Agreement between prospective diary data and retrospective questionnaire report of abdominal pain and stooling symptoms in children with irritable bowel syndrome

Clinical decision-making about diagnosis and treatment is routinely guided by patient report of symptoms, with symptom recall critically influencing medical management particularly for conditions lacking objective indicators such as functional gastrointestinal disorders (FGIDs). In research, patient-reported outcomes are increasingly conceptualized as primary endpoints1, 2, emphasizing patient report in evaluating potential treatments. Health-related information recalled via questionnaire has been questioned as unreliable or subject to bias.3 For example, peak and end effects (i.e., most intense and most recent pain) disproportionately influence pain recall and can undermine validity of retrospective self-reports.3–13 Recalled pain ratings are typically higher than momentary assessments, and lengthening recall interval magnifies bias.4,9, 11,13–19 Despite these limitations, retrospective questionnaires are used routinely and endorsed for assessing symptoms in FGID treatment trials1. Given this endorsement and the burden of diaries20, evaluating if questionnaires are a sufficient proxy is worthwhile. Most literature evaluating correspondence between recalled and recorded symptoms involves non-GI samples, but evidence suggests GI symptoms are subject to recall error, or that recalled and recorded bowel function diverge.21–27 This small literature is often limited by lack of correspondence between questionnaire and diary intervals (e.g., questionnaires preceding diary). This challenges interpretation, but adult IBS literature suggests that subtyping differs based on recalled versus recorded data26,27. Rome III Diagnostic Questionnaires may overestimate the frequency of abnormal stool form26, participants tend to recall more extreme stool forms as representative27,28, and those describing constipation underestimate stool frequency on questionnaire.23,24 Lackner and colleagues recently reported that, though as a group adults with IBS accurately recalled some IBS symptoms, individual correspondence varied with a subset of patients evidencing poor recall accuracy.28 Few studies examine correspondence between recalled and recorded pain in children, particularly in GI samples. Available research employs variable methods and yields variable interpretations concerning children’s recall accuracy. Accuracy usually increases with age, recalled pain ratings tend to be higher than momentary ratings (though pediatric findings are more mixed), and peak- and end-effects similarly bias ratings.14,18,29–32 Even less pediatric research concerns recalled versus recorded stool data, though evidence suggests recalled and recorded defecation frequency do not closely correspond and methodology affects diagnostic classification.33 One study by Chogle and colleagues34 examined correspondence between recalled and recorded pain in pediatric FGIDs, comparing a four-week pain diary to retrospective report of number of pain days. Results reflected a moderate positive correlation (Spearman correlation = 0.4), with 16% of children having perfect agreement. Interestingly, younger children had higher correspondence than adolescents. Group data reflected more pain days via diary than recall, but individual data indicated that 54% of children recalled fewer episodes than reported on diary, whereas 40% recalled more.34 To extend scant and often methodologically limited literature, we evaluated correspondence between retrospective questionnaire and prospective diary in pediatric IBS, evaluating pain and stool variables during the same reporting interval. Given the importance of parent perception in pediatrics and evidence that parent- and child-report differ35, correspondence between parent questionnaire and child diary was also examined. Relationship of child age to correspondence between child diary and questionnaire was also assessed.