes-of-death-does-childhood-misfortune-reduce-tU0aO3s02d
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
1079-5014
eISSN
1758-5368
D.O.I.
10.1093/geronb/gby039
Publisher site
See Article on Publisher Site

Abstract

Abstract Objectives Although previous research reveals the detrimental effects of early misfortune on the development of chronic diseases in later life, few studies have investigated its effects on remaining disease free. This study draws on cumulative inequality theory to investigate whether experiencing childhood misfortune reduces the likelihood of remaining disease free over time. Method This study utilizes five waves of data from the Health and Retirement Study to test whether five domains of childhood misfortune predict being disease free at baseline (2004) and developing disease over time (2004–2012). Results Respondents reporting risky parental behaviors during childhood were less likely to be disease free at baseline and had an increased risk of disease onset over time, the latter driven by having a guardian who smoked in combination with more pack-years smoked in adulthood. Furthermore, we find that adult resources, that is wealth, help to mitigate the noxious effects of other misfortunes, notably poor socioeconomic conditions. Discussion Consistent with cumulative inequality theory, these findings reveal that experiencing multiple types of misfortune during childhood decreases the likelihood of remaining disease free in later life, but engaging in health behaviors, such as physical activity, can help to ameliorate some of the noxious effects of early misfortune. Childhood disadvantage, Cumulative inequality theory, Successful aging Nearly 86% of Americans aged 65 and older live with at least one chronic disease, whereas more than 60% have two or more diseases (Ward, Schiller, & Goodman, 2014). Many chronic diseases are also the leading causes of death for older Americans: Two thirds of their deaths are due to heart disease, cancer, chronic lower respiratory disease, stroke, Alzheimer’s disease, and diabetes (Heron, 2016; Xu, Murphy, Kochanek, & Arias, 2016). Although chronic disease in later life is widespread and studied extensively, relatively few investigators examine how people are able to avoid any of the chronic “killer” diseases during later life. Most gerontologists do not consider being disease free a sine qua non for “successful aging,” but avoiding the leading causes of death greatly improves one’s chances for exceptional longevity (Evert, Lawler, Bogan, & Perls, 2003; Rowe & Kahn, 1987). Studying the antecedents of remaining free of chronic diseases is a logical way to identify risks that are common to multiple diseases and fundamental factors that promote health and well-being. Epidemiology’s search for risk factors common across life-threatening diseases also has led to greater consideration of experiences from childhood. Although hundreds of studies have identified adult risk factors of various chronic diseases, a now well-established body of research has unveiled the influence of early-life risk factors on adult diseases, ranging from lung cancer (Brown et al., 2010) to sarcopenia (Sayer, Syddall, Gilbody, Dennison, & Cooper, 2004). Kuh and Ben-Shlomo (2016) describe the shift in epidemiology from an adult life style model to one that also accounts for life course influences. In this article, we examine life course influences on the likelihood of being disease free in later life. We ask whether childhood misfortune reduces the likelihood that older adults will be able to avoid major, life-threatening diseases. Although reaching later life free of chronic disease is both rare and remarkable, to our knowledge, there is only one study that systematically examines early-life influences on this outcome (Schafer & Ferraro, 2012). The present study examines the enduring effects of childhood misfortune on health in later life and is distinct from parallel studies in several ways outlined in the following section. Childhood Misfortune and Avoiding Disease Dozens of studies during the past two decades reveal how early-life insults elevate the risk of specific diseases during adulthood. There is compelling evidence that childhood misfortune raises the risk of heart disease (Morton, Mustillo, & Ferraro, 2014; O’Rand & Hamil-Luker, 2005), diabetes (Thomas, Hyppönen, & Power, 2008), chronic lung disease (Svanes et al., 2010), and cancer (Brown et al., 2010; Kemp, Ferraro, Morton, & Mustillo, 2018) directly, through stress processes, and indirectly, through poor environments and health behaviors. These findings have been valuable for identifying life course influences on conditions often identified as “age-related diseases.” Although these diseases are prevalent among older people, the etiology of these and other conditions begins much earlier. Investigating individual diseases is fruitful for advancing life course epidemiology, but it is plausible that some risk factors are pluripotent. For example, Blackwell, Hayward, and Crimmins (2001) found that poor childhood health is associated with cancer, lung disease, cardiovascular conditions, and arthritis, even after adjusting for life course socioeconomic status. Furthermore, research examining disease accumulation reveals that early-life misfortune raises the risk of comorbidity in later life (Ferraro, Schafer, & Wilkinson, 2016; Pavela & Latham, 2016). Failure to account for related diseases in studies of single diseases, however, may lead to misdirected conclusions. For example, a person may be cancer-free giving the appearance of relative health, yet suffer from other, equally life-threatening conditions. Although studying the early origins of disease-free aging is a promising line of inquiry to identify factors that promote longevity and overall health, it is surprising that so few life course studies have examined being disease free as an outcome. The study that directly motivated the present investigation used data from the Midlife Development in the United States Study (MIDUS). Schafer and Ferraro (2012) found that a simple sum of childhood misfortune indicators, ranging from physical and mental abuse to financial strain, decreased the likelihood of being disease free in adulthood, both at baseline and 10 years later (1995 and 2015). Although being disease free at the follow-up survey was infrequent (<7%), the influence of early misfortune did not wane. The findings from this study are important but limited in several ways. First, given its focus on midlife, the MIDUS sample has a baseline age ceiling of 74 years, whereas the Health and Retirement Study (HRS) does not, enabling us to examine the highly selective situation where septuagenarians and octogenarians are disease free. Second, although Schafer and Ferraro (2012) examined disease prevalence and incidence across two waves of data, separated by 8 years, we use five waves of HRS data collected at 2-year intervals to permit a more fine-grained analysis of transitions involving disease incidence. This longitudinal approach enables us to examine trends and the risk of getting a major disease over time. Third, the authors incorporated an extensive list of conditions in their conceptualization of disease free, but some of these conditions are symptoms and often temporary (e.g., teeth problems, constipation). Instead, we focus on leading causes of death. Fourth, rather than adding all types of misfortune into a single variable, we examine specific domains of childhood misfortune to identify which types of misfortune are most detrimental in a parsimonious way. Finally, an oversampling of minority respondents in the HRS allows for in-depth investigation into whether the relationship between childhood misfortune and remaining disease free in later life varies by race and ethnicity. Black adults have higher prevalence of several diseases considered to be leading causes of death and higher mortality rates than White adults (National Center for Health Statistics, 2016). Evidence regarding health disparities between Hispanic and non-Hispanic Americans is less consistent but merits study nonetheless. Theoretical Background Cumulative inequality (CI) theory informs this study of the early origins of adult health (Schafer, Ferraro, & Mustillo, 2011), integrating elements of the life course perspective (Elder, 1998), cumulative disadvantage theory (Dannefer, 2003), and stress process theories (Pearlin, 1989). Instead of focusing on proximal factors such as adult socioeconomic status and health behaviors, the theory emphasizes how social stratification manifests over the life course, especially how early disadvantage shapes risks and resources over time (Ferraro & Shippee, 2009; Ferraro et al., 2016). Early disadvantage may alter health trajectories in several ways, and four elements of the theory are germane for the present study. First, recognizing the role of family lineage, CI theory draws on the life course perspective and “linked lives” to argue that childhood conditions, including inequality across families of origin, are integral to stratification processes (Ferraro, Shippee, & Schafer, 2009). Second, the theory emphasizes how inequality diffuses across life domains—“disadvantage in one domain (e.g., family life) may spill over to other domains (e.g., mental and physical health)” (Ferraro & Shippee, 2009, p. 339), resulting in stress proliferation (Pearlin, 1989). Rather than add all indicators of misfortune into one variable, a formerly common practice, the theory calls for examining multiple domains of misfortune, when such measures are available. Third, a novel contribution of CI theory is that it emphasizes the plasticity of human development and the ways in which disadvantage does not accumulate. Although experiencing misfortune in childhood can be a strong predictor of health problems in later life by influencing continued exposure to stressors and behaviors unconducive to adult health (e.g., smoking), human agency and resource activation can mitigate the unfavorable consequences of early misfortune (Ferraro et al., 2009, 2016). The theory cautions about “runaway cumulative disadvantage” and calls for identifying factors that can avert the presumed health consequences. Finally, cumulative inequality often leads to premature mortality among those most exposed to adversity, creating compositional change in a population that can masquerade as decreasing inequality (Ferraro & Shippee, 2009). It may be more difficult to detect the potential influence of early misfortune on persons belonging to racial and ethnic minorities because of selective mortality. Thus, we systematically examine racial and ethnic differences and mortality selection in our analyses. Research Questions The present study is innovative in several ways and addresses the following research questions: Does childhood misfortune decrease the likelihood of older adults being disease free? We address this question at the initial survey and across 8 years of longitudinal follow-ups, anticipating that childhood misfortune is a threat to disease-free aging. Do adult resources and lifestyle factors moderate the relationship between childhood misfortune and being disease free in later life? We anticipate that the relationship between childhood misfortune and being disease free will be moderated by adult resources (e.g., education), and lifestyle factors (e.g., smoking). Method Sample Data were obtained from the HRS—a multistage, probability study of adults aged 51 years and older, with an oversampling of Black and Hispanic adults and Florida residents. Waves 7–11 (2004–2012) were used for this research because a detailed battery of childhood misfortune indicators were first measured on the full sample in 2004. For this study, respondents were excluded if they were age ineligible, deceased as of 2004, or entered the study in 2010. In addition, those who had proxy responses or fell 2 SD below the average cognition score were excluded to preserve the validity of the responses, particularly for the retrospective childhood misfortune variables. These restrictions resulted in a final, analytic sample of 13,921 adults. Disease Free There are different ways to define being disease free. One could give particular attention to chronic diseases and life-threatening conditions or could include every potential health condition experienced throughout life. However, the majority of adults aged 50 and older experience an array of minor health symptoms and conditions, yet continue to live long and fairly healthy lives until faced with one of the major causes of death. We focus on those diseases that are both chronic and life threatening. As of 2014, the top chronic disease-related causes of death for those aged 50 and older included heart disease, cancer, chronic lower respiratory disease, cerebrovascular disease, diabetes, Alzheimer’s disease, kidney disease, and liver disease (National Center for Health Statistics, 2015); we include the aforementioned diseases to create our indicators of being disease free at 2004 (prevalence) and developing disease (incidence) from 2004 to 2012. Respondents were asked if a doctor had ever told them that they have diabetes, lung disease, stroke, heart problems, hypertension, or cancer (excluding skin). Respondents also were asked if they had ever been diagnosed with a “memory-related disease.” In 2010 and 2012, this question was dropped in favor of two questions addressing Alzheimer’s disease and dementia separately. Respondents also were asked if they had other diseases or medical conditions not addressed. We used these responses to gather information about chronic lower respiratory disease, liver disease, and kidney disease. We also included decedents from any of the nine aforementioned diseases using the National Death Index linkage. As specified in axiom 5 of CI theory, failure to account for selective mortality may underestimate disease incidence for the most virulent conditions (i.e., diagnosed with a disease and died between interviews). For the baseline analyses of prevalence, a binary variable was created: one or more diseases (=0) and disease free (=1). Respondents who reported a memory disease prior to 2010 or Alzheimer’s disease or dementia during or after 2010 were coded as 0. Respondents reporting one or more diseases were coded as 0, regardless of missingness on other disease variables. However, respondents reporting no diseases were coded as 1 only if they had valid answers for at least half (four of seven), physician-diagnosed diseases. For the longitudinal analyses of incidence, we inverted the coding to be consistent with the widely used concept of disease incidence (0 = disease free; 1 = one or more diseases between 2004 and 2012). Childhood Misfortune Childhood misfortune was operationalized differently in the literature. Instead of adding indicators of misfortune into an overall construct as others have done, we draw from CI theory to identify which domains of childhood misfortune are consequential to remaining disease free (proposition 2.b: Ferraro & Shippee, 2009). In preliminary inquiry, we used both exploratory and confirmatory factor analyses to finalize the indicators used to specify the domains, eliminating two potential indicators. We categorized misfortunes experienced before the age of 18 into five domains: socioeconomic status (SES), risky parental behaviors, infectious diseases, impairments, and risky adolescent behaviors, using 20 retrospective questions collected between 2004 and 2008. Socioeconomic status (SES) misfortune consisted of five indicators: reporting poor or fair finances in relation to others, moving due to finances, having a father with an unskilled labor occupation, mother having less than a high school education, and/or father having less than a high school education. Risky parental behavior was assessed by having a parent who smoked, abused substances, and/or was physically abusive. Infectious disease included measles, mumps, and/or chicken pox. The presence of a childhood impairment included head injury, disability for more than 6 months, learning problems, visual impairment, and/or speech impairment. Risky adolescent behavior was assessed by having trouble with police, substance abuse, depressive symptoms, and/or other psychological issues. Each indicator was coded dichotomously and summed to create a count of misfortune within each domain, top coded at 2 (0, 1, and 2+) to examine a threshold effect of accumulating two or more misfortunes as compared to none. Several domain coding schemes were examined (e.g., dichotomous, top coded at 3+); however, differentiating 0, 1, and 2+ indicators in each domain yielded the most useful information. Multiple imputation by Chained Equations (MICE) was used to handle missing data among childhood misfortune domains with missing exceeding 5% (SES = 30.19%; risky parental behavior = 20.69%; risky adolescent behavior = 20.21%); observations on other variables that initially had missing were excluded prior to analysis (Von Hippel, 2007). All variables included in the analysis were used to impute 20 datasets using the “mi impute chained” command in Stata with the appropriate link functions for each variable. Adult Resources and Lifestyle Factors Although a challenging childhood can predict health problems for older adults, not everyone who experiences childhood misfortune suffers from poor health in later life. Childhood circumstances can have lasting impacts on adult health by influencing continued exposure to stressors and behaviors unconducive to adult health. Yet, resilience through human agency and resource activation can help to protect against some of these health risks (Ferraro et al., 2009). For example, smoking cessation has been associated with substantial reduction in risk of all-cause mortality among those with coronary heart disease (Critchley & Capewell, 2003). Thus, we include both adult resources and lifestyle factors that may modify health risks. We examined four indicators of adult resources: marital status, education, wealth, and health insurance status. Marital status was coded as never married, married (reference), widowed, and divorced/separated. Education was measured in years and top coded at 17+. Wealth is presented in tens of thousands of dollars and cube rooted to correct for skewness. A four-category indicator of health insurance status was created: no insurance (reference), Medicaid only, Medicare only, and private insurance (private only and/or Medigap plans). Five indicators of adult health characteristics were included in the models: depressive symptoms, smoking status, heavy drinking, body mass index (BMI), and physical activity. Depressive symptoms were measured using an eight-item Center for Epidemiological Studies–Depression (CES-D) scale. Smoking was assessed by a pack-years variable of cigarette smoking (one pack-year is equivalent to smoking 20 cigarettes daily for 1 year) where never smokers were coded as 0. A heavy drinker was dichotomously coded as having an average of five or more drinks per day for men and four or more drinks per day for women (Dawson, 2011). BMI was based on self-reports and categorized into underweight or normal weight (BMI < 25), overweight (25 ≥ BMI < 30), and obese (BMI ≥ 30). A self-reported physical activity scale accounts for both frequency and intensity. Respondents were asked the frequency in which they participated in mild, moderate, and vigorous activity; possible answers included never/rarely (=0), one to three times per month (=1), once per week (=2), two or more times per week (=3), and every day (=4). Drawing from other researchers, the physical activity scale was created by weighting the type of physical activity by intensity (mild = 1.2, moderate = 1.4, and vigorous = 1.8) based on metabolic equivalent recommendations (Latham & Williams, 2015; Umstattd Meyer, Janke, & Beaujean, 2014). Possible scores ranged from 0 (no physical activity) to 17.6 (mild, moderate, and vigorous physical activity daily). Demographics Finally, models adjusted for demographic characteristics reported at baseline including age (in years), gender (female = 1), and race/ethnicity (non-Hispanic white = 0, non-Hispanic black, and Hispanic). Other races were excluded as the number of cases was insufficient for meaningful comparisons (n = 303). Analysis The analysis was conducted in three parts. First, logistic regression was used to estimate the prevalence of being disease free in 2004 in the full sample. Second, Cox proportional hazards models were estimated, using time-on-study with wave (every 2 years) as our time scale, to investigate disease incidence (reporting one or more diseases) between 2004 and 2012 among a more selective, healthy sample—those who were disease free in the prevalence model (2004). The proportional hazards assumption was tested for potential time-varying covariates with no violation encountered, and ties were handled with the Breslow’s (1974) method. Finally, a series of theoretically informed interactions were tested to assess the potential moderating effects of adult risks and resources. Models utilized 2004 HRS weights and accounted for the survey design. Analyses were conducted using Stata ST/SE 14.1. Additional information is presented in Supplementary Material. Results Sample Characteristics The weighted descriptive statistics for focal variables are presented in Table 1 (see Supplementary Table S1 for descriptive statistics for all variables in the final analyses). At baseline (2004), approximately 35.30% of the sample was disease free. By 2012, only 15.92% of the sample remained disease free. For the childhood misfortune variables, respondents were most likely to report infectious disease (1.74; SD = 0.57) and socioeconomic misfortunes (1.45; SD = 0.78). Table 1. Select Descriptive Statistics for the Analytical Sample (N = 13,921)a Variable Range Percent Mean (SE) Disease free (prevalence)b  2004 0, 1 35.30  2006 0, 1 28.02  2008 0, 1 23.02  2010 0, 1 18.81  2012 0, 1 15.92 Childhood misfortune  Socioeconomic 0–2 1.45 (0.78)  Risky parental behavior 0–2 0.90 (0.67)  Infectious disease 0–2 1.74 (0.57)  Impairments 0–2 0.22 (0.48)  Risky adolescent behavior 0–2 0.29 (0.54) Variable Range Percent Mean (SE) Disease free (prevalence)b  2004 0, 1 35.30  2006 0, 1 28.02  2008 0, 1 23.02  2010 0, 1 18.81  2012 0, 1 15.92 Childhood misfortune  Socioeconomic 0–2 1.45 (0.78)  Risky parental behavior 0–2 0.90 (0.67)  Infectious disease 0–2 1.74 (0.57)  Impairments 0–2 0.22 (0.48)  Risky adolescent behavior 0–2 0.29 (0.54) Note:aWeighted values presented. bDiseases include cancer, heart disease, hypertension, chronic lower respiratory disease, cerebrovascular disease, diabetes, liver disease, memory disease, and kidney disease. View Large Table 1. Select Descriptive Statistics for the Analytical Sample (N = 13,921)a Variable Range Percent Mean (SE) Disease free (prevalence)b  2004 0, 1 35.30  2006 0, 1 28.02  2008 0, 1 23.02  2010 0, 1 18.81  2012 0, 1 15.92 Childhood misfortune  Socioeconomic 0–2 1.45 (0.78)  Risky parental behavior 0–2 0.90 (0.67)  Infectious disease 0–2 1.74 (0.57)  Impairments 0–2 0.22 (0.48)  Risky adolescent behavior 0–2 0.29 (0.54) Variable Range Percent Mean (SE) Disease free (prevalence)b  2004 0, 1 35.30  2006 0, 1 28.02  2008 0, 1 23.02  2010 0, 1 18.81  2012 0, 1 15.92 Childhood misfortune  Socioeconomic 0–2 1.45 (0.78)  Risky parental behavior 0–2 0.90 (0.67)  Infectious disease 0–2 1.74 (0.57)  Impairments 0–2 0.22 (0.48)  Risky adolescent behavior 0–2 0.29 (0.54) Note:aWeighted values presented. bDiseases include cancer, heart disease, hypertension, chronic lower respiratory disease, cerebrovascular disease, diabetes, liver disease, memory disease, and kidney disease. View Large Disease-Free Prevalence Table 2 presents the logistic regression models predicting being disease free in 2004. When adjusting for age, sex, and race in Model 1, respondents who experienced 2+ socioeconomic misfortunes during childhood had 31.2% lower odds of being disease free in 2004, relative to those who experienced no misfortune. Furthermore, respondents who experienced one or 2+ risky parental behaviors (odds ratios [ORs] = 0.882, 0.737, respectively), impairments (ORs = 0.863, 0.712), or risky adolescent behaviors (ORs = 0.815, 0.706) also had lower odds of being disease free in 2004 than those who experienced none. As expected, the odds of being disease free decreased with age, and Black adults had about half the odds of being disease free as White adults. We present Model 1 to highlight the relationships between types of childhood misfortune and disease-free status without adjusting for adult resources but urge caution when interpreting the results. Whereas Model 1 omits variables related to being disease free, and omitted variables bias the ORs for the variables that are included in the model (Mood, 2010), we interpret the ORs recognizing that the estimates are probably biased downward. Estimates from Model 2 will be closer to the true effects. Table 2. Logistic Regression Predicting Prevalence of Being Disease Free at Baseline (2004)a Model 1 Model 2 OR 95% CI OR 95% CI Childhood misfortuneb  Socioeconomic (=1) 0.865 0.717–1.043 0.974 0.784–1.209  Socioeconomic (=2+) 0.688† 0.577–0.820 0.870 0.707–1.069  Risky parent behavior (=1) 0.882* 0.794–0.979 0.908 0.805–1.023  Risky parent behavior (=2+) 0.737† 0.635–0.855 0.791‡ 0.667–0.937  Infectious disease (=1) 1.111 0.866–1.427 0.950 0.714–1.264  Infectious disease (=2+) 0.923 0.752–1.132 0.806 0.644–1.009  Impairment (=1) 0.863* 0.772–0.965 0.904 0.787–1.039  Impairment (=2+) 0.712‡ 0.564–0.898 0.843 0.637–1.115  Risky adolescent behavior (=1) 0.815‡ 0.704–0.943 0.951 0.804–1.127  Risky adolescent behavior (=2+) 0.706* 0.528–0.944 0.935 0.681–1.282 Demographics  Age (years at baseline) 0.945† 0.940–0.950 0.936† 0.930–0.942  Femalec 1.053 0.957–1.159 1.085 0.965–1.219  Blackd 0.504† 0.418–0.608 0.682† 0.558–0.833  Hispanicd 0.915 0.780–1.072 1.120 0.901–1.394 Adult resources and lifestyle factors  Divorcede 1.091 0.930–1.280  Widowede 1.229‡ 1.077–1.403  Never marriede 1.165 0.888–1.529  Education (years) 1.001 0.979–1.022  Wealth (∛) 1.004† 1.002–1.005  Medicaidf 0.424† 0.309–0.581  Medicare/VAf 0.757 0.571–1.003  Private/Medigapf 0.739‡ 0.578–0.944  Depressive symptoms 0.884† 0.851–0.918  Smoking (pack-years) 0.994† 0.991–0.996  Heavy drinkerg 1.250 0.914–1.709  25 ≤ BMI < 30h 0.586† 0.523–0.658  BMI ≥ 30h 0.335† 0.289–0.389  Physical activity 1.046† 1.032–1.061 Constant 32.389† 20.877–50.249 74.139† 40.656–135.198 F 51.51† 67.98† N 13,824 12,023 Model 1 Model 2 OR 95% CI OR 95% CI Childhood misfortuneb  Socioeconomic (=1) 0.865 0.717–1.043 0.974 0.784–1.209  Socioeconomic (=2+) 0.688† 0.577–0.820 0.870 0.707–1.069  Risky parent behavior (=1) 0.882* 0.794–0.979 0.908 0.805–1.023  Risky parent behavior (=2+) 0.737† 0.635–0.855 0.791‡ 0.667–0.937  Infectious disease (=1) 1.111 0.866–1.427 0.950 0.714–1.264  Infectious disease (=2+) 0.923 0.752–1.132 0.806 0.644–1.009  Impairment (=1) 0.863* 0.772–0.965 0.904 0.787–1.039  Impairment (=2+) 0.712‡ 0.564–0.898 0.843 0.637–1.115  Risky adolescent behavior (=1) 0.815‡ 0.704–0.943 0.951 0.804–1.127  Risky adolescent behavior (=2+) 0.706* 0.528–0.944 0.935 0.681–1.282 Demographics  Age (years at baseline) 0.945† 0.940–0.950 0.936† 0.930–0.942  Femalec 1.053 0.957–1.159 1.085 0.965–1.219  Blackd 0.504† 0.418–0.608 0.682† 0.558–0.833  Hispanicd 0.915 0.780–1.072 1.120 0.901–1.394 Adult resources and lifestyle factors  Divorcede 1.091 0.930–1.280  Widowede 1.229‡ 1.077–1.403  Never marriede 1.165 0.888–1.529  Education (years) 1.001 0.979–1.022  Wealth (∛) 1.004† 1.002–1.005  Medicaidf 0.424† 0.309–0.581  Medicare/VAf 0.757 0.571–1.003  Private/Medigapf 0.739‡ 0.578–0.944  Depressive symptoms 0.884† 0.851–0.918  Smoking (pack-years) 0.994† 0.991–0.996  Heavy drinkerg 1.250 0.914–1.709  25 ≤ BMI < 30h 0.586† 0.523–0.658  BMI ≥ 30h 0.335† 0.289–0.389  Physical activity 1.046† 1.032–1.061 Constant 32.389† 20.877–50.249 74.139† 40.656–135.198 F 51.51† 67.98† N 13,824 12,023 Note: BMI = body mass index; CI = confidence interval; OR = odds ratio. aAnalyses weighted to adjust for clustering. bReference group for each of the five domains is no misfortune. cReference group is male. dReference group is non-Hispanic White. eReference group is married. fReference group is no insurance. gReference group is non-heavy drinker. hReference group is BMI < 25. *p < .05; ‡p < .01, †p < .001. View Large Table 2. Logistic Regression Predicting Prevalence of Being Disease Free at Baseline (2004)a Model 1 Model 2 OR 95% CI OR 95% CI Childhood misfortuneb  Socioeconomic (=1) 0.865 0.717–1.043 0.974 0.784–1.209  Socioeconomic (=2+) 0.688† 0.577–0.820 0.870 0.707–1.069  Risky parent behavior (=1) 0.882* 0.794–0.979 0.908 0.805–1.023  Risky parent behavior (=2+) 0.737† 0.635–0.855 0.791‡ 0.667–0.937  Infectious disease (=1) 1.111 0.866–1.427 0.950 0.714–1.264  Infectious disease (=2+) 0.923 0.752–1.132 0.806 0.644–1.009  Impairment (=1) 0.863* 0.772–0.965 0.904 0.787–1.039  Impairment (=2+) 0.712‡ 0.564–0.898 0.843 0.637–1.115  Risky adolescent behavior (=1) 0.815‡ 0.704–0.943 0.951 0.804–1.127  Risky adolescent behavior (=2+) 0.706* 0.528–0.944 0.935 0.681–1.282 Demographics  Age (years at baseline) 0.945† 0.940–0.950 0.936† 0.930–0.942  Femalec 1.053 0.957–1.159 1.085 0.965–1.219  Blackd 0.504† 0.418–0.608 0.682† 0.558–0.833  Hispanicd 0.915 0.780–1.072 1.120 0.901–1.394 Adult resources and lifestyle factors  Divorcede 1.091 0.930–1.280  Widowede 1.229‡ 1.077–1.403  Never marriede 1.165 0.888–1.529  Education (years) 1.001 0.979–1.022  Wealth (∛) 1.004† 1.002–1.005  Medicaidf 0.424† 0.309–0.581  Medicare/VAf 0.757 0.571–1.003  Private/Medigapf 0.739‡ 0.578–0.944  Depressive symptoms 0.884† 0.851–0.918  Smoking (pack-years) 0.994† 0.991–0.996  Heavy drinkerg 1.250 0.914–1.709  25 ≤ BMI < 30h 0.586† 0.523–0.658  BMI ≥ 30h 0.335† 0.289–0.389  Physical activity 1.046† 1.032–1.061 Constant 32.389† 20.877–50.249 74.139† 40.656–135.198 F 51.51† 67.98† N 13,824 12,023 Model 1 Model 2 OR 95% CI OR 95% CI Childhood misfortuneb  Socioeconomic (=1) 0.865 0.717–1.043 0.974 0.784–1.209  Socioeconomic (=2+) 0.688† 0.577–0.820 0.870 0.707–1.069  Risky parent behavior (=1) 0.882* 0.794–0.979 0.908 0.805–1.023  Risky parent behavior (=2+) 0.737† 0.635–0.855 0.791‡ 0.667–0.937  Infectious disease (=1) 1.111 0.866–1.427 0.950 0.714–1.264  Infectious disease (=2+) 0.923 0.752–1.132 0.806 0.644–1.009  Impairment (=1) 0.863* 0.772–0.965 0.904 0.787–1.039  Impairment (=2+) 0.712‡ 0.564–0.898 0.843 0.637–1.115  Risky adolescent behavior (=1) 0.815‡ 0.704–0.943 0.951 0.804–1.127  Risky adolescent behavior (=2+) 0.706* 0.528–0.944 0.935 0.681–1.282 Demographics  Age (years at baseline) 0.945† 0.940–0.950 0.936† 0.930–0.942  Femalec 1.053 0.957–1.159 1.085 0.965–1.219  Blackd 0.504† 0.418–0.608 0.682† 0.558–0.833  Hispanicd 0.915 0.780–1.072 1.120 0.901–1.394 Adult resources and lifestyle factors  Divorcede 1.091 0.930–1.280  Widowede 1.229‡ 1.077–1.403  Never marriede 1.165 0.888–1.529  Education (years) 1.001 0.979–1.022  Wealth (∛) 1.004† 1.002–1.005  Medicaidf 0.424† 0.309–0.581  Medicare/VAf 0.757 0.571–1.003  Private/Medigapf 0.739‡ 0.578–0.944  Depressive symptoms 0.884† 0.851–0.918  Smoking (pack-years) 0.994† 0.991–0.996  Heavy drinkerg 1.250 0.914–1.709  25 ≤ BMI < 30h 0.586† 0.523–0.658  BMI ≥ 30h 0.335† 0.289–0.389  Physical activity 1.046† 1.032–1.061 Constant 32.389† 20.877–50.249 74.139† 40.656–135.198 F 51.51† 67.98† N 13,824 12,023 Note: BMI = body mass index; CI = confidence interval; OR = odds ratio. aAnalyses weighted to adjust for clustering. bReference group for each of the five domains is no misfortune. cReference group is male. dReference group is non-Hispanic White. eReference group is married. fReference group is no insurance. gReference group is non-heavy drinker. hReference group is BMI < 25. *p < .05; ‡p < .01, †p < .001. View Large When further adjusting for adult resources and lifestyle factors in Model 2, the magnitude of several relationships between childhood misfortune and being disease free is reduced; the ORs for socioeconomic, impairments, and risky adolescent behavior misfortunes are no longer significant. However, experiencing 2+ risky parental behaviors in childhood was still associated with being disease free in 2004 (OR = 0.791). Age (OR = 0.936) and race (Black; OR = 0.682) were associated with lower odds of being disease free. We used linear probability models to examine potential interactions between race/ethnicity and each of the childhood misfortune domains (not shown) but did not find any evidence that childhood misfortune differentially affects the odds of being disease free by race. Among adult resources, wealth (OR = 1.004) and being widowed as opposed to married (OR = 1.229) increased one’s odds of being disease free. Conversely, older adults with Medicaid exclusively (OR = 0.424) and private and/or Medigap insurance plans (OR = 0.739) were less likely to report being disease free compared with those with no insurance. Depressive symptoms (OR = 0.884), smoking (OR = 0.994), and being overweight (OR = 0.586) or obese (OR = 0.335) decrease the odds, whereas physical activity increases those odds of being disease free (OR = 1.046). Disease Incidence Table 3 presents the longitudinal results for disease incidence among those who were disease free in 2004. Cox proportional hazard models are used to predict disease onset by 2012. When adjusting for demographic factors only (Model 1), respondents who experienced 2+ SES or risky adolescent misfortunes in childhood, had 17.1% and 34.1%, respectively, higher hazard of developing disease between 2004 and 2012 compared with those who experienced no misfortune within those domains. In addition, age (hazard ratio [HR] = 1.014) increased the hazard of getting one or more diseases, whereas being female lowered the hazard by 15.6%. Finally, Black and Hispanic adults had 26.3% and 25.4% greater hazard of developing disease than White adults. Table 3. Cox Proportional Hazards Predicting Disease Incidence (2004–2012)a Model 1 Model 2 Model 3 HR 95% CI HR 95% CI HR 95% CI Childhood misfortuneb  Socioeconomic (=1) 1.105 0.946–1.292 1.118 0.946–1.323 1.114 0.940–1.322  Socioeconomic (=2+) 1.171* 1.017–1.349 1.110 0.932–1.322 1.107 0.928–1.322  Risky parent b. (=1) 1.046 0.935–1.171 1.057 0.939–1.191 0.977 0.852–1.121  Risky parent b. (=2+) 1.063 0.927–1.217 1.072 0.930–1.235 1.022 0.849–1.230  Infectious disease (=1) 1.081 0.894–1.306 1.164 0.936–1.447 1.162 0.936–1.443  Infectious disease (=2+) 1.048 0.881–1.247 1.075 0.885–1.306 1.077 0.888–1.307  Impairment (=1) 1.015 0.896–1.151 0.962 0.840–1.101 0.968 0.846–1.107  Impairment (=2+) 1.084 0.865–1.358 1.134 0.880–1.462 1.147 0.888–1.480  Risky adolescent b. (=1) 1.058 0.931–1.204 1.049 0.910–1.209 1.049 0.912–1.206  Risky adolescent b. (=2+) 1.341* 1.033–1.740 1.265 0.948–1.687 1.239 0.924–1.663 Demographics  Age (years at baseline) 1.014† 1.009–1.019 1.017† 1.011–1.022 1.017† 1.011–1.022  Female 0.844‡ 0.761–0.935 0.873* 0.785–0.969 0.873* 0.786–0.971  Race/ethnicity (White = rf)   Black 1.263‡ 1.107–1.440 1.137 0.981–1.318 1.132 0.975–1.314   Hispanic 1.254‡ 1.089–1.443 1.166 0.990–1.375 1.179* 1.005–1.384 Adult resources and lifestyle  Marital status (married = rf)   Divorced 0.925 0.812–1.054 0.929 0.815–1.060   Widowed 1.011 0.873–1.170 1.001 0.866–1.158   Never married 1.097 0.906–1.328 1.102 0.909–1.337  Education (years) 1.012 0.992–1.033 1.014 0.994–1.034  Wealth (∛) 0.998* 0.997–1.000 0.998* 0.997–1.000  Insurance (none = rf)   Medicaid 1.152 0.839–1.581 1.203 0.891–1.623   Medicare/VA 1.096 0.876–1.371 1.093 0.874–1.366   Private/Medigap 1.022 0.858–1.218 1.023 0.857–1.222  Depressive symptoms 1.059† 1.034–1.084 1.059 1.034–1.084  Smoking (pack-years) 1.000 0.998–1.002 0.996* 0.992–1.000  Heavy drinker 1.365* 1.078–1.729 1.371‡ 1.089–1.727 BMI (<25 = rf)   25 ≤ BMI < 30 1.205† 1.096–1.324 1.206 1.097–1.325   BMI ≥ 30 1.611† 1.441–1.801 1.619 1.448–1.810  Physical activity 0.982* 0.966–0.998 0.982* 0.966–0.998 1 Risky parent × smoking 1.007‡ 1.002–1.012 2+ Risky parent × smoking 1.005 0.998–1.012 F 6.96† 16.36† 18.84† N 4,101 3,594 3,594 Model 1 Model 2 Model 3 HR 95% CI HR 95% CI HR 95% CI Childhood misfortuneb  Socioeconomic (=1) 1.105 0.946–1.292 1.118 0.946–1.323 1.114 0.940–1.322  Socioeconomic (=2+) 1.171* 1.017–1.349 1.110 0.932–1.322 1.107 0.928–1.322  Risky parent b. (=1) 1.046 0.935–1.171 1.057 0.939–1.191 0.977 0.852–1.121  Risky parent b. (=2+) 1.063 0.927–1.217 1.072 0.930–1.235 1.022 0.849–1.230  Infectious disease (=1) 1.081 0.894–1.306 1.164 0.936–1.447 1.162 0.936–1.443  Infectious disease (=2+) 1.048 0.881–1.247 1.075 0.885–1.306 1.077 0.888–1.307  Impairment (=1) 1.015 0.896–1.151 0.962 0.840–1.101 0.968 0.846–1.107  Impairment (=2+) 1.084 0.865–1.358 1.134 0.880–1.462 1.147 0.888–1.480  Risky adolescent b. (=1) 1.058 0.931–1.204 1.049 0.910–1.209 1.049 0.912–1.206  Risky adolescent b. (=2+) 1.341* 1.033–1.740 1.265 0.948–1.687 1.239 0.924–1.663 Demographics  Age (years at baseline) 1.014† 1.009–1.019 1.017† 1.011–1.022 1.017† 1.011–1.022  Female 0.844‡ 0.761–0.935 0.873* 0.785–0.969 0.873* 0.786–0.971  Race/ethnicity (White = rf)   Black 1.263‡ 1.107–1.440 1.137 0.981–1.318 1.132 0.975–1.314   Hispanic 1.254‡ 1.089–1.443 1.166 0.990–1.375 1.179* 1.005–1.384 Adult resources and lifestyle  Marital status (married = rf)   Divorced 0.925 0.812–1.054 0.929 0.815–1.060   Widowed 1.011 0.873–1.170 1.001 0.866–1.158   Never married 1.097 0.906–1.328 1.102 0.909–1.337  Education (years) 1.012 0.992–1.033 1.014 0.994–1.034  Wealth (∛) 0.998* 0.997–1.000 0.998* 0.997–1.000  Insurance (none = rf)   Medicaid 1.152 0.839–1.581 1.203 0.891–1.623   Medicare/VA 1.096 0.876–1.371 1.093 0.874–1.366   Private/Medigap 1.022 0.858–1.218 1.023 0.857–1.222  Depressive symptoms 1.059† 1.034–1.084 1.059 1.034–1.084  Smoking (pack-years) 1.000 0.998–1.002 0.996* 0.992–1.000  Heavy drinker 1.365* 1.078–1.729 1.371‡ 1.089–1.727 BMI (<25 = rf)   25 ≤ BMI < 30 1.205† 1.096–1.324 1.206 1.097–1.325   BMI ≥ 30 1.611† 1.441–1.801 1.619 1.448–1.810  Physical activity 0.982* 0.966–0.998 0.982* 0.966–0.998 1 Risky parent × smoking 1.007‡ 1.002–1.012 2+ Risky parent × smoking 1.005 0.998–1.012 F 6.96† 16.36† 18.84† N 4,101 3,594 3,594 Note: b. = behavior; BMI = body mass index; CI = confidence interval; HR = hazard ratio; rf = reference group. aAnalyses weighted to adjust for clustering. bReference group is no misfortune. *p < .05; ‡p < .01, †p < .001. View Large Table 3. Cox Proportional Hazards Predicting Disease Incidence (2004–2012)a Model 1 Model 2 Model 3 HR 95% CI HR 95% CI HR 95% CI Childhood misfortuneb  Socioeconomic (=1) 1.105 0.946–1.292 1.118 0.946–1.323 1.114 0.940–1.322  Socioeconomic (=2+) 1.171* 1.017–1.349 1.110 0.932–1.322 1.107 0.928–1.322  Risky parent b. (=1) 1.046 0.935–1.171 1.057 0.939–1.191 0.977 0.852–1.121  Risky parent b. (=2+) 1.063 0.927–1.217 1.072 0.930–1.235 1.022 0.849–1.230  Infectious disease (=1) 1.081 0.894–1.306 1.164 0.936–1.447 1.162 0.936–1.443  Infectious disease (=2+) 1.048 0.881–1.247 1.075 0.885–1.306 1.077 0.888–1.307  Impairment (=1) 1.015 0.896–1.151 0.962 0.840–1.101 0.968 0.846–1.107  Impairment (=2+) 1.084 0.865–1.358 1.134 0.880–1.462 1.147 0.888–1.480  Risky adolescent b. (=1) 1.058 0.931–1.204 1.049 0.910–1.209 1.049 0.912–1.206  Risky adolescent b. (=2+) 1.341* 1.033–1.740 1.265 0.948–1.687 1.239 0.924–1.663 Demographics  Age (years at baseline) 1.014† 1.009–1.019 1.017† 1.011–1.022 1.017† 1.011–1.022  Female 0.844‡ 0.761–0.935 0.873* 0.785–0.969 0.873* 0.786–0.971  Race/ethnicity (White = rf)   Black 1.263‡ 1.107–1.440 1.137 0.981–1.318 1.132 0.975–1.314   Hispanic 1.254‡ 1.089–1.443 1.166 0.990–1.375 1.179* 1.005–1.384 Adult resources and lifestyle  Marital status (married = rf)   Divorced 0.925 0.812–1.054 0.929 0.815–1.060   Widowed 1.011 0.873–1.170 1.001 0.866–1.158   Never married 1.097 0.906–1.328 1.102 0.909–1.337  Education (years) 1.012 0.992–1.033 1.014 0.994–1.034  Wealth (∛) 0.998* 0.997–1.000 0.998* 0.997–1.000  Insurance (none = rf)   Medicaid 1.152 0.839–1.581 1.203 0.891–1.623   Medicare/VA 1.096 0.876–1.371 1.093 0.874–1.366   Private/Medigap 1.022 0.858–1.218 1.023 0.857–1.222  Depressive symptoms 1.059† 1.034–1.084 1.059 1.034–1.084  Smoking (pack-years) 1.000 0.998–1.002 0.996* 0.992–1.000  Heavy drinker 1.365* 1.078–1.729 1.371‡ 1.089–1.727 BMI (<25 = rf)   25 ≤ BMI < 30 1.205† 1.096–1.324 1.206 1.097–1.325   BMI ≥ 30 1.611† 1.441–1.801 1.619 1.448–1.810  Physical activity 0.982* 0.966–0.998 0.982* 0.966–0.998 1 Risky parent × smoking 1.007‡ 1.002–1.012 2+ Risky parent × smoking 1.005 0.998–1.012 F 6.96† 16.36† 18.84† N 4,101 3,594 3,594 Model 1 Model 2 Model 3 HR 95% CI HR 95% CI HR 95% CI Childhood misfortuneb  Socioeconomic (=1) 1.105 0.946–1.292 1.118 0.946–1.323 1.114 0.940–1.322  Socioeconomic (=2+) 1.171* 1.017–1.349 1.110 0.932–1.322 1.107 0.928–1.322  Risky parent b. (=1) 1.046 0.935–1.171 1.057 0.939–1.191 0.977 0.852–1.121  Risky parent b. (=2+) 1.063 0.927–1.217 1.072 0.930–1.235 1.022 0.849–1.230  Infectious disease (=1) 1.081 0.894–1.306 1.164 0.936–1.447 1.162 0.936–1.443  Infectious disease (=2+) 1.048 0.881–1.247 1.075 0.885–1.306 1.077 0.888–1.307  Impairment (=1) 1.015 0.896–1.151 0.962 0.840–1.101 0.968 0.846–1.107  Impairment (=2+) 1.084 0.865–1.358 1.134 0.880–1.462 1.147 0.888–1.480  Risky adolescent b. (=1) 1.058 0.931–1.204 1.049 0.910–1.209 1.049 0.912–1.206  Risky adolescent b. (=2+) 1.341* 1.033–1.740 1.265 0.948–1.687 1.239 0.924–1.663 Demographics  Age (years at baseline) 1.014† 1.009–1.019 1.017† 1.011–1.022 1.017† 1.011–1.022  Female 0.844‡ 0.761–0.935 0.873* 0.785–0.969 0.873* 0.786–0.971  Race/ethnicity (White = rf)   Black 1.263‡ 1.107–1.440 1.137 0.981–1.318 1.132 0.975–1.314   Hispanic 1.254‡ 1.089–1.443 1.166 0.990–1.375 1.179* 1.005–1.384 Adult resources and lifestyle  Marital status (married = rf)   Divorced 0.925 0.812–1.054 0.929 0.815–1.060   Widowed 1.011 0.873–1.170 1.001 0.866–1.158   Never married 1.097 0.906–1.328 1.102 0.909–1.337  Education (years) 1.012 0.992–1.033 1.014 0.994–1.034  Wealth (∛) 0.998* 0.997–1.000 0.998* 0.997–1.000  Insurance (none = rf)   Medicaid 1.152 0.839–1.581 1.203 0.891–1.623   Medicare/VA 1.096 0.876–1.371 1.093 0.874–1.366   Private/Medigap 1.022 0.858–1.218 1.023 0.857–1.222  Depressive symptoms 1.059† 1.034–1.084 1.059 1.034–1.084  Smoking (pack-years) 1.000 0.998–1.002 0.996* 0.992–1.000  Heavy drinker 1.365* 1.078–1.729 1.371‡ 1.089–1.727 BMI (<25 = rf)   25 ≤ BMI < 30 1.205† 1.096–1.324 1.206 1.097–1.325   BMI ≥ 30 1.611† 1.441–1.801 1.619 1.448–1.810  Physical activity 0.982* 0.966–0.998 0.982* 0.966–0.998 1 Risky parent × smoking 1.007‡ 1.002–1.012 2+ Risky parent × smoking 1.005 0.998–1.012 F 6.96† 16.36† 18.84† N 4,101 3,594 3,594 Note: b. = behavior; BMI = body mass index; CI = confidence interval; HR = hazard ratio; rf = reference group. aAnalyses weighted to adjust for clustering. bReference group is no misfortune. *p < .05; ‡p < .01, †p < .001. View Large When adjusting for adult resources and lifestyle factors in Model 2, the relationship between domains of childhood misfortune and disease incidence attenuated, such that the HRs for 2+ SES and risky adolescent misfortunes are no longer significant. As in Model 1, both age (HR = 1.017) and being female (HR = 0.873) were associated with disease incidence, but Black and Hispanic adults no longer differed from White adults. We tested interaction terms for race and childhood misfortune domains, but none were significant (not shown). Depressive symptoms (HR = 1.059), heavy drinking (HR = 1.365), and being overweight (HR = 1.205) or obese (HR = 1.611) as compared to normal weight increased the risk of incident disease; however, more wealth (HR = 0.998) and physical activity (HR = 0.982) decreased the risk of disease. Tests of Moderation To address our second research question, we estimated a series of models with product terms to assess the moderating effects of adult resources and lifestyle factors on the relationship between childhood misfortune and disease incidence. None of the product terms were significant in the fully adjusted prevalence model. The process was repeated for the incidence analyses, where we identified one significant interaction. As shown in Model 3 of Table 3, among those who were disease free in 2004, disease incidence by 2012 was higher for persons with one risky parental behavior and a high level of pack-years smoking (HR = 1.007). Most HRs remained stable from Model 2 except for the effect for Hispanic adults (HR = 1.179, p < .05). Finding that one risky parental behavior was significant, we subsequently tested product terms between each of the three indicators of risky parental behavior and pack-years. Analyses showed that parental smoking was the source of the interaction. Specifically, having a parent who smoked, in combination with the number of pack-years smoked in adulthood, increased the risk of disease onset by 2012 (HR = 1.006, p < .01; not shown). Discussion This research contributes to the literature on the early origins of adult health by identifying types of childhood misfortune associated with being free of any disease during middle and later life. Our first research question asked whether experiencing childhood misfortune decreased the odds of being disease free in later life. Results from logistic regression of baseline data revealed that experiencing childhood misfortune from several domains—socioeconomic, risky parental behavior, impairments, and/or risky adolescent behavior—reduces the odds of being disease free in later life. Even when adjusting for adult behaviors and resources, experiencing 2+ risky parental behaviors reduced the odds of being disease free. These findings add to the literature demonstrating the long-term influence of early misfortune. Growing up in a risky household, experiencing poor socioeconomic conditions, and/or exhibiting risky behavior in adolescence are the types of childhood misfortune most consequential to being disease free. This is consistent with conclusions of prior studies examining selected diseases among HRS respondents (O’Rand & Hamil-Luker, 2005) as well as those using the MIDUS (Schafer & Ferraro, 2012), Wisconsin Longitudinal Study (Springer, 2009), and Adverse Childhood Experiences (ACE) Study (Brown et al., 2010; Felitti et al., 1998). This study extends the early origins of health literature by identifying types of childhood misfortune that reduce the likelihood of being disease free. CI theory specifies pathways by which disadvantage may or may not accumulate, highlighting how the consequences of early insults are partly contingent on resource activation. Prior to including adult resources and lifestyles in our models, many domains of misfortune were negatively associated with being disease free. However, several of these findings were attenuated after including these adult factors, highlighting the importance of resources and human agency in shaping health trajectories. For instance, remaining physically active increases the likelihood of being disease free at baseline and remaining so over time, consistent with work by others (Barengo, Antikainen, Borodulin, Harald, & Jousilahti, 2017), despite experiencing considerable childhood misfortune. In addition, SES resources are central to overcoming the anticipated deleterious sequelae of early misfortune. In both the prevalence and incidence models, we observed that childhood SES misfortune was related to disease-free status but that these associations were no longer significant after adjusting for adult wealth. This suggests that some of the negative health consequences associated with financial strain or poverty during childhood can be offset by sufficient socioeconomic assets during adulthood. Our second research question asked if and to what extent adult lifestyle and resources moderate the relationship between childhood misfortune and developing disease over time. Our analyses revealed an interaction associated within the risky parental behavior domain: having a parent who smoked during one’s childhood—in combination with the number of pack-years smoked in adulthood—significantly increased the risk of developing disease over time. To our knowledge, no previous study of the early origins of adult health reports such an interaction. To illustrate the dose–response relationship when a parent smoked during one’s childhood, the risk of getting more or more disease rises 12% between a 40 pack-year smoker and a 60 pack-year smoker (i.e., 10 additional cigarettes daily for 40 years). This finding shines a light on the intergenerational and life course implications of tobacco consumption. Parental smoking may influence children’s health through environmental exposure to tobacco smoke (i.e., second-hand smoke), by modeling a lifestyle of tobacco use, and perhaps biologically by inducing symptoms of early nicotine dependence in the womb (Alberg & Korte, 2014). Any sustained exposure to tobacco smoke is harmful, but our analyses show the potency of lifelong exposure to smoking. Some mothers and fathers stop smoking during pregnancy, which is certainly beneficial, yet smoking resumption among parents after birth remains a serious threat to their progeny’s likelihood of disease-free aging. In addition to the general influence of one’s upbringing, the imprint of family lineage is clear (proposition 1.c, Ferraro & Shippee, 2009): risky parental behaviors have long-term consequences on the health of the next generation. Although future research may be able to probe more carefully how much of the family lineage effect is due to genetic and environmental influence, CI theory and the results of this study call for more attention to family lineage as a core element of life course epidemiology. In addition to our main findings, we briefly discuss those pertaining to gender and race. Gender did not predict being disease free at baseline, but predicted disease incidence with women less likely to develop disease. This finding stands in contrast to much of the literature that reports women are more likely to develop chronic diseases (Rieker & Bird, 2005). Although men and women suffer from similar diseases in later life, women live longer lives and more likely to suffer from more chronic, nonlife-threatening diseases (i.e., arthritis) that are not accounted for in our measure. Black adults had significantly lower odds of being disease free in 2004, even in the fully adjusted model. These prevalence results are consistent with much of the disparities literature showing that stark Black–White racial inequalities develop early in the life course, which may be understated in studies of older adults because of early-life mortality selection (Ferraro, Kemp, & Williams, 2017; Woolf & Braveman, 2011). By contrast, although both Black and Hispanic adults had higher risk of developing disease by 2012, these effects became nonsignificant after accounting for adult resources and lifestyle factors—confirming that these factors can ameliorate or exacerbate racial disparities, even during later life. These findings should be interpreted with limitations in mind. First, childhood misfortune was measured using retrospective data, which may involve recall bias. We attempted to reduce the effect of this bias by adjusting for depressive symptoms and socioeconomic resources as recommended by Vuolo, Ferraro, Morton, and Yang (2014). In addition, we restricted the sample to adults who fell within 2 SD of the mean cognition score. Nevertheless, results should be interpreted with care. Second, interpretation of the moderation findings regarding risky parental behaviors, particularly smoking, is limited because of the crudeness of the measure. Although it taps exposure to family tobacco use during childhood, there is no information on duration or quantity of smoking. Future studies that capture more detailed information on parental smoking may help identify the processes at work, including potential mediation. Third, self-reports of a number of variables used here can be limiting. For example, in a review of physical activity measures, Prince and colleagues (2008) found that correlations between self-report and direct measures of physical activity were only low to moderate, creating potential issues of reliability. Furthermore, self-reported, physician-diagnosed disease may underestimate prevalence and incidence due to limited access to medical care, particularly for uninsured and minority adults. Although we adjusted for insurance status and wealth, these three preceding limitations probably lead to a conservative bias, underestimating the influence of early origins on adult health. Finally, studies of disease-free aging have to grapple with the issue of which diseases should be included in one’s definition of this highly favorable health status. Although some include an exhaustive list of diseases and symptoms (see Schafer & Ferraro, 2012), we chose to focus on the leading causes of death in the United States. Empirical generalizations across studies need to consider the types of conditions examined. Conclusion Kuh and Ben-Shlomo (2016) call for more “study of common risk and protective factors” (p. 117). Knowledge about risk factors for specific diseases is vitally important; yet, the public is often confused by conflicting reports about health practices derived from “single-disease epidemiology.” Consequently, identifying risk factors associated with any of the leading causes of death is important for fundamental scientific and public understanding of health. Our findings show that risky parental behaviors, especially smoking, have long-term health consequences on their children. Perhaps more attention on the intergenerational transmission of health risks would aid smoking cessation programs. This study extends the findings of Schafer and Ferraro (2012), demonstrating the early origins of remaining disease-free in middle and later life. Family upbringing characterized by SES misfortune and risky parental behaviors place notable long-term health risks on the children. One might expect that the effect of childhood misfortune on disease onset would be limited to early and middle adulthood, but these results show elevated risk decades after the insults. We do not claim that avoiding a noxious home environment guarantees that children will be able to escape disease altogether, but avoiding or even delaying these killer diseases begins long before symptoms are identified. Our findings may seem disheartening; yet, it is valuable to remember that the trajectories set into motion by experiencing early adversity are not determined for life. Engaging in regular physical activity, abstaining from—or even reducing—tobacco consumption, and accumulating socioeconomic resources may help to interrupt or counteract the likely sequelae of childhood misfortune. Supplementary Material Supplementary data is available at The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences online. Funding This research was supported by a grant from the National Institute on Aging to K. Ferraro (AG043544). Conflict of Interest None reported. Acknowledgments M. M. Williams drafted the manuscript and performed the statistical analyses. B. R. Kemp assisted with the study’s design and analyses. K. F. Ferraro planned the study, supervised the data analysis, and wrote sections of the manuscript. S. A. Mustillo assisted with the study’s analysis and interpretation of results. All authors contributed to revising the manuscript for publication. References Alberg , A. , & Korte , J . ( 2014 ). Parental smoking as a risk factor for adult tobacco use: Can maternal smoking during pregnancy be distinguished from the social environmental influence during childhood ? American Journal of Epidemiology , 179 , 1418 – 1421 . doi: 10.1093/aje/kwu079 Google Scholar CrossRef Search ADS PubMed Barengo, N., Antikainen, R., Borodulin, K., Harald, K., & Jousilahti, P. (2017). Leisure-time physical activity reduces total and cardiovascular mortality and cardiovascular disease incidence in older adults. Journal of the American Geriatrics Society, 65, 504–510. doi:10.1111/jgs.14694 Blackwell , D. L. , Hayward , M. D. , & Crimmins , E. M . ( 2001 ). Does childhood health affect chronic morbidity in later life ? Social Science & Medicine , 52 , 1269 – 1284 . doi:10.1016/S0277-9536(00)00230-6 Google Scholar CrossRef Search ADS Breslow , N . ( 1974 ). Covariance analysis of censored survival data . Biometrics , 30 , 89 – 99 . doi:10.2307/2529620 Google Scholar CrossRef Search ADS PubMed Brown , D. W. , Anda , R. F. , Felitti , V. J. , Edwards , V. J. , Malarcher , A. M. , Croft , J. B. , & Giles , W. H . ( 2010 ). Adverse childhood experiences are associated with the risk of lung cancer: A prospective cohort study . BMC Public Health , 10 , 20 . doi: 10.1186/1471-2458-10-20 Google Scholar CrossRef Search ADS PubMed Critchley , J. A. , & Capewell , S . ( 2003 ). Mortality risk reduction associated with smoking cessation in patients with coronary heart disease: A systematic review . JAMA , 290 , 86 – 97 . doi: 10.1001/jama.290.1.86 Google Scholar CrossRef Search ADS PubMed Dannefer , D . ( 2003 ). Cumulative advantage/disadvantage and the life course: Cross-fertilizing age and social science theory . The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences , 58 , S327 – S337 . Google Scholar CrossRef Search ADS Dawson , D. A . ( 2011 ). Defining risk drinking . Alcohol Research and Health , 34 , 144 – 156 . Google Scholar PubMed Elder , G. H. , Jr . ( 1998 ). The life course as developmental theory . Child Development , 69 , 1 – 12 . Google Scholar CrossRef Search ADS PubMed Evert , J. , Lawler , E. , Bogan , H. , & Perls , T . ( 2003 ). Morbidity profiles of centenarians: Survivors, delayers, and escapers . The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences , 58 , 232 – 237 . Google Scholar CrossRef Search ADS Felitti , V. J. , Anda , R. F. , Nordenberg , D. , Williamson , D. F. , Spitz , A. M. , Edwards , V. ,… Marks , J. S . ( 1998 ). Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults. The Adverse Childhood Experiences (ACE) Study . American Journal of Preventive Medicine , 14 , 245 – 258 . Google Scholar CrossRef Search ADS PubMed Ferraro , K. F. , Kemp , B. R. , & Williams , M. M . ( 2017 ). Diverse aging and health inequality by race and ethnicity . Innovation in Aging , 1 , 1 – 11 . doi: 10.1093/geroni/igx002 Ferraro , K. F. , Schafer , M. H. , & Wilkinson , L. R . ( 2016 ). Childhood disadvantage and health problems in middle and later life: Early imprints on physical health ? American Sociological Review , 81 , 107 – 133 . doi: 10.1177/0003122415619617 Google Scholar CrossRef Search ADS PubMed Ferraro , K. F. , & Shippee , T. P . ( 2009 ). Aging and cumulative inequality: How does inequality get under the skin ? The Gerontologist , 49 , 333 – 343 . doi: 10.1093/geront/gnp034 Google Scholar CrossRef Search ADS PubMed Ferraro , K. F. , Shippee T. P. , & Schafer M. H . ( 2009 ). Cumulative inequality theory for research on aging and the life course . In V. L. Bengtson , D. G. Gans , N. M. Putney , & M. Silverstein (Eds.), Handbook of theories of aging (pp. 413 – 434 ). New York, NY : Springer . Heron , M . ( 2016 ). Deaths: Leading causes for 2014 . National Vital Statistics Report , 65 , 1 – 96 . Kemp , B. R. , Ferraro , K. F. , Morton , P. M. , & Mustillo , S. A . ( 2018 ). Early origins of adult cancer risk among men and women: Influence of childhood misfortune ? Journal of Aging and Health , 30 , 140 – 163 . doi: 10.1177/0898264316670049 Google Scholar CrossRef Search ADS Kuh , D. , & Ben-Shlomo , Y . ( 2016 ). Early life origins of adult health and ageing . In L. K. George & K. F. Ferraro (Eds.), Handbook of aging and the social sciences ( 8th ed ., pp. 101 – 122 ). San Diego, CA : Academic Press . Latham , K. , & Williams , M. M . ( 2015 ). Does neighborhood disorder predict recovery from mobility limitation? Findings from the health and retirement study . Journal of Aging and Health , 27 , 1415 – 1442 . doi: 10.1177/0898264315584328 Google Scholar CrossRef Search ADS PubMed Mood , C . ( 2010 ). Logistic regression: Why we cannot do what we think we can do, and what we can do about it . European Sociological Review , 26 , 67 – 82 . doi: 10.1093/esr/jcp006 Google Scholar CrossRef Search ADS Morton , P. M. , Mustillo , S. A. , & Ferraro , K. F . ( 2014 ). Does childhood misfortune raise the risk of acute myocardial infarction in adulthood ? Social Science & Medicine , 104 , 133 – 141 . doi: 10.1016/j.socscimed.2013.11.026 Google Scholar CrossRef Search ADS National Center for Health Statistics . ( 2015 ). 10 Leading causes of death by age group, United States, 2015 . Hyattsville, MD : Author . National Center for Health Statistics . ( 2016 ). Health, United States, 2015: With special feature on racial and ethnic health disparities . Hyattsville, MD : Author . O’Rand , A. M. , & Hamil-Luker , J . ( 2005 ). Processes of cumulative adversity: Childhood disadvantage and increased risk of heart attack across the life course . The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences , 60 , 117 – 124 . Google Scholar CrossRef Search ADS Pavela , G. , & Latham , K . ( 2016 ). Childhood conditions and multimorbidity among older adults . The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences , 71 , 889 – 901 . doi: 10.1093/geronb/gbv028 Google Scholar CrossRef Search ADS Pearlin , L. I . ( 1989 ). The sociological study of stress . Journal of Health and Social Behavior , 30 , 241 – 256 . Google Scholar CrossRef Search ADS PubMed Prince , S. A. , Adamo , K. B. , Hamel , M. E. , Hardt , J. , Connor Gorber , S. , & Tremblay , M . ( 2008 ). A comparison of direct versus self-report measures for assessing physical activity in adults: A systematic review . The International Journal of Behavioral Nutrition and Physical Activity , 5 , 56 . doi: 10.1186/1479-5868-5-56 Google Scholar CrossRef Search ADS PubMed Rieker , P. P. , & Bird , C. E . ( 2005 ). Rethinking gender differences in health: Why we need to integrate social and biological perspectives . The Journals of Gerontology. Series B: Psychological Sciences and Social Sciences , 60 , 40 – 47 . Google Scholar CrossRef Search ADS Rowe , J. W. , & Kahn , R. L . ( 1987 ). Human aging: Usual and successful . Science , 237 , 143 – 149 . Google Scholar CrossRef Search ADS PubMed Sayer , A. A. , Syddall , H. E. , Gilbody , H. J. , Dennison , E. M. , & Cooper , C . ( 2004 ). Does sarcopenia originate in early life? Findings from the Hertfordshire cohort study . The Journals of Gerontology. Series A: Biological Sciences and Medical Sciences , 59 , M930 – M934 . Google Scholar CrossRef Search ADS Schafer , M. H. , & Ferraro , K. F . ( 2012 ). Childhood misfortune as a threat to successful aging: Avoiding disease . The Gerontologist , 52 , 111 – 120 . doi: 10.1093/geront/gnr071 Google Scholar CrossRef Search ADS PubMed Schafer , M. H. , Ferraro , K. F. , & Mustillo , S. A . ( 2011 ). Children of misfortune: Early adversity and cumulative inequality in perceived life trajectories . American Journal of Sociology , 116 , 1053 – 1091 . Google Scholar CrossRef Search ADS Springer , K. W . ( 2009 ). Childhood physical abuse and midlife physical health: Testing a multi-pathway life course model . Social Science & Medicine , 69 , 138 – 146 . doi: 10.1016/j.socscimed.2009.04.011 Google Scholar CrossRef Search ADS Svanes , C. , Sunyer , J. , Plana , E. , Dharmage , S. , Heinrich , J. , Jarvis , D. ,… Antó , J. M . ( 2010 ). Early life origins of chronic obstructive pulmonary disease . Thorax , 65 , 14 – 20 . doi: 10.1136/thx.2008.112136 Google Scholar CrossRef Search ADS PubMed Thomas , C. , Hyppönen , E. , & Power , C . ( 2008 ). Obesity and type 2 diabetes risk in midadult life: The role of childhood adversity . Pediatrics , 121 , e1240 – e1249 . doi: 10.1542/peds.2007-2403 Google Scholar CrossRef Search ADS PubMed Umstattd Meyer , M. R. , Janke , M. C. , & Beaujean , A. A . ( 2014 ). Predictors of older adults’ personal and community mobility: Using a comprehensive theoretical mobility framework . The Gerontologist , 54 , 398 – 408 . doi: 10.1093/geront/gnt054 Google Scholar CrossRef Search ADS PubMed Von Hippel , P . ( 2007 ). Regression with missing Ys: An improved strategy for analyzing multiply imputed data . Sociological Methodology , 37 , 83 – 117 . doi: 10.1111/j.1467-9531.2007.00180.x Google Scholar CrossRef Search ADS Vuolo , M. , Ferraro , K. F. , Morton , P. M. , & Yang , T. Y . ( 2014 ). Why do older people change their ratings of childhood health ? Demography , 51 , 1999 – 2023 . doi: 10.1007/s13524-014-0344-3 Google Scholar CrossRef Search ADS PubMed Ward , B. W. , Schiller , J. S. , & Goodman , R. A . ( 2014 ). Multiple chronic conditions among US adults: A 2012 update . Preventing Chronic Disease , 11 , E62 . doi: 10.5888/pcd11.130389 Google Scholar PubMed Woolf , S. H. , & Braveman , P . ( 2011 ). Where health disparities begin: The role of social and economic determinants – And why current policies may make matters worse . Health Affairs (Project Hope) , 30 , 1852 – 1859 . doi: 10.1377/hlthaff.2011.0685 Google Scholar CrossRef Search ADS PubMed Xu , J. , Murphy , S. L. , Kochanek , K. D. , & Arias , E . ( 2016 ). Mortality in the United States, 2015 (NCHS Data Brief No. 267). Atlanta, GA : National Center for Health Statistics , 1 – 8 . © The Author(s) 2018. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

The Journals of Gerontology Series B: Psychological Sciences and Social SciencesOxford University Press

Published: Apr 3, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off