Background Clinical understanding of gastrointestinal symptoms is commonly based on patient reports of symptom experience. For diagnosis and treatment choices to be appropriate, symptom reports need to be accurate. We examined the agreement between questionnaire recall and prospective diary enumeration of symptoms relevant to the irritable bowel syndrome.

Patients and methods Data are reported from a randomly selected general population sample (n=238) and also a primary healthcare centre (PHC) sample (n=503, 10 PHCs). All the patients completed the questionnaires, which included Rome III-qualifying irritable bowel syndrome items and a stool and symptom diary over either 7 or 14 days. Agreement between retrospective questionnaire reports and prospective diaries was evaluated.

Results Concordance between questionnaires and diaries was highest for the simple construct of the occurrence of abdominal pain, although after adjusting for possible chance, agreement was only moderate in the general population sample. More complex constructs, such as pain relieved by defecation, yielded poorer concordance. In general, concordance was stronger among PHC respondents than in the general population sample.

Conclusion Concordance between questionnaires and diaries was generally poor and related to the complexity of the symptom construct and the type of respondent. The information used to classify individuals based on patient self-report may be unreliable, and therefore, more effort is needed to develop data collection instruments.