P298 Validation of the mobile Health Index (mHI) for remote monitoring of IBD disease activity
W.K. van Deen*1, A.E. van der Meulen- de Jong2, N.K. Parekh3, Y. Muyshondt3, E. Kane1, L. Eimers1, E.K. Inserra1, A. Zand1, C.A. DiNicola1, S. Bhatia3, J.M. Choi1, C.Y. Ha1, M.G. van Oijen1, E. Esrailian1, D.W. Hommes1
1UCLA Center for Inflammatory Bowel Diseases, Division of Digestive Diseases, University of California, Los Angeles, Los Angeles, United States, 2Leiden University Medical Center, Gastroenterology and Hepatology, Leiden, Netherlands, 3UC Irvine, Digestive Disease Center, Irvine, United States
Mobile health technologies are increasingly used to monitor patients remotely in chronic disease management such as diabetes mellitus and congestive heart failure. In IBD several monitoring systems have been studied but the use of mobile devices is still limited. We developed two 4-question indices to monitor disease activity using a smartphone in Crohn's disease (CD) and ulcerative colitis (UC) patients; the mobile Health Index (mHI) -CD and mHI-UC. The aim of this study was to validate the mHI in an independent cohort of IBD patients.
Patients with a diagnosis of CD or UC were included in 3 specialized IBD centers. During clinic visits, clinical disease activity indices (Harvey Bradshaw index for CD, partial Mayo score for UC) were completed by the physician, and patients filled out the mHI and the short-IBDQ for quality of life (QoL) assessment. During endoscopic visits the physician also completed an endoscopic score (Mayo for UC, SES-CD for CD). Spearman rho was calculated to assess correlation of the mHI with clinical and endoscopic disease activity indices. Patients were followed over time to assess responsiveness to change; a Spearman rho was calculated to estimate the correlation between changes in scores. A subset of patients filled out a second mHI within 24 hours after the clinic visit, and the intraclass correlation (ICC) was calculated to assess test-retest reliability.
In total 194 UC patients (19% active) and 217 CD patients (19% active) were included. The correlation of the mHI with clinic scores was 0.73 (p<0.001) for CD and 0.70 (p<.0001) for UC. Sensitivity and specificity to detect active disease were 93% and 66% for CD and 67% and 93% for UC, respectively. Both scores were responsive to change with a correlation of 0.37 (n=46) for CD and 0.65 (n=27) for UC between the change in scores. Test-retest reliability was good with an ICC of 0.91 (n=28) for the UC mHI and 0.95 (n=23) for the CD mHI. The CD mHI did correlate with endoscopic healing, though predictive values were poor (sensitivity 69%, specificity 51%, rho=0.31 (p=0.0062)). For UC the mHI correlated strongly with mucosal healing (rho=0.57 (p<0.0001), sensitivity 56%, specificity 88%). Furthermore, both scores have a strong inverse correlation with QoL (rhoi=-0.78 for CD and rho=-0.80 for UC, p<.0001)
The developed mHI scores for CD and UC have excellent test characteristics to monitor patients' symptoms remotely. Because the scores consist of 4 simple questions answered by patients, the score is ideal for implementation in a mobile smartphone app for home monitoring. As previously shown CD symptoms do correlate poorly with mucosal healing, while UC symptoms are strongly correlated with mucosal lesions.