Childhood Conditions Predict Chronic Diseases and Functional Limitations Among Older Adults: The Case of Indonesia

Abstract

Objective: Limited knowledge exists about whether childhood health and socioeconomic status (SES) conditions influence health among older adults in Indonesia. Method: Data from Wave 5 (2014/2015) of the Indonesian Family Life Survey (N = 6,530) was used to examine associations between childhood health and SES and hypertension, lung conditions, diabetes, lower body functional limitations (LBFL), and instrumental activities of daily living limitations (IADL). Results: Poor childhood health was associated with 34% higher odds of hypertension (odds ratio [OR] = 1.34, p < .05), 37% higher odds of diabetes (OR = 1.37, p < .05), and 32% higher odds of lung conditions (OR = 1.32, p < .05). Household overcrowding was associated with 22% higher odds of a LBFL (OR = 1.22, p < .05) and 24% higher odds of an IADL (OR = 1.24, p < .01). Lacking a household toilet was associated with 18% higher odds of a LBFL (OR = 1.18, p < .05). Adjusting for adult SES and current health did not account for these relationships. Discussion: Childhood conditions appear to shape older adult health in Indonesia.

Keywords

early-life conditions disability health inequalities Indonesia

Introduction

There is a large body of literature linking early-life circumstances to health later in life. Older adults who experienced harsh conditions in childhood are at an increased risk for heart attacks (Hamil-Luker & O’Rand, 2007), report more functional limitations (Henretta & McCrory, 2016; Montez & Hayward, 2014), are more likely to suffer from chronic conditions such as hypertension (Blackwell, Hayward, & Crimmins, 2001; McEniry & McDermott, 2015), and have a higher probability of engaging in unhealthy behaviors such as smoking (Non et al., 2014). Early-life circumstances are thought to influence health in later life through biological and social pathways. On one hand, childhood conditions may directly impact later-life health through biological imprinting or scarring (Barker, 1997). On the other hand, childhood conditions may indirectly impact later-life health via social pathways such as hindering economic opportunities throughout the life course (Umberson, Williams, Thomas, Liu, & Thomeer, 2014).

The relationships between childhood conditions, defined as health and socioeconomic status (SES) conditions, and the likelihood of developing health conditions later in life have been well documented in developed countries. Less is known about the influence of childhood factors on chronic conditions and functional limitations among older adults in rapidly aging, less-developed countries such as Indonesia. Investigation of the long-term impacts of childhood conditions on older adult health in the developing world is needed for two key reasons. First, as less-developed countries undergo epidemiological transitions, it is important to understand how childhood conditions may contribute to their increasing chronic disease and physical disability burdens (McEniry & McDermott, 2015). In contrast to older adults in developed countries, those in less-developed countries are characterized by their endurance of poor early-life circumstances, including impoverished living conditions, malnutrition, and heightened exposure to infectious diseases. Thus, they are potentially more vulnerable to the impacts of these conditions at older ages. Second, from a demographic perspective, there are implications for population aging. Indonesia, one of the most populous countries in the world with more than 261 million people, is expected to contribute heavily to worldwide population aging. According to the World Health Organization (WHO), by 2050 more than 73 million people in Indonesia (20% of the population) will be 60 years and older (WHO, 2015). Using Wave 5 of the Indonesian Family Life Survey (IFLS), this study extends prior literature by examining the associations between childhood health and SES conditions and chronic conditions and functional limitations among Indonesian adults ages 50 and older.

Pathways Linking Childhood Conditions and Chronic Diseases

The long reach of childhood is well documented in the current literature, whereby early-life conditions are shown to impact later-life health outcomes (Elo & Preston, 1992; Ferraro, Schafer, & Wilkinson, 2016). Poor health in childhood may directly impact the likelihood of having chronic conditions later in life through biological scarring where organ systems suffer damage during a critical period in childhood (Barker, 1997; Ferraro & Shippee, 2009). Childhood respiratory infections (Barker, 1997; Chan et al., 2015), infectious diseases (Chan et al., 2015), and childhood malnutrition (van Abeelen et al., 2013) have been linked to lung conditions, hypertension, and other chronic diseases in adulthood in both developed and less-developed countries. Using the 1996 U.S. Health and Retirement Study (HRS), Blackwell et al. (2001) documented that poor childhood health was associated higher risks of developing lung and cardiovascular diseases, even after adjustment for adult SES and health risk behaviors. Similarly, using data from the Institute of Nutrition of Central America and Panama (INCAP) Longitudinal Study and the 2002 to 2004 Human Capital Study, Margolis (2010) found that childhood infectious diseases were associated with a higher risk of developing cardiovascular disease in Guatemala.

Childhood conditions may also affect the development of chronic conditions indirectly via social pathways such as limiting social and economic resources over the life course (Umberson et al., 2014). Using the Swedish Level of Living Survey and Study of Living Conditions of the Oldest Old, Berndt and Fors (2016) found that low SES in childhood was associated with lower educational attainment in adulthood, and that lower education then directly influenced the number of health problems later in life. This phenomenon is often referred to as “chains of risk,” where the health consequences of childhood conditions operate through a series of connected exposures throughout the life course (Ben-Shlomo & Kuh, 2002; Umberson et al., 2014). These chains of risk are thought to accumulate over the life course with each risk exposure increasing vulnerability to disease and death. However, it is important to consider the type, severity, and duration of exposure. Childhood conditions may place individuals on specific trajectories that increase their susceptibility to health problems as they age (Ben-Shlomo & Kuh, 2002). For example, using the 2005 to 2007 New England Family Study, Non et al. (2014) found that older adults who grew up in low-SES households were more likely to smoke cigarettes and consume excessive amounts of alcohol in adulthood. These risky health behaviors then placed them at a higher risk of cardiometabolic disease later in life (Non et al., 2014).

Pathways Linking Childhood Conditions and Functional Limitations

Childhood circumstances may also be directly associated with functional limitations through biological scarring in childhood or indirectly through chains of risk throughout the life course, including the development of chronic conditions. For example, malnutrition in childhood may negatively impact bone health, increasing the risk of osteoporotic fractures and raising the susceptibility of developing physical mobility limitations (Langley-Evans, 2015). In their study of childhood conditions and functional limitations in Mexico, Huang, Soldo, and Elo (2011) found that childhood malnutrition was associated with an increased risk of adult lower body functional limitation (LBFL). However, adjustment for education and occupation in adulthood largely attenuated the association between low childhood SES and developing an LBFL (Huang et al., 2011). Similarly, Henretta and McCrory (2016), using the Irish Longitudinal Study of Aging (TILDA), documented that while poor childhood health was directly associated with worse physical functioning later in life, high levels of education in adulthood mediated the impact of low childhood SES. Results from these studies suggest that it is important to separate childhood conditions into measures of health and family SES. Although poor childhood health and nutrition may have direct impacts on health due to biological scarring, poor family SES may influence later-life health indirectly by impacting SES in adulthood.

Other studies have emphasized that adult SES, typically measured as education, wealth, and/or income, may attenuate the associations between childhood SES and functional limitations. For example, using the 1998 to 2006 U.S. Health and Retirement Survey (HRS), Bowen and González (2010) found that adult SES completely mediated the association between low SES in childhood and having an instrumental activities of daily living (IADL) later in life. In another study using the 1998 to 2008 HRS, older adults who grew up in low-SES households, but achieved high levels of education had active life expectancies (i.e., years without functional limitations) that were similar to those adults who grew up in higher SES households (Montez & Hayward, 2014).Within the context of a less-developed country, Shen and Zeng (2014) documented that high adult SES fully mediated the impact of poor childhood SES on the odds of developing an IADL limitation among older adults in China.

Some studies have pointed to current health conditions as being in the pathway between childhood conditions and functional limitations (Elo, 2009; Huang et al., 2011). In their study examining Puerto Rico and seven cities in Latin America and the Caribbean, Monteverde, Noronha, and Palloni (2009) documented that childhood conditions influenced the risk of developing functional limitations indirectly through the pathway of chronic diseases. These chronic diseases were then associated with a higher likelihood of having functional limitations (Monteverde et al., 2009). However, the authors emphasized the need to be cautious in interpreting these results because they were unable to ascertain the order of the occurrence of chronic conditions and functional limitations (Monteverde et al., 2009).

The Indonesian Setting

Indonesia, the world’s fourth most populous country with more than 261 million people, has witnessed dramatic epidemiological and economic change over the past four decades. Gross national income per capita has risen from US$560 in 2000 to US$3,603 in 2016 (World Bank Group, 2017). Public health initiatives and programs have contributed to a notable increase in life expectancy at birth from 38 years in 1950 to 71 years in 2010 (Guilmoto & Jones, 2016). By 1997, more than half of the population had access to public health care, and by 2010 four fifths of the population had access to clean drinking water (Guilmoto & Jones, 2016). These rapid improvements have substantially improved childhood conditions for more recent birth cohorts, whereas older Indonesians were exposed to much higher rates of infectious diseases, malnutrition, and substandard living conditions (Guilmoto & Jones, 2016). Due to the types and severity of these childhood conditions, older Indonesians may be more susceptible to the influences of these conditions on health compared with older adults in more-developed countries.

Few studies have examined the influence of childhood health and SES on health outcomes among older adults in Indonesia. The IFLS included questions on childhood conditions for the first time in the most recent wave, which was fielded in 2014/2015. One prior study using the 2000 and 2007 waves of the IFLS found that poor early-life conditions, measured using adult height as a proxy, was positively associated with hypertension later in life (Sohn, 2017), suggesting that exposure to infectious diseases and nutritional deprivation in childhood may play an important role in the development of chronic diseases among older Indonesian adults. Indeed, rates of noncommunicable diseases are rising in Indonesia, with hypertension, diabetes, and chronic lung conditions being now in the Top 5 causes of death (WHO, 2015).

Contributions of the Current Study

As previously noted, there are relatively few studies examining the linkages between childhood conditions and adult health outcomes in rapidly aging, less-developed countries. The current study makes the following contributions to this literature. First, this study focuses on Indonesia, a populous, less-developed country that to date has received little attention in the literature connecting childhood conditions to health later in life. Second, this study takes advantage of the new module available in the most recent wave of the IFLS, which contains a rich set of questions about childhood health and SES conditions. Third, this study examines multiple health outcomes including hypertension, having a lung condition, diabetes, having a LBFL, and having an IADL to provide a more holistic picture of the influence of childhood circumstances on the health of older Indonesians.

Summary of Hypotheses

Based on the extant literature, this study examines the following hypotheses specific to childhood health and family SES.

Childhood Health

Childhood infectious diseases will be associated with greater odds of having hypertension, diabetes, and poor lung functioning, net of adult SES and current health factors.

Childhood hunger will be associated with greater odds of all three chronic conditions, a LBFL, and an IADL, net of adult SES and current health factors.

Childhood SES

Low childhood SES will be associated with greater odds of having all three chronic conditions, a LBFL and an IADL.

Adult SES and current health factors will largely account for any associations between childhood SES and the five distinct health outcomes, but not measures of childhood health.

Method

Data

Data comes from the IFLS, which is an ongoing longitudinal survey based on a sample of households representing 83% of the Indonesian population living in 13 of the nation’s 27 provinces. The first wave was conducted in 1993/1994 and collected data from more than 30,000 individuals living in more than 7,400 households. Follow-up waves were fielded in 1997, 2000, 2007, and 2014/2015. Response follow-up rates were high: nearly 95% in the second and third wave, and 91% in the fourth wave. Of the original respondents in the first wave, 82% were successfully recontacted for the fifth wave, IFLS5 (Strauss, Witoelar, & Sikoki, 2016). The sampling scheme stratified on provinces and urban/rural location, then randomly sampled within these strata (Strauss et al., 2016). Within each province, enumeration areas were randomly chosen from a nationally representative sample frame used in Indonesia’s 1993 National Socioeconomic Survey (SUSENAS). Within each enumeration area, households were randomly chosen based upon the 1993 SUSENAS listings. In each household, information was collected on the household head and his or her spouse, two randomly selected children of the head and spouse aged 0 to 14 years, a randomly selected individual age 50 years or older and his or her spouse, and then a randomly selected individual age 15 to 49 years and his or her spouse. IFLS5 marks a change from pencil and paper questionnaire to a computer-assisted personal interview (CAPI). IFLS collects detailed information on health status, labor force participation, family structure and mobility, migration, and various aspects of economic well-being. IFLS5 is the first wave to include questions on childhood SES and health.

The sample in this analysis was limited to Indonesian adults who were 50 years or older during the time of the interview. This is the typical age cutoff for studies on Indonesia because people tend to retire before age 60 (Guilmoto & Jones, 2016). For IFLS5, 7,594 adults aged 50 years and older completed the adult questionnaire. Questionnaires completed by proxies (n = 947) were excluded because proxies did not answer questions on childhood conditions. All cases with missing information were dropped (n = 117) after results from multiple imputation (MI) revealed that results from models that included and excluded missing values were nearly identical. The final sample size was 6,530 respondents, 3,067 males and 3,463 females.

Dependent Variables

Each respondent was asked whether a doctor ever told him or her that he or she had selected chronic conditions such as hypertension, diabetes, heart attack, stroke, lung condition, and so forth. The following conditions were selected for investigation in the present study: (a) hypertension, (b) chronic lung condition, and (c) diabetes. These three conditions were selected because they are among the Top 5 leading causes of death among older adults in Indonesia (WHO, 2015). These three conditions have also been linked to childhood conditions in previous literature (Blackwell et al., 2001). Other conditions, such as stroke, were not used due to a very small number of respondents reporting being diagnosed with these conditions.

Two dichotomous measures of functional limitations, having a LBFL and having an instrumental activities of daily living limitation (IADL) were also included in this study in an effort to capture multiple dimensions of health. LBFLs included physical mobility tasks such as climbing stairs and standing up from sitting in a chair. IADLs included cognition-related tasks such as balancing a checkbook and preparing a meal. These basic skills are necessary for performing routine tasks in daily life, including the ability to get out of the house, work, and participate in social activities (Verbrugge & Jette, 1994). Respondents were asked about eight LBFL items and six IADL items. Respondents who reported having difficulty with or were unable to do at least one of the eight LBFL activities were coded as having a LBFL. The same coding procedure was followed for having an IADL.

Independent Variables

The independent variables of interest were childhood (i.e., <16 years old) health and family SES. Two dichotomous measures of childhood health were employed: (a) having experienced an infectious disease in childhood and (b) going hungry in childhood. Previous studies have found associations between childhood infectious diseases and the development of hypertension, diabetes, lung conditions, and functional limitations later in life (Blackwell et al., 2001; Chan et al., 2015). Childhood hunger was chosen because malnutrition has been linked to a wide variety of adult health outcomes, including diabetes, hypertension, and functional limitations, in both developed and less-developed countries (Huang et al., 2011).

Three dichotomous measures of low childhood family SES conditions were employed: (a) no toilet in the family home, (b) no books in family home, and (c) whether the family home was overcrowded. These family living situation measures capture information about the SES of the participant when she or he was growing up. Lacking a toilet in the household was chosen because unsanitary public waste facilities exposes children to the risk of illness and the spread of disease through fecal matter (Fink, Günther, & Hill, 2014). Growing up in a home without any books was included as a measure of childhood SES to capture illiteracy and general level of education in the household (Guilmoto & Jones, 2016). Household overcrowding is typically calculated as the number of persons living in the household per the number of rooms (Galobardes, Shaw, Lawlor, Lynch, & Smith, 2006). Overcrowding is usually then defined two or more people per room (Galobardes et al., 2006). Given the large proportion of respondents in this sample (more than 30%) who reported more than three people per room in childhood, a cut-off of four people per room was used in the study. Children growing up in overcrowded households may receive fewer economic resources from their parents and have increased exposure to infectious diseases through shared living quarters (Fink et al., 2014; Galobardes et al., 2006). Although father’s occupation and mother’s education are typical measures of childhood SES, neither of these measures was a significant predictor of the outcomes in this sample, and thus were not included in the analyses. This may be the case because when the respondents were children educational opportunities were limited and a large portion of the sample was self-employed (Guilmoto & Jones, 2016).

Adult SES

Adult SES was measured by (a) education and (b) monthly household per capita expenditure. Education was measured using three categories: no education, completed primary school, and completed secondary school or above. Given that older Indonesian adults tend to reside in multigenerational households, personal income is typically not a good measure of current SES. Household expenditure is a better measure of household resources than current income, particularly in low-income rural settings where income may vary year to year due to weather, pests, and plant diseases (Lei, Smith, Sun, & Zhao, 2014). In the IFLS, the individual in the household most knowledgeable about household expenses was asked to report the monthly expenditure on a range of goods and services. Average consumption/expenditure per capita was then calculated as the sum of all the expenditure aggregates divided by the number of household members. Due to skewness, a logged version of the monthly household expenditure variable was used in the analysis.

Adulthood Health

Two health risk behaviors were included in the analyses: smoking behavior and body mass index (BMI). Smoking behavior included three categories: current smoker, former smoker, and never smoked. BMI included three categories: normal (BMI < 25), overweight (BMI ⩾ 25), and obese (BMI ⩾ 30). Another common risk factor, drinking, was not included because the IFLS does not ask questions regarding alcohol consumption due to the sensitivity of asking these questions. For the LBFL and IADL models, hypertension, having a lung condition, and diabetes were also included as three dichotomous variables.

Control Variables

Multiple other explanatory variables were included as controls, including age, age-squared, sex (1 = male), marital status (1 = currently married), and whether the respondent lives in an urban area.

Analytic Strategy

A series of empirical models were estimated to examine the relationship between childhood conditions and hypertension, having a lung condition, diabetes, having a LBFL, and having an IADL. This study used logistic regression models because all five outcomes were dichotomous variables. For each of the three chronic conditions, Model 1 included the five measures of childhood conditions while also controlling for age, age-squared, sex, marital status, and whether the respondent lived in an urban area. Model 2 included all of the previous variables in addition to the two adult SES measures of education and household expenditure. Model 3 included all of the variables in the previous models in addition to the two health behavior variables, smoking behavior and BMI. The analyses of having a LBFL and having an IADL followed a similar strategy except that Model 3 also included variables for hypertension, having a lung condition, and diabetes.

It is possible that retrospective reporting of childhood health conditions was influenced by the respondent’s current health status. As such, any relationship found between childhood and adult health conditions would be biased. It is useful to examine the internal consistency of the responses to the questions on overall health status and health problems in childhood. If the reporting of childhood health conditions is valid, then there should be a highly significant correlation between these two measures. The relationship between self-rated health in childhood (excellent, very good, good, fair, or poor) and whether the respondent experienced an infectious disease in childhood was evaluated using Pearson chi-square and log-likelihood chi-square tests (results not shown). The results of both tests were highly significant, which indicates a strong relationship between self-rated childhood health and whether the respondent reported an infectious disease in childhood. Thus, the respondents appeared to be consistent in their reporting of childhood health conditions.

Results

Sample Characteristics

Table 1 presents the sample characteristics. Approximately 26% of the sample had been diagnosed with hypertension, 8% diagnosed with a chronic lung condition, and 6% with diabetes. About 42% had at least one LBFL and 21% had at least one IADL. Concerning childhood conditions, about 17% reported going hungry in childhood, 23% had an infectious disease, 52.27% grew up in a household without a toilet, 19% grew up in overcrowded households, and 89% grew up in a household without any books. Regarding adult SES, approximately 14% had no education, 55% completed primary school, and 31% completed secondary school or more. The nonlogged mean household expenditure was about 418,109 Indonesian rupiah (about US$30) per household member per month. Among the health risk factors, 60% had never smoked, 9% had quit smoking, and 31% were current smokers. Furthermore, 66% had a normal BMI while 26% were overweight and 8% were obese. About 47% of the sample was male and the average age of the sample was 60 years old. Three quarters of the sample was married and about 56% of the sample lived in an urban area.

Table 1.

Sample Characteristics for Indonesian Adults 50 and Older, Indonesian Family Life Survey Wave 5 (N = 6,530).

Characteristics	%
Adult health outcomes
Hypertension diagnosis	26.11
Lung condition diagnosis	8.07
Diabetes diagnosis	6.40
Lower body functional limitation (LBFL)	41.99
Instrumental activity of daily living limitation (IADL)	20.92
Demographic characteristics
Age, M (SD)	59.84 (8.08)
Male	46.97
Currently married	74.98
Lives in an urban area	55.82
Childhood health
Experienced hunger	16.83
Had infectious disease	23.37
Childhood socioeconomic status
No toilet at age 12	52.27
House overcrowded (>4 people per room) at age 12	18.58
Home did not have books at age 12	89.36
Adult socioeconomic status
Respondent’s education
No formal education	13.52
Primary school	55.30
Secondary school or more	31.18
Household per capita expenditure logged, M (SD)	12.22 (1.07)
Household per capita expenditure not logged, M (SD)	418,108.70 (1,170,351)
Health behaviors
Smoking behavior
Never smoked	59.89
Former smoker	9.33
Current smoker	30.78
Body mass index (BMI)
Normal weight	66.14
Overweight	25.87
Obese	7.99

Hypertension

Table 2 presents odds ratio (OR) estimates (and 95% confidence intervals) for hypertension. Model 1 includes all of the childhood conditions adjusted for demographic characteristics. Only one measure of childhood health, infectious disease, was significantly associated with higher odds of hypertension (OR = 1.43 p < .001). This association remained significant in Model 2 after adjustment for adult SES, that is, educational attainment and household per capita expenditure. In addition, household overcrowding (OR = 1.17, p = .037) and lacking books in the home (OR = 1.29, p = .010) during childhood became significantly associated with higher odds of hypertension in Model 2. These suppression effects suggest that low childhood SES was negatively associated with SES in adulthood, but adulthood SES was positively associated with the likelihood of having hypertension. With the introduction of the health risk behaviors in Model 3, childhood infectious disease, household overcrowding, and growing up in a household without books all remained significantly associated with higher odds of hypertension.

Table 2.

ORs From Logistic Regression for Hypertension, Ages 50 and Above, Indonesian Family Life Survey Wave 5 (N = 6,530).

Characteristics	Model 1		Model 2		Model 3
Characteristics	OR	95% CI	OR	95% CI	OR	95% CI
Demographics
Age	1.19**	[1.08, 1.30]	1.17**	[1.07, 1.28]	1.18**	[1.07, 1.29]
Age-squared	1.00**	[0.99, 1.00]	1.00**	[0.99, 1.00]	1.00**	[0.99, 1.00]
Male	0.50**	[0.44, 0.56]	0.47**	[0.41, 0.53]	0.59**	[0.49, 0.71]
Currently married (ref: not married)	1.06	[0.92, 1.22]	1.04	[0.91, 1.20]	0.99	[0.86, 1.14]
Lives in an urban area (ref: rural area)	1.28**	[1.14, 1.43]	1.16*	[1.03, 1.31]	1.09	[0.95, 1.22]
Childhood health
Experienced hunger	0.98	[0.84, 1.15]	1.01	[0.87, 1.19]	1.03	[0.88, 1.21]
Had infectious disease	1.43**	[1.26, 1.63]	1.37**	[1.20, 1.56]	1.34**	[1.17, 1.52]
Childhood socioeconomic status
No toilet at age 12	0.91	[0.81, 1.02]	0.97	[0.86, 1.10]	0.98	[0.87, 1.11]
House overcrowded (>4 people per room) at age 12	1.13^†	[0.98, 1.31]	1.16*	[1.00, 1.34]	1.16*	[1.00, 1.34]
Home did not have books at age 12	1.20^†	[0.99, 1.45]	1.29*	[1.06, 1.57]	1.27*	[1.05, 1.55]
Adult socioeconomic status
Education (ref: none)
Primary school			1.29**	[1.08, 1.56]	1.28*	[1.06, 1.54]
Secondary school and above			1.49**	[1.20, 1.86]	1.39**	[1.11, 1.73]
Household per capita expenditure (logged)			1.11**	[1.05, 1.17]	1.06*	[1.00, 1.13]
Health behaviors
Smoking behaviors (never)
Former smoker					1.50**	[1.19, 1.89]
Current smoker					0.69**	[0.56, 0.84]
Body mass index (normal)
Overweight					1.49**	[1.30, 1.70]
Obese					1.56**	[1.59, 2.38]
Constant	0.00**	[0.00, 0.02]	0.00**	[0.00, 0.01]	0.00**	[0.00, 0.01]
Log-likelihood	−3,631.09		−3,614.05		−3,554.83

Note. OR = odds ratio; CI = confidence interval.

†

p < .10. *p < .05. **p < .01.

Lung Condition

Table 3 presents OR estimates for having a lung condition. Consistent with the previous models for hypertension, only childhood infectious disease was associated with an increased risk of having a lung condition in Model 1 (OR = 1.24, p = .036). Adjustment for adult SES in Model 2 attenuated the association between childhood infectious disease and having a lung condition; and it was no longer significant. Higher education and household expenditure were strongly associated with higher odds of having a lung condition, suggesting that childhood infectious disease may be influencing the odds of having a lung condition through adult SES. Childhood hunger became significantly associated with higher odds of having a lung condition in Model 2 (OR = 1.31, p = .024). This suppression effect suggests that going hungry in childhood was negatively associated with SES in adulthood, but that higher SES in adulthood was associated with increased odds of having a lung condition. In Model 3, childhood hunger remained a significant predictor after adjusting for health risk behaviors (OR = 1.32; p = .020).

Table 3.

ORs From Logistic Regression for Having a Lung Condition, Ages 50 and Above, Indonesian Family Life Survey Wave 5 (N = 6,530).

Characteristics	Model 1		Model 2		Model 3
Characteristics	OR	95% CI	OR	95% CI	OR	95% CI
Demographics
Age	1.16^†	[1.00, 1.34]	1.15^†	[0.99, 1.33]	1.14	[0.98, 1.32]
Age-squared	1.00^†	[0.99, 1.00]	1.00	[0.99, 1.00]	1.00	[0.99, 1.00]
Male	0.89	[0.73, 1.08]	0.83^†	[0.68, 1.01]	0.82	[0.61, 1.10]
Currently married (ref: not married)	1.10	[0.88, 1.38]	1.08	[0.86, 1.36]	1.06	[0.84, 1.33]
Lives in an urban area (ref: rural area)	1.40**	[1.16, 1.69]	1.23*	[1.01, 1.50]	1.19^†	[0.97, 1.45]
Childhood health
Experienced hunger	1.25^†	[0.99, 1.57]	1.31*	[1.04, 1.65]	1.32*	[1.05, 1.67]
Had infectious disease	1.24*	[1.01, 1.52]	1.17	[0.95, 1.44]	1.13	[0.92, 1.39]
Childhood socioeconomic status
No toilet at age 12	0.90	[0.75, 1.09]	0.99	[0.81, 1.20]	0.99	[0.82, 1.20]
House overcrowded (>4 people per room) at age 12	1.06	[0.84, 1.33]	1.10	[0.87, 1.38]	1.07	[0.85, 1.35]
Home did not have books at age 12	0.94	[0.71, 1.25]	1.04	[0.78, 1.39]	1.02	[0.77, 1.37]
Adult socioeconomic status
Education (ref: none)
Primary school			1.20	[0.87, 1.64]	1.19	[0.87, 1.63]
Secondary school and above			1.55*	[1.08, 2.21]	1.46*	[1.02, 2.10]
Household per capita expenditure (logged)			1.16**	[1.06, 1.27]	1.15**	[1.04, 1.26]
Health behaviors
Smoking behaviors (never)
Former smoker					2.37**	[1.72, 3.25]
Current smoker					0.65**	[0.47, 0.89]
Body mass index (normal)
Overweight					0.89	[0.72, 1.11]
Obese					1.11	[0.81, 1.53]
Constant	0.00**	[0.00, 0.06]	0.00 **	[0.00, 0.01]	0.00**	[0.00, 0.02]
Log-likelihood	−1,814.67		−1,802.07		−1,765.82

Note. OR = odds ratio; CI = confidence interval.

†

p < .10. *p < .05. **p < .01.

Diabetes

In Table 4, consistent with Model 1 for hypertension and having a lung condition, childhood infectious disease was associated with a higher risk of diabetes (OR = 1.55, p < .001). This association was slightly attenuated, but remained significant in Model 2 after adjustment for adult SES (OR = 1.40; p = .003). Following the pattern in Model 2 for hypertension and having a lung condition, both measures of adult SES were also strongly associated with higher odds of diabetes. Taken together, these results suggest that adult SES may be partially mediating the influence of childhood infectious disease on diabetes. The positive association between childhood infectious disease and diabetes remained robust with the introduction of health behaviors in Model 3 (OR = 1.37, p = .005).

Table 4.

ORs From Logistic Regression for Diabetes, Ages 50 and Above, Indonesian Family Life Survey Wave 5 (N = 6,530).

Characteristics	Model 1		Model 2		Model 3
Characteristics	OR	95% CI	OR	95% CI	OR	95% CI
Demographics
Age	1.84**	[1.46, 2.32]	1.81**	[1.44, 2.29]	1.82**	[1.44, 2.30]
Age-squared	1.00**	[0.99, 1.00]	1.00**	[0.99, 1.00]	1.00**	[0.99, 1.00]
Male	0.96	[0.78, 1.20]	0.86	[0.69, 1.07]	1.02	[0.75, 1.39]
Currently married (ref: not married)	1.28^†	[0.99, 1.67]	1.22	[0.94, 1.59]	1.19	[0.91, 1.56]
Lives in an urban area (ref: rural area)	2.55**	[2.01, 3.23]	2.04**	[1.60, 2.61]	1.96**	[1.54, 2.51]
Childhood health
Experienced hunger	0.98	[0.75, 1.28]	1.08	[0.82, 1.42]	1.10	[0.83, 1.45]
Had infectious disease	1.55**	[1.25, 1.92]	1.40**	[1.13, 1.75]	1.37**	[1.10, 1.71]
Childhood socioeconomic status
No toilet at age 12	0.88	[0.71, 1.08]	1.05	[0.84, 1.30]	1.03	[0.83, 1.28]
House overcrowded (>4 people per room) at age 12	0.88	[0.67, 1.16]	0.95	[0.72, 1.25]	0.94	[0.71, 1.24]
Home did not have books at age 12	0.94	[0.69, 1.27]	1.13	[0.83, 1.55]	1.10	[0.80, 1.50]
Adult socioeconomic status
Education (ref: none)
Primary school			1.39	[0.90, 2.15]	1.39	[0.90, 2.15]
Secondary school and above			2.17**	[1.36, 3.46]	2.06**	[1.29, 3.29]
Household per capita expenditure (logged)			1.31**	[1.18, 1.44]	1.29*	[1.17, 1.43]
Health behaviors
Smoking behaviors (never)
Former smoker					1.24	[0.86, 1.79]
Current smoker					0.65*	[0.46, 0.90]
Body mass index (normal)
Overweight					1.21	[0.96, 1.52]
Obese					0.92	[0.63, 1.33]
Constant	0.00**	[0.00, 0.00]	00.0 **	[0.00, 0.00]	0.00**	[0.00, 0.00]
Log-likelihood	−1,482.96		−1,451.22		−1,441.54

Note. OR = odds ratio; CI = confidence interval.

†

p < .10. *p < .05. **p < .01.

LBFL

Results in Table 5 show that in contrast to the models of chronic conditions, childhood infectious disease was not a significant predictor of having a LBFL. However, childhood hunger (OR = 1.24; p = .003), household overcrowding (OR = 1.24; p = .002), and growing up in a home without a toilet (OR = 1.19; p = .002) were all associated with an increased risk of having a LBFL. Adjustment for adult SES in Model 2 had little impact on the risks associated with the above three childhood conditions. None of the adult SES measures were significantly associated with having a LBFL, suggesting that education and household expenditure were not in the pathway between childhood conditions and having a LBFL. In Model 3, the coefficients for childhood hunger, household overcrowding, and lacking a toilet remained robust compared with those in Model 2, implying that current health factors did not account for the associations between these childhood conditions and having a LBFL.

Table 5.

ORs From Logistic Regression for Having a Lower Body Functional Limitation, Ages 50 and Above, Indonesian Family Life Survey Wave 5 (N = 6,530).

Characteristics	Model 1		Model 2		Model 3
Characteristics	OR	95% CI	OR	95% CI	OR	95% CI
Demographics
Age	0.99	[0.91, 1.08]	0.99	[0.91, 1.08]	0.96	[0.88, 1.05]
Age-squared	1.00	[0.99, 1.00]	1.00	[0.99, 1.00]	1.00*	[1.00, 1.00]
Male	0.37**	[0.33, 0.41]	0.37**	[0.33, 0.42]	0.35**	[0.30, 0.42]
Currently married (ref: not married)	1.01	[0.89, 1.15]	1.01	[0.89, 1.15]	0.99	[0.86, 1.13]
Lives in an urban area (ref: rural area)	0.91^†	[0.82, 1.02]	0.94	[0.84, 1.05]	0.88*	[0.79, 0.99]
Childhood health
Experienced hunger	1.24**	[1.07, 1.42]	1.22**	[1.06, 1.41]	1.21*	[1.05, 1.39]
Had infectious disease	1.05	[0.93, 1.19]	1.06	[0.93, 1.19]	1.00	[0.88, 1.14]
Childhood socioeconomic status
No toilet at age 12	1.19**	[1.07, 1.33]	1.16**	[1.04, 1.30]	1.18**	[1.06, 1.33]
House overcrowded (>4 people per room) at age 12	1.24**	[1.09, 1.42]	1.23**	[1.08, 1.41]	1.22**	[1.07, 1.40]
Home did not have books at age 12	1.14	[0.96, 1.36]	1.10	[0.92, 1.31]	1.09	[0.91, 1.30]
Adult socioeconomic status
Education (ref: none)
Primary school			1.17^†	[0.99, 1.38]	1.13	[0.96, 1.34]
Secondary school and above			1.00	[0.82, 1.22]	0.92	[0.75, 1.12]
Household per capita expenditure (logged)			0.96	[0.91, 1.01]	0.92**	[0.87, 0.97]
Health behaviors
Smoking behaviors (never)
Former smoker					1.36**	[1.09, 1.71]
Current smoker					1.18^†	[0.99, 1.41]
Body mass index (normal)
Overweight					1.16*	[1.02, 1.32]
Obese					1.53**	[1.25, 1.87]
Hypertension diagnosis					1.46**	[1.30, 1.65]
Lung condition diagnosis					1.84**	[1.52, 2.24]
Diabetes diagnosis					1.70**	[1.36, 2.11]
Constant	0.22	[0.01, 3.37]	0.40	[0.02, 6.56]	1.20	[0.07, 20.78]
Log-likelihood	−4,109.04		−4,102.86		−4,025.01

Note. OR = odds ratio; CI = confidence interval.

†

p < .10. *p < .05. **p < .01.

IADL

Table 6 presents OR estimates for having an IADL. In Model 1, childhood hunger (OR = 1.33; p < .001), household overcrowding (OR = 1.28; p < .001), and growing up in a home without any books (OR = 1.50; p = .001) were associated with increased odds of having an IADL. Although education and household expenditure were significantly associated with lower odds of having an IADL in Model 2, the inclusion of these adult SES measures had little impact on the associations between the above three childhood conditions and the risk of having an IADL. Similar to the LBFL results, the coefficients for childhood hunger, household overcrowding, and lacking books in the home remained robust in Model 3, suggesting that health behaviors and chronic conditions did not account for the associations between the childhood SES and health measures and having an IADL.

Table 6.

ORs From Logistic Regression for Having an Instrumental Activity of Daily Living Limitation, Ages 50 and Above, Indonesian Family Life Survey Wave 5 (N = 6,530).

Characteristics	Model 1		Model 2		Model 3
Characteristics	OR	95% CI	OR	95% CI	OR	95% CI
Demographics
Age	0.95	[0.87, 1.04]	0.96	[0.88, 1.05]	0.94	[0.86, 1.02]
Age-squared	1.00*	[1.00, 1.00]	1.00^†	[1.00, 1.00]	1.00*	[1.00, 1.00]
Male	1.24**	[1.08, 1.42]	1.30**	[1.13, 1.49]	1.14	[0.93, 1.40]
Currently married (ref: not married)	0.98	[0.84, 1.15]	0.99	[0.85, 1.16]	0.99	[0.84, 1.15]
Lives in an urban area (ref: rural area)	0.83**	[0.73, 0.94]	0.91	[0.80, 1.03]	0.88^†	[0.77, 1.00]
Childhood health
Experienced hunger	1.33**	[1.14, 1.56]	1.31**	[1.12, 1.53]	1.28**	[1.09, 1.50]
Had infectious disease	0.97	[0.84, 1.13]	1.02	[0.88, 1.18]	0.98	[0.84, 1.14]
Childhood socioeconomic status
No toilet at age 12	1.04	[0.92, 1.19]	0.99	[0.57, 1.13]	1.00	[0.88, 1.15]
House overcrowded (>4 people per room) at age 12	1.28**	[1.10, 1.48]	1.25**	[1.08, 1.45]	1.24**	[1.06, 1.44]
Home did not have books at age 12	1.50**	[1.19, 1.89]	1.41**	[1.12, 1.77]	1.41**	[1.11, 1.78]
Adult socioeconomic status
Education (ref: none)
Primary school			0.85^†	[0.71, 1.02]	0.82*	[0.68, 0.98]
Secondary school and above			0.76*	[0.61, 0.95]	0.71**	[0.59, 0.89]
Household per capita expenditure (logged)			0.87**	[0.82, 0.93]	0.85**	[0.80, 0.91]
Health behaviors
Smoking behaviors (never)
Former smoker					1.50**	[1.17, 1.91]
Current smoker					1.27*	[1.03, 1.55]
Body mass index (normal)
Overweight					0.97	[0.83, 1.13]
Obese					1.27^†	[0.99, 1.61]
Hypertension diagnosis					1.36**	[1.18, 1.57]
Lung condition diagnosis					1.75**	[1.42, 2.15]
Diabetes diagnosis					1.41**	[1.10, 1.80]
Constant	0.22	[0.01, 3.69]	1.24	[0.07, 23.28]	3.30	[0.17, 65.17]
Log-likelihood	−3,229.68		−3,214.65		−3,173.29

Note. OR = odds ratio; CI = confidence interval.

†

p < .10. *p < .05. **p < .01.

Discussion

Childhood health and SES were found to be significant predictors of multiple health outcomes among older adults in Indonesia, a rapidly aging, less-developed country. These results were consistent with findings from several studies on developed and less-developed countries that have linked childhood conditions to later-life chronic conditions and functional limitations (e.g., Henretta & McCrory, 2016; Shen & Zeng, 2014). The results also partially supported the first proposed hypothesis, as childhood infectious disease was found to be associated with increased odds of hypertension and diabetes, net of adult SES and current health factors. These findings may reflect biological scarring, in which childhood infectious diseases are indicative of physiological changes to the body that put people at higher risks of chronic conditions later in life (Barker, 1997). Margolis (2010) found similar evidence in Guatemala, as children who experienced infectious diseases matured into adults who had higher risks of developing cardiovascular disease, regardless of their achieved SES in adulthood. These results may also reflect chains of risk throughout the life course in Indonesia (Umberson et al., 2014). Children who were sick in childhood may have consequently experienced stunted educational achievement in adulthood that then influenced their chances of developing hypertension and diabetes.

Interestingly, childhood infectious disease was not a significant predictor of lung health. The results may have varied across the chronic conditions because while respondents in the IFLS were asked whether they had experienced an infectious disease in childhood, but they were not asked about specific diseases. Some childhood infectious diseases raise the risk of developing certain chronic conditions, but not others. For example, rheumatic heart fever has been linked to adult heart disease (Elo & Preston, 1992), while lower respiratory tract infections have been linked to chronic lung conditions (Chan et al., 2015). The older Indonesians in this sample may have had childhood infectious diseases that have been linked to hypertension and diabetes, but not to chronic lung conditions.

Partial support was also found for the second hypothesis, as childhood hunger was associated with increased risks of poor lung functioning, and developing a LBFL and an IADL, even after adjustment for adult SES and current health factors. This link between childhood hunger and lung health was consistent with results from a previous study that found evidence that older adults who experienced malnutrition as children during the 1944-1945 Dutch Famine were more likely to develop chronic obstructive pulmonary disease (van Abeelen et al., 2013). Furthermore, consistent with findings from Huang et al.’s (2011) study on older adults in Mexico, childhood malnutrition was positively associated with developing a functional limitation in Indonesia, suggesting that the influence of childhood nutritional deprivation may extend to adulthood and increase the risk of physical disability at older ages.

It was further hypothesized that low childhood SES would be associated with greater odds of all five distinct health outcomes, and that adult SES and current health factors would largely account for these associations. This was hypothesized due to a large body of literature documenting that adult SES (Bowen & González, 2010; Henretta & McCrory, 2016; Montez & Hayward, 2014) and chronic conditions (Monteverde et al., 2009) may partially or fully mediate the associations between childhood SES and functional limitations. Although low childhood SES was associated with higher odds of having hypertension, a LBFL, and an IADL, little support was found for the hypothesis that adult SES and current health factors would largely account for these associations. The strength of these associations remained robust after the addition of the adult SES measures, health risk behaviors, and chronic conditions, suggesting that childhood SES directly influenced hypertension and functional health among older adults in Indonesia.

These results differed from those of Bowen and González (2010), who found that adult SES completely accounted for the association between childhood SES and having an IADL among older U.S. adults. The dissimilarities in these results may reflect differences in the power of adult SES to attenuate the impact of poor childhood SES on functional health in Indonesia compared with the United States. Prior research has documented that the United States has a steep SES-health gradient, whereby older adults with higher SES experience lower mortality and have lower risks of developing health problems (Elo, 2009). However, a recent study found that the SES-health gradient in Indonesia is relatively flat, and that SES differences in adult health are smaller in Indonesia compared with the United States (Sudharsanan, 2017). Results from the present study supported this recent finding from Indonesia, as higher adult SES was associated with decreased odds of having an IADL, but the strength of this association was not strong enough to outweigh the health consequences of poor SES experienced in childhood.

Furthermore, that current health factors did not account for the associations between childhood SES and having a LBFL and an IADL in Indonesia was inconsistent with results that Monteverde et al. (2009) found in Latin America and the Caribbean. Childhood conditions were measured differently in Monteverde et al. (2009) compared with the present study. Monteverde et al. (2009) only used one dichotomous variable that was equal to one if the respondent experienced poor childhood SES or episodes of poor health in childhood. These episodes of poor childhood health were defined as being confined to a bed for a month or more, or being sick often, or experiencing at least one childhood infectious disease (Monteverde et al., 2009). Using this combined measure of childhood SES and health conditions overlooked that different measures of childhood conditions may have dissimilar impacts on functional health later in life. By separating childhood conditions in health and SES measures, the present study was able to show how different measures of childhood conditions were associated with some health outcomes later in life, but not others. Expressly, results from the present study showed that childhood infectious disease was associated with hypertension and diabetes, but not with either of the functional limitation measures. Including childhood infectious diseases in their measure of childhood conditions may have rendered Monteverde et al.’s (2009) measure more predictive of chronic conditions than of functional limitations.

Somewhat unanticipated was the finding that low childhood SES and high adult SES were both associated with increased odds of hypertension. For example, growing up in an overcrowded household was used to indicate low SES in childhood. Thus, it was unsurprising that this measure was associated with increased odds of developing hypertension, a LBFL, and an IADL. However, education and household expenditure were both positively associated with hypertension. Although these results may seem counterintuitive, perhaps higher SES Indonesians had the knowledge and resources to see doctors and be diagnosed with hypertension. In their study of the awareness of hypertension in less-developed countries, Chow et al. (2013) found that less education was associated with lower rates of awareness and diagnosis of hypertension. Poorer, less-educated Indonesians in this sample may actually have had hypertension, diabetes, or a lung condition, but lacked formal diagnoses.

Limitations

First, since this study is cross-sectional, it does not draw any conclusions about causality, especially with regard to measures of current health status. For example, it may be the case that hypertension, diabetes, and having a lung condition are in the pathways between childhood conditions and LBFLs and IADLs, or that LBFLs and IADLs are in the pathway between childhood conditions and the three chronic conditions. Second, there may be response bias due to using retrospectively reported measures of childhood conditions. Although some studies (Haas, 2007; Krieger, Okamoto, & Selby, 1998; Smith, 2009) have found evidence that retrospective childhood reported are reasonably reliable, other studies (i.e., Batty, Lawlor, Macintyre, Clark, & Leon, 2005) have found that retrospective reporting may underestimate the real effects. Third, it is important to consider that mortality selection may have biased the results. Concerning the Indonesians in this sample, surviving an infectious disease in childhood to age 50 or older suggests that the health of respondents was robust enough to survive through adulthood and into older ages. Finally, these results may not be generalizable to other less-developed countries beyond Indonesia. Older adults in other less-developed countries may have had different types and severity of childhood conditions that may or may not have shaped their health in different ways.

Conclusion

This research underscored the enduring influence of childhood health and SES on health later in life. These findings have important implications for policy, especially in less-developed countries that are increasingly contending with population aging. Specifically, the present study found that children who have poor health and malnutrition were more likely to develop chronic conditions and functional limitations later in life. Based on this finding, programs that subsidize food and medical care may help children born into low-SES families avoid excessive risk of poor health in adulthood. One such program, Indonesia’s Midwife in the Village Program, improved the health of children by placing more than 50,000 midwives into impoverished areas of the country from 1989 to 1998 to provide services such as preventative care, nutrition counseling, and immunizations. Several studies (i.e., Frankenberg, Suriastini, & Thomas, 2005) have documented the success of this program in improving children’s health. The present study provided additional support for such programs because health interventions in childhood may have long-lasting impacts on later-life health. Furthermore, for Indonesians who are already middle-aged and older, it is important to ensure access to high-quality health care to contend with the potential health consequences of poor childhood conditions. The Indonesian government unveiled an ambitious plan in 2013 to provide universal health care for all citizens by 2019 (Fossati, 2017). Although it is not yet clear whether this goal will be achieved, access to medical care is becoming increasingly available and accessible in Indonesia (Fossati, 2017). Given the deleterious childhood circumstances experienced by Indonesia’s aging population, it is vital to provide health care to address potential later-life health consequences.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This material is based upon work supported by the National Science Foundation Graduate Research Fellowship Program under Grant No. DGE-1845298. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation.

ORCID iD

Morgan E. Peele

References

Barker

D. J.

(1997). Maternal nutrition, fetal nutrition, and disease in later life. Nutrition, 13, 807-813.

Batty

G. D.

Lawlor

D. A.

Macintyre

Clark

Leon

D. A.

(2005). Accuracy of adults’ recall of childhood social class: Findings from the Aberdeen children of the 1950s study. Journal of Epidemiology & Community Health, 59, 898-903.

Ben-Shlomo

Kuh

(2002). A life course approach to chronic disease epidemiology: Conceptual models, empirical challenges and interdisciplinary perspectives. International Journal of Epidemiology, 31, 285-293.

Berndt

Fors

(2016). Childhood living conditions, education and health among the oldest old in Sweden. Ageing & Society, 36, 631-648.

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.

Bowen

M. E.

González

H. M.

(2010). Childhood socioeconomic position and disability in later life: Results of the health and retirement study. American Journal of Public Health, 100(Suppl. 1), S197-S203.

Chan

J. Y.

Stern

D. A.

Guerra

Wright

A. L.

Morgan

W. J.

Martinez

F. D.

(2015). Pneumonia in childhood and impaired lung function in adults: A longitudinal study. Pediatrics, 135, 607-616.

Chow

C. K.

Teo

K. K.

Rangarajan

Islam

Gupta

Avezum

. . . Kazmi

(2013). Prevalence, awareness, treatment, and control of hypertension in rural and urban communities in high-, middle-, and low-income countries. Journal of the American Medical Association, 310, 959-968.

Elo

I. T.

(2009). Social class differentials in health and mortality: Patterns and explanations in comparative perspective. Annual Review of Sociology, 35, 553-572.

10.

Elo

I. T.

Preston

S. H.

(1992). Effects of early-life conditions on adult mortality: A review. Population Index, 58, 186-212.

11.

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.

12.

Ferraro

K. F.

Shippee

T. P.

(2009). Aging and cumulative inequality: How does inequality get under the skin? The Gerontologist, 49, 333-343.

13.

Fink

Günther

Hill

(2014). Slum residence and child health in developing countries. Demography, 51, 1175-1197.

14.

Fossati

(2017). From periphery to centre: Local government and the emergence of universal healthcare in Indonesia. Contemporary Southeast Asia: A Journal of International and Strategic Affairs, 39, 178-203.

15.

Frankenberg

Suriastini

Thomas

(2005). Can expanding access to basic healthcare improve children’s health status? Lessons from Indonesia’s “midwife in the village” programme. Population Studies, 59, 5-19.

16.

Galobardes

Shaw

Lawlor

D. A.

Lynch

J. W.

Smith

G. D.

(2006). Indicators of socioeconomic position (part 1). Journal of Epidemiology & Community Health, 60(1), 7-12.

17.

Guilmoto

C. Z.

Jones

G. W.

(Eds.). (2016). Contemporary demographic transformations in China, India and Indonesia. Dordrecht, Netherlands: Springer.

18.

Haas

S. A.

(2007). The long-term effects of poor childhood health: An assessment and application of retrospective reports. Demography, 44, 113-135.

19.

Hamil-Luker

O’Rand

A. M.

(2007). Gender differences in the link between childhood socioeconomic conditions and heart attack risk in adulthood. Demography, 44, 137-158.

20.

Henretta

J. C.

McCrory

(2016). Childhood circumstances and mid-life functional mobility. Journal of Aging and Health, 28, 440-459.

21.

Huang

Soldo

B. J.

Elo

I. T.

(2011). Do early-life conditions predict functional health status in adulthood? The case of Mexico. Social Science & Medicine, 72, 100-107.

22.

Krieger

Okamoto

Selby

J. V.

(1998). Adult female twins’ recall of childhood social class and father’s education: A validation study for public health research. American Journal of Epidemiology, 147, 704-708.

23.

Langley-Evans

S. C.

(2015). Nutrition in early life and the programming of adult disease: A review. Journal of Human Nutrition and Dietetics, 28(Suppl. 1), 1-14.

24.

Lei

Smith

J. P.

Sun

Zhao

(2014). Gender differences in cognition in China and reasons for change over time: Evidence from CHARLS. The Journal of the Economics of Ageing, 4, 46-55.

25.

Margolis

(2010). Childhood morbidity and health in early adulthood: Life course linkages in a high morbidity context. Advances in Life Course Research, 15, 132-146.

26.

McEniry

McDermott

(2015). Early-life conditions, rapid demographic changes, and older adult health in the developing world. Biodemography and Social Biology, 61, 147-166.

27.

Monteverde

Noronha

Palloni

(2009). Effect of early conditions on disability among the elderly in Latin America and the Caribbean. Population Studies, 63(1), 21-35.

28.

Montez

J. K.

Hayward

M. D.

(2014). Cumulative childhood adversity, educational attainment, and active life expectancy among U.S. adults. Demography, 51, 413-435.

29.

Non

A. L.

Rewak

Kawachi

Gilman

S. E.

Loucks

E. B.

Appleton

A. A.

. . . Kubzansky

L. D.

(2014). Childhood social disadvantage, cardiometabolic risk, and chronic disease in adulthood. American Journal of Epidemiology, 180, 263-271.

30.

Shen

Zeng

(2014). Direct and indirect effects of childhood conditions on survival and health among male and female elderly in China. Social Science & Medicine, 119, 207-214.

31.

Smith

J. P.

(2009). Reconstructing childhood health histories. Demography, 46, 387-403.

32.

Sohn

(2017). The association between height and hypertension in Indonesia. Economics & Human Biology, 27, 74-83.

33.

Strauss

Witoelar

Sikoki

(2016). The fifth wave of the Indonesia family life survey: Overview and field report. Santa Monica, CA, USA: Rand.

34.

Sudharsanan

(2017). The association between socioeconomic status and adult mortality in a developing country: Evidence from a nationally representative longitudinal survey of Indonesian adults. The Journals of Gerontology, Series B: Psychological Sciences & Social Sciences.

35.

Umberson

Williams

Thomas

P. A.

Liu

Thomeer

M. B.

(2014). Race, gender, and chains of disadvantage: Childhood adversity, social relationships, and health. Journal of Health and Social Behavior, 55, 20-38.

36.

van Abeelen

A. F.

Elias

S. G.

De Jong

P. A.

Grobbee

D. E.

Bossuyt

P. M.

Van Der Schouw

Y. T.

. . . Uiterwaal

C. S

. (2013). Famine in the young and risk of later hospitalization for COPD and asthma. PLoS ONE, 8(12), e82636.

37.

Verbrugge

L. M.

Jette

A. M.

(1994). The disablement process. Social Science & Medicine, 38, 1-14.

38.

World Bank Group. (2017). Indonesia overview: Contexts. Washington, DC: World Bank Publications. Retrieved from https://www.worldbank.org/en/country/indonesia/overview

39.

World Health Organization. (2015). World report on ageing and health. Geneva, Switzerland: Author.