Abstract
Background:
The long-term health effects of physical child abuse are well documented in self-report, retrospective studies. However, there have been few longitudinal, multimethod studies on physiological responses to stress and the onset of chronic disease, thereby slowing the advancement of prevention and intervention programs.
Objectives:
This study used survey data from an extended longitudinal study to examine prospective and retrospective associations between measures of physical child abuse and adult health in the 40s.
Participants and Setting:
Data are from an ongoing longitudinal study of the correlates and consequences of child maltreatment that began in the 1970s with a sample of 457 children.
Methods:
Bivariate correlations and multiple regression models with covariates were used to assess associations between measures of physical child abuse and outcomes of self-reported health in adulthood.
Results:
Physical child abuse, measured retrospectively, was moderately related to reports of overall health, as well as a number of adult health problems and conditions, such as back and chest pain, hypertension, and certain forms of cancer. Associations were also observed for lifetime alcohol problems and past-year doctor and emergency room visits. Fewer associations between prospective parent self-report measures of physical child abuse and adult health were identified, although child welfare (official record) reports performed similarly to retrospective measures.
Conclusions:
This study adds important information on the long-term health effects of child physical abuse, as well as measurement differences in the prediction of adult health outcomes. Conclusions drawn from prospective and retrospective studies of abuse are at best inconsistent, and possibly incompatible.
Keywords: child maltreatment, physical child abuse, adult health, prospective, retrospective, measurement
Introduction
Studies of physical health and disease typically focus on older adults. However, certain conditions and undiagnosed symptoms can occur earlier in the 40s, or even earlier (Shonkoff, Boyce, & McEwen, 2009). This is particularly true of vulnerable populations (Belsky et al., 2017; Belsky et al., 2015) including those with histories of child abuse (Johnson et al., 2017). Vulnerability to chronic health conditions among abused children is heightened by the effects of toxic stress and “weathering,” an acceleration of the normal aging process (Belsky et al., 2017; Belsky et al., 2015) and impairment of immune functioning associated with the onset of chronic disease (Middlebrooks & Audage, 2008; Shonkoff et al., 2009).
A well-known study of adult health outcomes and adverse childhood experiences (ACEs), including abuse, found that adults, ages 19–92 (average age 56 years), who experienced an increasing number of childhood adversities were at higher risk for illnesses and diseases, such as diabetes, stroke, and heart disease (Felitti et al., 1998). In another study by Merrick and colleagues {, 2019 #5} using data from the Behavior Risk Factor Surveillance System for 25 U.S. states, ACEs were related to chronic disease risk in a dose-response pattern; higher ACE scores predicted more disease risk.
Findings from other studies point to physical child abuse as a potent risk factor for poor health and disease (T. I. Herrenkohl, Hong, Klika, Herrenkohl, & Russo, 2013; Johnson et al., 2017; Lansford et al., 2021; Springer, Sheridan, Kuo, & Carnes, 2007; Widom, Horan, & Brzustowicz, 2015), with links to hypertension, diabetes (Thomas, Hypponen, & Power, 2008), obesity (Ouchi, Parker, Lugus, & Walsh, 2011; Pretty, O’Leary, Cairney, & Wade, 2013; Thomas et al., 2008), lung disease, arthritis, autoimmune disorders (Goodwin & Stein, 2004), and coronary heart disease (Wegman & Stetle, 2009). Notably, not all studies find that abuse is strongly predictive of adult health conditions. For example, a prospective study by Johnson and colleagues (2017) found no direct association between prospectively measured physical abuse and indicators of adult health and health risk behaviors among midlife adults, suggesting that associations based on retrospective and cross-sectional data may be inflated, or that adults who recall past abuse are at higher risk (Newbury et al., 2018).
Notably, most findings published in the literature on the long-term health effects of child abuse are primarily based on retrospective data, the limitations of which are well documented (Henry, Moffitt, Caspi, Langley, & Silva, 1994; Lansford et al., 2021; Newbury et al., 2018; Widom & Shepard, 1996). Of primary concern is the accuracy of retrospective reports of child abuse, which involve memories of distant childhood events that can be selectively recalled, forgotten, or shaped by later life experiences (Tajima, Herrenkohl, Huang, & Whitney, 2004; Widom & Shepard, 1996). Prospective measures of child abuse are thought to be less biased, but not all prospective measures are the same (McKinney, Harris, & Caetano, 2009; Tajima et al., 2004). For example, prospective measures based on official record data are thought to underestimate the incidence of abuse because relatively few cases are actually reported to child protection agencies (Bank & Burraston, 2001; Lansford et al., 2021; McGee, Wolfe, Yuen, Wilson, & Carnochan, 1995; Tajima et al., 2004). Collecting child abuse data directly from parents is a viable (Lansford et al., 2021) but complicated alternative because of reporting requirements and then sensitivity of the subject (Tajima et al., 2004).
In the current study, we examine how retrospective and prospective measures of physical child abuse relate to self-reported adult health at midlife. Data are from an extended longitudinal study focused on the correlates and long-term consequences of child abuse and neglect (T. I. Herrenkohl et al., 2013). We view this work as an important contribution to the literature because it furthers understanding of how prospective and retrospective measures derived from different data sources relate to long-term health outcomes and how substantive findings can vary as a function of which measures are used.
Methods
Sample
Data are from an ongoing, prospective investigation of the correlates and consequences of child maltreatment (R. C. Herrenkohl, Herrenkohl, Egolf, & Wu, 1991; T. I. Herrenkohl et al., 2013). The study began in the 1970s with a gender-balanced sample of 457 children, 18 months to 6 years of age. Participants were selected from child welfare caseloads and various group settings (e.g., Head Start centers, daycare centers) within a two-county area of eastern Pennsylvania in the United States. Subsequent assessments of the sample were completed when children of the study were in elementary school (1980–1982), adolescence (1990–1992), and adulthood (2008–2010 and 2019–2020). Extensive data on child abuse and neglect were obtained at all waves of the study, including from parents and child welfare agencies at the start of the study in the 1970s and 1980s. The original child sample was gender balanced and predominantly White (80.7% White, 11.2% more than one race, and 6.8% Black/African American, Native Hawaiian, or other Pacific Islander, or unknown) and socioeconomically diverse; more than 60% of families were living below the federal poverty level upon entering the study. Data collection and analysis procedures were approved by Institutional Review Boards at several institutions (Lehigh University, University of Washington, and University of Michigan).
The most recent data collection for the study was completed in 2020 when child participants had reached the age of 46 years on average. Of 303 participants retained in this most recent assessment, 52.5% were female, and 83.8% were white. Eighty percent of participants had graduated from high school or earned a GED (high school equivalency diploma); 44% had earned at least one higher education degree; and 76% were employed at least part-time. The median household income of the sample was between $60,000 and $70,000 and around a quarter of households (23%) were receiving food assistance (e.g., food stamps). Over a third of participants (36%) were receiving medical assistance because of low incomes.
Measures
Prospective Parent Reports of Physical Child Abuse
Prospective data on physical child abuse were collected from parents self-reports and child welfare (official) records. Official record reports of child maltreatment were obtained at the start of the study in the 1970s and were used in recruitment of families to the study. Just over half of all child participants (54%) were involved with child welfare agencies for abuse and neglect at the time of recruitment.
Additionally, in the preschool and schoolage assessments, all primary caregivers (mostly mothers) reported on the disciplining practices they and other caregivers used with their children. In these two (preschool and schoolage) assessments, disciplining questions were anchored in different time periods; in the preschool assessment, parents were asked about their discipline practices in (a) the past 3 months and (b) prior to the last 3 months (e.g., Did you use spanking prior to the last 3 months? Did your spouse (partner)? Did you slap your child’s hands or legs prior to the last 3 months? Did your spouse (partner)?). In the schoolage assessment, parents were asked about their discipline practices in the past year (e.g., Have you used spanking as a means of disciplining your child in the last 12 months? Has your spouse (partner) used this method?). Each time period is represented by its own variable in the analyses: (1) preschool abuse past 3 months; (2) preschool abuse prior to past 3 months; 3. schoolage abuse past year.
Variables for each time period were developed by summing the severity ratings of 12 discipline practices considered severely punishing and abusive (e.g., burning a child, biting a child, or hitting a child with a hard object so as to bruise). Each of these practices received a severity rating of 4 or 5 on a scale ranging from 1 (rewarding) to 5 (abusive). Summed scores provide an overall severity rating for each child participant. Scores in the 4- to 5-point range indicate a child was exposed to 1 abusive practice; scores in the 8- to 10-point range indicate a child was exposed to two abusive practices; and scores in the 12- to 15-point range indicate a child was exposed to three abusive practices. Scores above 16 indicate a child was exposed to four or more abusive practices at for that assessment. Scores of 0 indicate a child was not abused.
Retrospective Reports of Physical Child Abuse
Child participants, as adolescents and adults, retrospectively reported on their earlier experiences of physical abuse. In the adolescent assessment, child participants reported on the same 12 abusive physical disciplining practices reported earlier by their parents. Questions asked of adolescents referred to the discipline they received from birth to the age of 12 (e.g., Has your [mother], [father], or [anyone else] used the following methods to influence your behavior—spanking you; slapping your face; slapping your hands or legs?). As with the parent-self report measures, practices reported by adolescents were assigned severity ratings and these ratings were then summed to provide an overall severity score for each child participant. Finally, an adult retrospective measure of abuse was developed by scoring responses to a single question: “As a child, were you ever badly beaten up by your parents or the people who raised you (yes/no)?”
Self-Reported Adult Health
In the 2020 adult assessment, participants were asked about their overall physical health (e.g., “In general, would you say your health is excellent, very good, good, fair, poor?” “My health is excellent [definitely true, mostly true, mostly false, definitely false],” “Compared to one year ago, how would you rate your health in general now?”). Participants were also asked whether they were bothered during the past 4 weeks by a range of recent physical health problems, such as stomach pain, back pain, pain in your arms, legs, or joints, headaches, chest pain, dizziness, fainting spells, feeling your heart pound or race, shortness of breath, pain or problems during sexual intercourse, constipation, loose bowels or diarrhea, nausea, gas or indigestion, feeling tired or low energy, or trouble sleeping. These variables were scored 1, not at all; 2, a little; and 3, a lot. Additionally, participants were asked whether they ever had any lifetime conditions, such as problems with alcohol, arthritis, asthma, breast or colon cancer, chronic obstructive pulmonary disease, diabetes, heart disease, hepatitis or liver disease, high blood pressure, high cholesterol, and stroke or near stroke in their life. All lifetime variables were scored 0 (no) and 1 (yes). Physical health problems and lifetime illnesses were analyzed separately and then combined in overall measures (summative indexes) to capture all recent (past 4 weeks) and lifetime problems or conditions. Additionally, participants were asked whether, in the past year, they had been to a doctor or other medical professional (yes/no); had been to a hospital emergency room (yes/no); and had been hospitalized (yes/no).
Demographics
Covariates included age, race (white/non-white), and gender (female=0, male=1) to account for differences on these variables (Widom, Czaja, Bentley, & Johnson, 2012).
Analyses
Bivariate correlations and multiple regression models with covariates were used to assess associations between the various measures of child physical abuse relate to outcomes of self-reported health in adulthood. Correlation results appearing in the tables that follow for Pearson’s r (Pearson product-moment correlations). Nonparametric tests for categorical variables were also conducted and results were highly similar.
Results
Descriptive statistics for analysis variables are provided in Table 1. As shown in the table, recent (past 4 weeks) physical health problems were experienced, on average, “a little of the time” for those in the sample. Mean scores for back pain, pain in arms and legs, and feeling tired or having low energy indicate these problems were experienced more often than other problems. Twenty-five percent or more of the sample reported lifetime illnesses or conditions, including arthritis, high blood pressure, and high cholesterol, whereas relatively few participants reported conditions, such as breast cancer, colon cancer, glaucoma, and heart disease. Around 20% of the sample reported lifetime alcohol problems. Scores for overall lifetime illnesses or conditions ranged from 0–8, with an average of 1.72. Nearly three quarters of the sample (72%) had been to a doctor; 35% had been to a hospital emergency room; and 17% had been hospitalized in the past year, when participants were assessed in 2020.
Table 1.
Descriptive statistics for prospective and retrospective measures of physical child abuse and self-reported adult health
| N | Min | Max | M | S.D. | |
|---|---|---|---|---|---|
| Child welfare official record (prospective) | 303 | 0 | 1 | 0.49 | 0.5 |
| Preschool abuse (past 3 months) | 300 | 0 | 42.87 | 7.26 | 8.47 |
| Preschool abuse (prior to 3 months) | 302 | 0 | 80.44 | 18.22 | 17.25 |
| Schoolage abuse (past year) | 242 | 0 | 57.48 | 11.46 | 11.12 |
| Adolescent retrospective abuse (before age 12) | 283 | 0 | 114.3 | 32.66 | 26.01 |
| Adult retrospective abuse | 299 | 0 | 1 | 0.23 | 0.42 |
| Overall health | 302 | 1 | 4 | 2.79 | 0.91 |
| Health compared to one year ago | 302 | 1 | 5 | 3.29 | 0.93 |
| Physical health problems past 4 weeks (4w1-14) | 303 | 14 | 42 | 20.35 | 4.77 |
| Stomach pain (4w1) | 303 | 1 | 3 | 1.37 | 0.62 |
| Back pain (4w2) | 303 | 1 | 3 | 1.88 | 0.78 |
| Pain in arms, legs, or joints (4w3) | 303 | 1 | 3 | 1.86 | 0.74 |
| Headache (4w4) | 303 | 1 | 3 | 1.53 | 0.7 |
| Chest pain (4w5) | 303 | 1 | 3 | 1.17 | 0.42 |
| Dizziness (4w6) | 303 | 1 | 3 | 1.2 | 0.46 |
| Fainting spell (4w7) | 302 | 1 | 3 | 1.08 | 0.32 |
| Feeling your heart pound or race (4w8) | 303 | 1 | 3 | 1.28 | 0.51 |
| Shortness of breath (4w9) | 303 | 1 | 3 | 1.36 | 0.57 |
| Pain or problems during sexual intercourse (4w10) | 298 | 1 | 3 | 1.11 | 0.4 |
| Constipation, loose bowels, or diarrhea (4w11) | 303 | 1 | 3 | 1.42 | 0.63 |
| Nausea, gas, or indigestion (4w12) | 303 | 1 | 3 | 1.48 | 0.64 |
| Feeling tired or having low energy (4w13) | 303 | 1 | 3 | 1.86 | 0.71 |
| Trouble sleeping (4w14) | 303 | 1 | 3 | 1.78 | 0.77 |
| Illnesses or conditions—lifetime (L1–16) | 303 | 0 | 8 | 1.72 | 1.77 |
| Alcohol problems (L1) | 303 | 0 | 1 | 0.2 | 0.4 |
| Arthritis, excluding fibromyalgia (L2) | 303 | 0 | 1 | 0.28 | 0.45 |
| Asthma (L3) | 303 | 0 | 1 | 0.17 | 0.37 |
| Breast cancer (L4) | 303 | 0 | 1 | 0.01 | 0.11 |
| Colon cancer (L5) | 303 | 0 | 1 | 0 | 0.06 |
| COPD/emphysema (L6) | 303 | 0 | 1 | 0.06 | 0.23 |
| Diabetes (L7) | 303 | 0 | 1 | 0.1 | 0.3 |
| Glaucoma (L8) | 303 | 0 | 1 | 0.02 | 0.14 |
| Heart disease (L9) | 303 | 0 | 1 | 0.07 | 0.26 |
| Hepatitis or liver disease (L10) | 303 | 0 | 1 | 0.02 | 0.15 |
| High blood pressure (L11) | 303 | 0 | 1 | 0.3 | 0.46 |
| High cholesterol (L12) | 303 | 0 | 1 | 0.24 | 0.43 |
| Peptic ulcer (L13) | 303 | 0 | 1 | 0.04 | 0.19 |
| Seizures [epilepsy] (L14) | 303 | 0 | 1 | 0.06 | 0.23 |
| Stroke or near stroke (L15) | 303 | 0 | 1 | 0.05 | 0.22 |
| Thyroid condition (L16) | 303 | 0 | 1 | 0.1 | 0.3 |
| Been to a doctor (past year) | 302 | 0 | 1 | 0.72 | 0.45 |
| Been to an emergency room (past year) | 301 | 0 | 1 | 0.35 | 0.48 |
| Been hospitalized (past year) | 302 | 0 | 1 | 0.17 | 0.38 |
Correlations among the variables are provided in Tables 2–4. As shown in Table 2, retrospective measures of abuse from the most recent adult assessment and the earlier adolescent assessment were moderately correlated with overall health. Correlations of a similar magnitude were also observed for the official record measure of child maltreatment (r=−.18, p<.01 and r=−.21, p<.001). However, only one prospective parent self-report measure of physical abuse was significantly correlated with overall (excellent) health in these analyses.
Table 2.
Correlations of prospective and retrospective measures of physical child abuse and self-reported overall adult health
| Child Welfare | Preschool Abuse (past 3 months) | Preschool Abuse (prior to past 3 months) | Schoolage Abuse | Adolescent Retrospective | Adult Retrospective | |
|---|---|---|---|---|---|---|
| Overall health | −0.18** | −0.07 | −0.16** | −0.12^ | −0.19** | −0.21*** |
| Current vs Past year health | 0.04 | −0.09 | −0.08 | −0.01 | 0.01 | −0.09 |
Note: School age sample N is 242 due to higher attrition at that wave of the study; adolescent sample N is 283;
p < 0.001,
p < 0.01,
p < 0.05,
p < 0.1
Table 4.
Correlations of prospective and retrospective measures of physical child abuse and self-reported lifetime health conditions
| Child Welfare | Preschool Abuse (past 3 months) | Preschool Abuse (prior to past 3 months) | Schoolage Abuse | Adolescent Retrospective | Adult Retrospective | |
|---|---|---|---|---|---|---|
| Alcohol problems (1) | 0.09 | 0.05 | 0.08 | 0.07 | 0.25*** | 0.24*** |
| Arthritis (2) | 0.07 | 0.07 | 0.14* | 0.07 | 0.15* | 0.11^ |
| Asthma (3) | 0.08 | −0.02 | −0.01 | 0.06 | 0.06 | 0.02 |
| Breast cancer (4) | 0.002 | −0.04 | 0.01 | 0.03 | 0.06 | 0.21*** |
| Colon cancer (5) | 0.06 | 0.02 | 0.13* | 0.14* | 0.14* | 0.11^ |
| COPD (6) | 0.08 | −0.03 | 0.08 | −0.07 | 0.03 | 0.07 |
| Diabetes (7) | 0.16** | 0.06 | 0.04 | −0.06 | −0.005 | −0.05 |
| Glaucoma (8) | −0.05 | −0.03 | −0.05 | 0.05 | 0.01 | −0.02 |
| Heart disease (9) | 0.03 | 0.06 | 0.05 | 0.08 | 0.11^ | 0.03 |
| Hepatitis or liver disease (10) | 0.07 | 0.07 | 0.02 | −0.06 | 0.07 | 0.02 |
| High blood pressure (11) | 0.1^ | −0.02 | 0.01 | 0.09 | 0.13* | 0.04 |
| High cholesterol (12) | 0.09 | −0.05 | 0.05 | 0.07 | 0.06 | 0.1^ |
| Peptic ulcers (13) | 0.09 | 0.09 | −0.01 | 0.07 | 0.12* | 0.08 |
| Seizures (14) | −0.01 | −0.001 | 0.02 | −0.001 | 0.02 | 0.07 |
| Stroke (15) | 0.08 | −0.02 | 0.02 | 0.03 | 0.11^ | 0.1^ |
| Thyroid condition (16) | 0.15 | −0.07 | −0.003 | −0.03 | 0.14* | 0.1^ |
| Overall problems (17) | 0.19** | 0.02 | 0.09 | 0.09 | 0.24*** | 0.19** |
Note: School age sample N is 242 due to higher attrition at that wave of the study; adolescent sample N is 283;
p < 0.001,
p < 0.01,
p < 0.05,
p < 0.1
Table 3 includes correlations of the abuse measures with recent physical health problems (past 4 weeks). Here, the adult retrospective measure of abuse was significantly correlated with a number of recent problems (e.g., back pain, pain in arms and legs), as well as the additive count of overall problems (r=.26, p<.001). The adolescent retrospective report of abuse was correlated only with back pain (r=.14, p<.05) and the count of recent problems (r=.15, p<.05). The official record measure of child maltreatment was correlated with a number of health outcomes, including the overall count of physical health problems in the past 4 weeks (r=.20, p<.001). Here, too, the prospective measures of abuse based on parent self-reports were associated with only a few (e.g., arthritis, colon cancer) adult health measures.
Table 3.
Correlations of prospective and retrospective measures of physical child abuse and self-reported recent adult physical health problems
| Child Welfare | Preschool Abuse (past 3 months) | Preschool Abuse (prior to past 3 months) | Schoolage Abuse | Adolescent Retrospective | Adult Retrospective | |
|---|---|---|---|---|---|---|
| Stomach Pain (1) | 0.04 | −0.03 | −0.004 | −0.01 | 0.04 | 0.04 |
| Back pain (2) | 0.19** | 0.11^ | 0.1^ | 0.19** | 0.14* | 0.2** |
| Pain in arms and legs (3) | 0.12* | 0.12* | 0.12* | 0.09 | 0.17** | 0.21*** |
| Chest pain (4) | 0.09 | 0.07 | 0.08 | 0.001 | 0.13* | 0.19** |
| Pain during intercourse (5) | 0.06 | 0.03 | 0.08 | 0.06 | 0.02 | 0.07 |
| Headache (6) | 0.1^ | 0.002 | 0.02 | −0.01 | 0.02 | 0.11^ |
| Dizziness (7) | 0.1^ | −0.01 | 0.05 | −0.01 | 0.1^ | 0.05 |
| Fainting spells (8) | 0.12* | 0.07 | 0.06 | 0.07 | 0.06 | −0.11 |
| Heart pounding (9) | 0.18** | −0.03 | −0.01 | 0.16* | 0.15* | 0.21*** |
| Shortness of breath (10) | 0.06 | 0.04 | 0.02 | 0.09 | 0.05 | 0.13* |
| Constipation (11) | 0.11 | 0.04 | 0.03 | 0.04 | 0.11^ | 0.15** |
| Nausea (12) | 0.03 | −0.01 | 0.03 | 0.01 | 0.08 | 0.12* |
| Feeling tired (13) | 0.15* | −0.02 | 0.01 | −0.03 | 0.07 | 0.27*** |
| Trouble sleeping (14) | 0.24*** | 0.02 | 0.001 | 0.02 | 0.06 | 0.22*** |
| Overall problems 1–14 (15) | 0.2*** | 0.05 | 0.07 | 0.09 | 0.15* | 0.26*** |
Note: School age sample N is 242 due to higher attrition at that wave of the study; adolescent sample N is 283;
p < 0.001,
p < 0.01,
p < 0.05,
p < 0.1
Table 4 provides correlations of the child abuse variables and lifetime health conditions. In this case, comparatively more significant results were observed for the adolescent retrospective measure of abuse. For example, this measure of abuse was positively correlated with lifetime alcohol problems (r=.25, p<.001), arthritis (r=.15, p<.05), colon cancer (r=.14, p<.05), high blood pressure (r=.13, p<.05), peptic ulcers (r=.12, p<.05), thyroid condition (r=.14, p<.05), and the additive count of problems (r=.24, p<.001). The adult retrospective measure was correlated with alcohol problems (r=.24, p<.001), breast cancer (r=.21, p<.001), and the overall number (count) of lifetime problems (r=.19, p<.01). The official record measure of child maltreatment was correlated with diabetes (r=.16, p<.01) and the overall count of problems (r=.19, p<.01). Notably, significant correlations were also observed for the preschool measure of abuse referring to abuse prior to the past 3 months and arthritis (r=.14, p<.05) and colon cancer (r=.13, p<.05). Past-year abuse measured in the school age wave of the study was also correlated with colon cancer (r=.14, p<.05).
In the last set of correlations, we examined associations between the child abuse variables and past year doctor appointments, emergency room visits, and hospitalizations. Here, the adult and adolescent retrospective measures of abuse were significantly correlated with visits to the emergency room (r=.26, p<.001 and r=.16, p<.01, respectively). The official record measure of child maltreatment was similarly correlated with emergency room visits (r=.19, p<.01). Additionally, the adult retrospective measures of abuse was also significantly correlated with doctor visits in the past year (r=.13, p<.05). None of the prospective measures of child abuse was significantly correlated with past year doctor visits or hospitalizations in adulthood.
Regression results are shown in Table 6. Separate regression models were conducted for each of the child abuse measures and each outcome, controlling for age, race, and gender. Analyses here were limited to outcomes of overall health (overall health; health compared to 1 year ago), the two count variables for recent (past 4 weeks) and lifetime health problems and conditions, and the 3 past year measures for doctor visits, emergency room visits, and hospitalizations. Similar to the correlation results, the adult and adolescent retrospective measures of abuse and the official record measure of child maltreatment were significant predictors. With one exception (preschool reports for “prior to the past 3 months”), none of the prospective measures of abuse significantly predicted adult health.
Table 6.
Regression models for prospective and retrospective measures of physical child abuse and self-reported adult health
| Child Welfare | Preschool Abuse (past 3 months) | Preschool Abuse (prior to past 3 months) | Schoolage Abuse | Adolescent Retrospective | Adult Retrospective | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| b | se | β | b | se | β | b | se | β | b | se | β | b | se | β | b | se | β | |
| −0.34** | 0.12 | −0.2 | −.01 | 0.01 | −0.1 | −.01* | 0 | −0.1 | −.01^ | 0.01 | −0.1 | −.01* | 0 | −0.2 | −0.45** | 0.18 | −0.2 | |
| Current vs Past year health | 0.07 | 0.11 | 0.04 | −.01^ | 0.01 | −0.1 | −0.01 | 0 | −0.1 | −.001 | 0.01 | −0 | −0.001 | 0 | −0 | −0.17 | 0.13 | −0.1 |
| Recent physical health | 1.99*** | 0.53 | 0.21 | 0.04 | 0.03 | 0.07 | .02 | 0.02 | 0.07 | 0.04 | 0.03 | 0.1 | 0.02* | 0.01 | 0.05 | 2.71*** | 0.63 | 0.24 |
| Lifetime health conditions | 0.58** | 0.2 | 0.17 | −.01 | 0.01 | −0 | .01 | 0.01 | 0.06 | 0.02 | 0.01 | 0.1 | 0.01** | 0 | 0.19 | 0.69** | 0.24 | 0.17 |
| Past year been to a doctor | .37 | 0.27 | 1.45 | −.02 | 0.02 | 0.98 | −.001 | 0.01 | 1 | 0.01 | 0.01 | 1.01 | 0.01* | 0.01 | 1.01 | 0.68^ | 0.36 | 1.98 |
| Past year been to ER | .81** | 0.26 | 2.26 | .01 | 0.02 | 1.01 | .01 | 0.01 | 1.01 | 0.003 | 0.01 | 1 | 0.01* | 0.01 | 1.01 | 1.14*** | 0.29 | 3.11 |
| Past year hospitalized | −.04 | 0.31 | 0.97 | −.02 | 0.02 | 0.98 | −.002 | 0.01 | 1 | 0 | 0.02 | 1 | 0.004 | 0.01 | 1 | 0.41 | 0.35 | 1.51 |
Note: School age sample N is 242 due to higher attrition at that wave of the study; adolescent sample N is 283;
p < 0.001,
p < 0.01,
p < 0.05,
p < 0.1
Discussion
Results provide additional, but qualified, support for the connection between child abuse and adult health, particularly if basing conclusions on abuse assessed retrospectively. At the same time, not all measures of abuse included in analyses were significantly related to health outcomes and effect sizes were somewhat smaller than have been reported in other studies and reviews of the literature (Wegman & Stetle, 2009).
In this study, we compared several prospective and retrospective measures of abuse from the childhood, adolescent, and adult waves of a longitudinal study spanning several decades. Not only did self-report measures from the adult and adolescent assessments of the study correlate more strongly than the childhood measures with recent and lifetime adult health problems and conditions in the 40s, they also predicted more problems, according to two summative indexes. Interestingly, an official record measure of child maltreatment related more similarly to the two retrospective measures in bivariate and multivariate models. Whether these discrepancies capture substantive differences in the prediction of adult health outcomes requires further investigation.
In an earlier analysis of the data used in the current study, Tajima and colleagues (2004) found moderate correspondence between prospective (parent reports) and retrospective (youth reports) of physical child abuse and moderate consistency in their prediction of adolescent health risk behaviors (lifetime and past year adolescent violence perpetration, alcohol abuse, marijuana abuse, dropping of school, and teenage pregnancy). Retrospective reports of abuse from adolescents predicted all but depression, and prospective reports predicted fewer outcomes overall. The researchers hypothesized that underreporting by respondents may help to explain the discordance in their measures and predictive models.
Research using other samples and methods is also instructive. For example, Johnson et al. (2017) used prospective data from the Minnesota Longitudinal Study of Risk and Adaptation to examine child abuse predictors of adult health in midlife adults, ages 37–39 years. Their composite measure of physical abuse was derived from multiple sources, including parent-child observations, caregiver interviews, medical and child welfare records, and adolescent reports. Adult health measures included self-reports of health behaviors (eating, exercise, sleep), overall health, and a summed scale of health problems much like the one used in our study. They also assessed biomarkers for cardiometabolic risk (e.g., blood pressure, BMI). Notably, their analyses failed to show any direct correlation between child physical abuse and adult health or health risk behaviors.
In another prospective study, Widom and colleagues (2012) investigated the association between official record measures of child abuse and indicators of metabolic risk in an adult sample. They found evidence of the long-term adverse health effects of child abuse, including risk markers for kidney and liver disease and diabetes. In another publication, the same research team reported that official records of child maltreatment were positively associated with allostatic load, a measure of poor health in adulthood. Analyses in that study accounted for gender, race, and age, as well as some hypothesized mediators (e.g., internalizing/externalizing behaviors, social support) (Widom et al., (2015).
Further, Newbury and colleagues (2018) compared prospective informant-reports and retrospective self-reports of childhood maltreatment in the prediction of psychiatric problems in young adults. They found that both measures had predictive utility but that retrospective reports were more strongly related to the outcomes in question. They concluded that their measures capture non-overlapping groups and that young adults who recall having been abused are at highest risk.
Why findings from prospective and retrospective studies appear inconsistent remains unclear, but there are several possible explanations. According to Wegman and Stetle (2009), findings can differ not only according to the data sources that are used to derive measures of abuse and health, but also based on the composition of samples and methods of analysis that are applied in any given study. Findings can also differ according to how abuse and health outcomes are defined and operationalized. Given the variation in methodology across studies addressing these topics, direct comparison of results can be difficult and lead to erroneous conclusions. Accordingly, we recommend efforts to bring more transparency to the range of methods underlying research on child abuse and adult health, and for researchers to be more intentional about comparing their measures and methods to those used by other researchers, noting and explaining inconsistencies where and when they exist. While meta-analytic findings support a connection between child abuse and adult health (Wegman & Stetle, 2009), inconsistencies abound and require further exploration.
Like other studies, ours has strengths and weaknesses. The prospective design and use of official record and self-report data on child abuse are clear strengths. However, self-report measures used in this study are subject to recall bias and are thought to be less reliable than objective measures (Wegman & Stetle, 2009). In addition, the sample is relatively homogeneous and does not reflect the racial and ethnic diversity of the general population in the region of the United States from which the sample was drawn. The study is also limited by some attrition over several decades since its inception.
However, results help to advance knowledge of the adverse health consequences of child abuse and draw attention to important measurement considerations that are understudied and often overlooked in research on health and adverse childhood experiences. Results are particularly interesting in light of the mixed findings already documented in the child abuse literature when prospective and retrospective measures are compared. While both measures of child abuse are useful in understanding stress-related risks to adult health, results of this and other studies like those of Johnson and colleagues (2017) and Widom and colleagues (2012; 2015) underscore the need to proceed cautiously when drawing substantive conclusions. Whereas retrospective, youth and adult report measures of abuse and childhood adversity are shown to be robust predictors of adult health and chronic disease, associations based on prospective reports of abuse are less consistent.
Table 5.
Correlations of prospective and retrospective measures of physical child abuse and self-reported past year adult healthcare use
| Child Welfare | Preschool Abuse (past 3 months) | Preschool Abuse (prior to past 3 months) | Schoolage Abuse | Adolescent Retrospective | Adult Retrospective | |
|---|---|---|---|---|---|---|
| Been to a doctor | 0.07 | −0.11^ | −0.02 | 0.02 | 0.11^ | 0.13* |
| Been to the ER | 0.19** | 0.04 | 0.07 | 0.03 | 0.16** | 0.26*** |
| Been hospitalized | 0.01 | −0.06 | 0.01 | −0.004 | 0.07 | 0.09 |
Note: School age sample N is 242 due to higher attrition at that wave of the study; adolescent sample N is 283;
p < 0.001,
p < 0.01,
p < 0.05,
p < 0.1
Acknowledgements:
This research was supported by the grants from the National Institute on Child Health and Development (R21HD094961) and National Institute on Aging (R01AG059823). The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the National Institutes of Health.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Declarations of interest: None.
References
- Bank L, & Burraston B (2001). Abusive home environments as predictors of poor adjustment during adolescence and early adulthood. Journal of Community Psychology, 29(3), 195–217. doi:DOI 10.1002/jcop.1014.abs [DOI] [Google Scholar]
- Belsky DJ, Caspi A, Cohen HJ, Kraus WE, Ramrakha S, Poulton R, & Moffitt TE (2017). Impact of early personal-history characteristics on the Pace of Aging: implications for clinical trials of therapies to slow aging and extend healthspan. Aging Cell, 1–8. doi: 10.1111/acel.12591 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Belsky DJ, Caspi A, Houts R, Cohena HJ, Corcorane DL, Danesef A, … Moffitt TE (2015). Quantification of biological aging in adults. Proceedings of the National Academy of Sciences, e4014–e4410. doi: 10.1073/pnas.1506264112/-/DCSupplemental [DOI] [PMC free article] [PubMed] [Google Scholar]
- Felitti VJ, Anda RF, Nordenberg D, Williamson DF, Spitz AM, Edwards V, … Marks JS (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(4), 245–258. doi: 10.1016/s0749-3797(98)00017-8 [DOI] [PubMed] [Google Scholar]
- Goodwin RD, & Stein MB (2004). Association between childhood trauma and physical disorders among adults in the United States. Psychological Medicine, 34(3), 509–520. doi: 10.1017/S003329170300134x [DOI] [PubMed] [Google Scholar]
- Henry B, Moffitt TE, Caspi A, Langley J, & Silva PA (1994). On the “remembrance of things past”: A longitudinal evaluation of the retrospective method. Psychological Assessment, 6(2), 92–101. doi: 10.1037/1040-3590.6.2.92 [DOI] [Google Scholar]
- Herrenkohl RC, Herrenkohl EC, Egolf BP, & Wu P (1991). The developmental consequences of child abuse: The Lehigh Longitudinal Study. In Starr RHJ & Wolfe DA (Eds.), The effects of child abuse and neglect: Issues and research (pp. 57–81). New York: The Guilford Press. [Google Scholar]
- Herrenkohl TI, Hong S, Klika JB, Herrenkohl RC, & Russo MJ (2013). Developmental impacts of child abuse and neglect related to adult mental health, substance use, and physical health. Journal of Family Violence, 28(2), 191–199. doi: 10.1007/s10896-012-9474-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Johnson WF, Huelsnitz CO, Carlson EA, Roisman GI, Englund MM, Miller GE, & Simpson JA (2017). Childhood abuse and neglect and physical health at midlife: Prospective, longitudinal evidence. Development and Psychopathology, 29(5), 1935–1946. doi: 10.1017/S095457941700150X [DOI] [PubMed] [Google Scholar]
- Lansford JE, Godwin J, McMahon RJ, Crowley M, Pettit GS, Bates JE, … Dodge KA (2021). Early physical abuse and adult outcomes. Pediatrics, 147(1), e20200873. doi: 10.1542/peds.2020-0873 [DOI] [PMC free article] [PubMed] [Google Scholar]
- McGee RA, Wolfe DA, Yuen SA, Wilson SK, & Carnochan J (1995). The measurement of maltreatment: A comparison of approaches. Child Abuse and Neglect, 19(2), 233–249. [DOI] [PubMed] [Google Scholar]
- McKinney CM, Harris TR, & Caetano R (2009). Reliability of self-reported childhood physical abuse by adults and factors predictive of inconsistent reporting. Violence and Victims, 24(5), 653–668. doi: 10.1891/0886-6708.24.5.65 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Middlebrooks JS, & Audage NC (2008). The effects of childhood stress on health across the lifespan. Centers for Disease Control and Prevention, National Center for Injury Prevention and Control, Atlanta, GA. [Google Scholar]
- Newbury JB, Arseneault L, Moffitt TE, Caspi A, Danese A, Baldwin JR, & Fisher HL (2018). Measuring childhood maltreatment to predict early-adult psychopathology: Comparison of prospective informant-reports and retrospective self-reports. Journal of Psychiatric Research, 96(57–64). doi: 10.1016/j.jpsychires.2017.09.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ouchi N, Parker JL, Lugus JJ, & Walsh K (2011). Adipokines in inflammation and metabolic disease. Nature Reviews Immunology, 11(2), 85–97. doi: 10.1038/nri2921 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pretty C, O’Leary DD, Cairney J, & Wade TJ (2013). Adverse childhood experiences and the cardiovascular health of children: A cross-sectional study. BMC Pediatrics, 13:208, 208. doi: 10.1186/1471-2431-13-208 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Shonkoff JP, Boyce WT, & McEwen BS (2009). Neuroscience, molecular biology, and the childhood roots of health disparities: Building a new framework for health promotion and disease prevention. JAMA, 301(21), 2252–2260. doi: 10.1001/jama.2009.754 [DOI] [PubMed] [Google Scholar]
- Springer KW, Sheridan J, Kuo D, & Carnes M (2007). Long-term physical and mental health consequences of childhood physical abuse: Results from a large population-based sample of men and women. Child Abuse & Neglect, 31(5), 517–530. doi: 10.1016/j.chiabu.2007.01.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tajima EA, Herrenkohl TI, Huang B, & Whitney SD (2004). Measuring child maltreatment: A comparison of prospective parent reports and retrospective adolescent reports. American Journal of Orthopsychiatry, 74(4), 424–435. doi: 10.1037/0002-9432.74.4.424 [DOI] [PubMed] [Google Scholar]
- Thomas C, Hypponen E, & Power C (2008). Obesity and type 2 diabetes risk in midadult life: The role of childhood adversity. Pediatrics, 121(5), e1240–1249. doi: 10.1542/peds.2007-2403 [DOI] [PubMed] [Google Scholar]
- Wegman HL, & Stetle C (2009). A meta-analytic review of the effects of childhood abuse on medical outcomes in adulthood. Psychosomatic Medicine, 71(8), 805–812. doi: 10.1097/PSY.0b013e3181bb2b46 [DOI] [PubMed] [Google Scholar]
- Widom CS, Czaja SJ, Bentley T, & Johnson MS (2012). A prospective investigation of physical health outcomes in abused and neglected children: New findings from a 30-year follow-up. American Journal of Public Health, 102, 1135–1144. doi:10.2105%2FAJPH.2011.300636 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Widom CS, Horan J, & Brzustowicz L (2015). Childhood maltreatment predicts allostatic load in adulthood. Child Abuse & Neglect, 47, 59–69. doi: 10.1016/j.chiabu.2015.01.016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Widom CS, & Shepard RL (1996). Accuracy of adult recollections of childhood victimization: Part 1. Childhood physical abuse. Psychological Assessment, 8(4), 412–421. doi: 10.1037/1040-3590.8.4.412 [DOI] [Google Scholar]