This article is more than 5 years old. Information may no longer be current.

Prospective, retrospective assessments yield different mechanisms underlying childhood maltreatment

A large systematic review and meta-analysis showed poor agreement between prospective and retrospective measures of childhood maltreatment, indicating that these two measures should not be used interchangeably to study risk factors associated with childhood maltreatment.

Researchers, clinicians and public health professionals use retrospective reports to quickly and inexpensively determine the consequences of childhood maltreatment, according to Jessie R. Baldwin, PhD, from the Social, Genetic and Developmental Psychiatry Centre, King’s College London, and colleagues.

“All these practices rely on the assumption that retrospective reports and prospective measures identify the same, or at least similar, groups of individuals. However, qualitative reviews have raised concerns about the validity of this assumption,” they wrote.

To examine the agreement between prospective and retrospective measures of childhood maltreatment, researchers searched online clinical databases for peer-reviewed articles. They paired studies with prospective measures of childhood maltreatment with corresponding retrospective data to compute the Cohen kappa coefficient. Moderators of agreement were the measure used for prospective or retrospective assessment of maltreatment, age at retrospective report, sample size, sex distribution and study quality.

The analysis included 16 studies with data on 25,471 participants. Analysis indicated poor agreement between prospective and retrospective measures of childhood maltreatment (kappa = 0.19; 95% CI, 0.14-0.24). Baldwin and colleagues found that more than 50% of individuals with prospective observations of childhood maltreatment did not report it retrospectively and similarly, more than 50% of those retrospectively reporting childhood maltreatment did not have concordant prospective observations.

Analysis showed that agreement was higher when retrospective measures of maltreatment were based on interviews instead of questionnaires (Q = 4.1521; P = .04) and in studies with smaller samples (Q = 4.2251; P = .04). Also, agreement for any of the childhood maltreatment measures was lower than agreement for more obvious forms of adversity, like parental loss.

However, type of prospective measure used, participant age at retrospective report, sex distribution of the study sample and study quality did not affect agreement, according to Baldwin and colleagues.

“The mechanisms underlying disease risk in children identified as being maltreated through prospective assessments may be different from the mechanisms underlying disease risk in adults retrospectively reporting childhood maltreatment. If risk mechanisms are different, then the two groups will need different interventions to effectively prevent and treat disease,” they wrote. “As such, our findings provide a new framework for etiologic research on childhood maltreatment and intervention development.”

These findings offer health care professionals “a wake-up call” to help them identify the important differences between prospective and retrospective reports, Cathy Spatz Widom, PhD, from the psychology department, John Jay College, City University of New York, wrote in a related comment.

“From a clinical perspective, these new findings do not negate the importance of listening to what a patient says, but they suggest that caution should be used in assuming that these retrospective reports accurately represent experiences, rather than perceptions, interpretations or existential recollections,” she wrote. – by Savannah Demko

Disclosures: The authors and comment author report no relevant financial disclosures.