We appreciate the insightful comments by Koloski et al. on our paper.[1, 2] Experience of distress from physical symptoms or medical conditions is not equivalent to having a psychiatric diagnosis. It is therefore not likely that ‘effort after meaning’ may confound our assessment of the onset of these. These require the individual to meet specific diagnostic and impairment criteria, and would not be equivalent to the experience of distress from a stressful event.
Had time allowed, adding a rating of severity from time of onset may help clarify this issue. Recall bias may certainly limit the accuracy of onset data, which delimits causal inference. As such, our observation of coincident onset was one that invites further studies.
While recall bias is a known problem in cross-sectional studies, a vast literature on psychiatric epidemiology has been based on retrospective designs, where memory aids have been adopted to improve recall accuracy. For example, an interviewer may probe onset more accurately by asking about landmarks in the life course, such as ‘Was that when you were in high school? If so would it be in the last year?’
These probes, however, were barely feasible in time-limited telephone surveys as the additional time required for completing an interview is not only costly but would invite refusal and further reduce the participation rate. Face-to-face prospective cohorts are doubtless superior in this respect, but do have methodological and practical problems too, such as attrition and costs issues, making these studies on gastrointestinal disorders a rarity.
The recently published prospective population-based cohort on functional dyspepsia (FD) and psychiatric morbidities deserve special attention in supporting a bidirectional chronological association between anxiety symptoms and FD onset. As depression overlaps significantly with anxiety disorders on both diagnosis and symptom levels, the lack of association of FD with anxiety is intriguing and deserves further studies.