Accuracy and bias in retrospective symptom reporting.

Abstract

Self-reported bodily symptoms are of primary importance in healthcare and in health-related research. Typically, they are assessed in clinical interviews or by means of traditional questionnaire formats that require the respondent to provide retrospective symptom estimates rated along intuitive frequency and/or intensity standards and aggregated across varying or unspecified time windows. Retrospective symptom assessments are often biased when compared to (averaged) momentary assessments of symptoms. A variety of factors and conditions have been identified to influence the amount of bias in symptom reporting. Recent research has focused on the underlying mechanisms for the discrepancy between memory and experience. It is suggested that different types of questions and formats assess different types of information, and each may be relevant for different purposes. Knowledge of these underlying mechanisms also provides a relevant framework to better understand individual differences in symptom reporting, including somatoform and somatic symptom disorder. Accuracy of self-reported bodily symptoms is important for the clinician and the researcher. Understanding the mechanisms underlying bias may provide an interesting window to understand how symptom episodes are processed, encoded, and consolidated in memory and may also provide clues to modify symptom experiences.