Longitudinal association of mental-health problems with subsequent diagnosis of persistent somatic symptom disorders: A large-scale registry-based observational study in primary care.

BACKGROUND

Somatic symptoms of common persistent somatic symptom (PSS) syndromes like irritable bowel syndrome (IBS), chronic fatigue syndrome (CFS), and fibromyalgia (FM) are not fully attributed to well-established biomedical pathological processes. These syndromes are often under-recognized and diagnosis is often delayed. This study assessed the extent to which mental health problems precede the onset of IBS, CFS, and FM diagnoses using large-scale registry-based data from primary care settings.

METHOD: Data from 11,409 patients were anonymously extracted from primary care data-bases in the Netherlands. Cases (IBS, CFS, or FM) and non-cases were matched for age and sex using a 1:2 ratio. Associations with preceding mental health were available mental health-related registrations in the dataset (i.e., mental health-related ICPC-codes, referrals, and psychopharmaceuticals) registered prior to diagnosis. For predictive modeling, logistic LASSO regressions were applied.

RESULTS

A total of 27 variables were longitudinally associated with IBS, CFS or FM (IBS k = 25, CFS k = 10, and FM k = 20). Five variables were longitudinally associated with all three syndromes (i.e., anxiety, psychosis, addiction behavior, and concentration disorders had positive predictive value and mental health-related referrals had negative predictive value). The overall classification performance of the models was fair (AUCIBS = 0.77) to good (AUCCFS = 0.82, AUCFM = 0.88).

CONCLUSIONS

Findings indicate that mental health-related registrations in primary care are associated with, vary between, and can accurately predict IBS, CFS, and/or FM. Prediction rules derived from mental health-related registrations might be able to support GPs in identifying patients with PSS. Future studies should investigate whether distinct decision rules are needed for the different syndromes.