Changing trends in coding for depression among the UK inflammatory bowel disease population

We read with interest the findings of Frolkis et al and welcome the contribution to this challenging field of research.1 An important consideration in the methodology used is the case definition of incident depression, defined as any patient with a code for a depressive disorder in The Health Improvement Network (THIN) database between 1986 and 2012. The authors acknowledge a limitation of their study may be that patients with depression could have been misclassified. Over the study period used, Rait et al have previously demonstrated a significant downward trend in the use of diagnostic codes for depressive disorders in the THIN database.2 These have been replaced with symptom codes such as 'low mood'. We have found similar trends in the inflammatory bowel disease population using the Clinical Practice Research Database, a data source that has significant overlap with THIN3 (figure 1). These temporal changes in...

http://bit.ly/2WZle3E