TOPLINE:

Predictive models with patient-reported psychosocial measures added to clinical data achieved better predictive performance for flares in patients with inflammatory bowel disease (IBD) than models that used baseline clinical data alone.

METHODOLOGY:

• Traditional prediction models for IBD flares mostly rely on clinical and biomarker data and have limited accuracy. Emerging evidence indicates that psychological and lifestyle factors may influence disease activity and the risk for flares.
• Researchers conducted an observational cohort study using data from the myIBDcoach remote monitoring app from November 2022 to June 2024 to develop and compare models that integrated clinical data with patient-reported outcomes.
• They enrolled adults with IBD who used the app and collected comprehensive data including baseline clinical and demographic information; disease type; medications; history of surgery; patient-reported levels of stress, fatigue, anxiety, and depression; life events; smoking status; and physical activity.
• Five predictive models were developed using baseline data only, baseline data plus psychosocial factors, baseline data plus psychosocial and lifestyle factors, psychosocial factors only, and lifestyle factors only. The models were trained on 70% of the data and tested on the remaining 30%, and their performance was compared.
• The primary outcome was the occurrence of a flare, defined as having clinical symptoms with a Monitor IBD at Home score ≥ 6 on a 0-10 scale, combined with fecal calprotectin levels > 250 μg/g or endoscopic disease activity.

TAKEAWAY:

• The study included 429 patients (58% male) with a mean age of about 26 years at IBD diagnosis. Of them, 198 (46%) experienced a flare during follow-up and reported higher scores for stress, depression, fatigue, and pain than those in remission (P < .05 for all).
• The model that combined baseline data with psychosocial factors achieved the highest performance, with an accuracy of 73.1% and an area under the receiver operating characteristic curve (AUC) of 0.769. The baseline-only model achieved an accuracy of 63.0% and an AUC of 0.652.
• The psychosocial-only model performed well, with an AUC of 0.739, whereas the lifestyle-only model performed poorly, with an AUC of 0.559. Adding lifestyle measures to the baseline plus psychosocial model did not meaningfully improve discrimination.
• Male sex, the use of biologic therapy at inclusion, and having pancolitis or extensive illness were strong predictors of IBD flares in the baseline model, and fatigue emerged as a strong psychosocial predictor.

IN PRACTICE:

“Healthcare professionals can create more precise predictive algorithms to identify patients at high risk of negative outcomes and enable early interventions by combining psychosocial and lifestyle aspects. For example, identifying those who have higher levels of psychosocial stressors may lead to targeted psychological support, and lifestyle changes may be customized to meet the needs of everyone,” the authors of the study wrote.

SOURCE:

The study was led by Y. Okegunna, MD, Maastricht University Medical Centre, Maastricht, Netherlands. It was published online in Digestive Diseases and Sciences.

LIMITATIONS:

The study used patients from the same population, limiting external validity. The outcome was simplified to a yes or no flare, which may not have captured disease complexity. The CIs were wide and restricted the certainty in the findings.

DISCLOSURES:

The study received support from Takeda Netherlands. The authors declared having no competing interests.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.