Lest we forget: comparing retrospective and prospective assessments of adverse childhood experiences in the prediction of adult health

Abstract

Background: Adverse childhood experiences (ACEs; e.g. abuse, neglect, and parental loss) have been associated with increased risk for later-life disease and dysfunction using adults' retrospective self-reports of ACEs. Research should test whether associations between ACEs and health outcomes are the same for prospective and retrospective ACE measures.

Methods: We estimated agreement between ACEs prospectively recorded throughout childhood (by Study staff at Study member ages 3, 5, 7, 9, 11, 13, and 15) and retrospectively recalled in adulthood (by Study members when they reached age 38), in the population-representative Dunedin cohort (N = 1,037). We related both retrospective and prospective ACE measures to physical, mental, cognitive, and social health at midlife measured through both objective (e.g. biomarkers and neuropsychological tests) and subjective (e.g. self-reported) means.

Results: Dunedin and U.S. Centers for Disease Control ACE distributions were similar. Retrospective and prospective measures of adversity showed moderate agreement (r = .47, p < .001; weighted Kappa = .31, 95% CI: .27-.35). Both associated with all midlife outcomes. As compared to prospective ACEs, retrospective ACEs showed stronger associations with life outcomes that were subjectively assessed, and weaker associations with life outcomes that were objectively assessed. Recalled ACEs and poor subjective outcomes were correlated regardless of whether prospectively recorded ACEs were evident. Individuals who recalled more ACEs than had been prospectively recorded were more neurotic than average, and individuals who recalled fewer ACEs than recorded were more agreeable.

Conclusions: Prospective ACE records confirm associations between childhood adversity and negative life outcomes found previously using retrospective ACE reports. However, more agreeable and neurotic dispositions may, respectively, bias retrospective ACE measures toward underestimating the impact of adversity on objectively measured life outcomes and overestimating the impact of adversity on self-reported outcomes. Associations between personality factors and the propensity to recall adversity were extremely modest and warrant further investigation. Risk predictions based on retrospective ACE reports should utilize objective outcome measures. Where objective outcome measurements are difficult to obtain, correction factors may be warranted.

Keywords: Adverse childhood experiences; cognitive health; epidemiology; mental health; physical health.

Figure 1

Figure 1A. Distribution of ACEs in the Dunedin cohort, recorded prospectively and retrospectively, with comparison to ACE distributions reported in the CDC ACEs Study.¹ Notes. ¹Distribution of ACEs in the CDC ACEs Study from Table 1 of Felitti et al. (1998, p. 248). Figure 1B. Prevalence of individual ACEs in the Dunedin cohort, as recorded by prospective and retrospective measurement.

Figure 2

Personality characteristics of individuals who recall more or fewer ACEs than were recorded in their prospective records. Notes. “Retrospective > Prospective” represents individuals who recalled more ACEs than were prospectively recorded (N = 384) and “Retrospective < Prospective” represents individuals who recalled fewer ACEs than were prospectively recorded (N = 209). Analyses reported in the text are based on the complete distribution. Group means in the figure are adjusted for sex.