Childhood misfortune and adult health: enduring and cascadic effects on somatic and psychological symptoms?

Abstract

Objective: This article traces the influence of early misfortune on somatic and psychological symptoms, examining whether the health disadvantages endure and/or worsen over a decade.

Method: The study uses two waves of data from the Midlife Development in the United States study. Structural equation models are used to assess change in somatic and psychological symptoms.

Results: Findings reveal that childhood abuse is associated with more initial somatic and psychological symptoms among American adults, but that family structure and financial strain had little association with baseline adult symptoms. The effects of abuse were not observed at Wave 2 (W2), net of Wave 1 (W1) symptoms.

Discussion: The results suggest an enduring-but not cascadic-effect of childhood abuse on adult health. There was little evidence that the consequences of early misfortune waned over the course of time or were less substantial for older adults; but neither did these early life experiences contribute to worsened symptoms over a decade of observation.

Figure 1.. Schematic figure illustrating enduring and cascadic effects

Note: Diamond denotes measurement at W1; Circle denotes measurement at W2. Each line segment illustrates a hypothetical pattern observable for a single study subject.

Figure 2.. Structural equation model assessed with MIDUS W1 and W2

Note: This model conceptually represents the analyses presented in this article, but numerous parameters are not pictured for the sake of clarity. Specifically, the error term of each latent variable from W1 is allowed to correlate with all other latent variables at W1, and latent variables from W2 are treated similarly. The indicator for each manifest variable at W1 has its error term correlated with its counterpart at W2. Finally, the exogenous variables (childhood misfortune plus covariates) are allowed to correlate.

Figure 3.. Measurement model results, W1

Note: Model fit statistics: χ² = 475.23, (df = 55); RMSEA = .05; CFI = .97; TLI = .95. ^aAll covariances are standardized and significant at p < .001.

Figure 4.. Measurement model results, W2

Note: Model fit statistics: χ² = 333.13, (df = 55); RMSEA = .05; CFI = .96; TLI = .94. ^aAll covariances are standardized and significant at p < .001.