Search in the Abstract Database

Abstracts Search 2014

* = Presenting author

NO003. Development of an assessment method based on Health Belief Model to identify patients with ulcerative colitis at high risk of non-adherence to aminosalicylates

A. Kawakami1, M. Tanaka2, M. Nishigaki2, N. Yamamoto-Mitani2, N. Yoshimura3, R. Suzuki4, S. Maeda5, R. Kunisaki1, 1Yokohama City University Medical centre, IBD Centre, Yokohama, Japan, 2The Univercity of Tokyo, School of Health Sciences and Nursing, Tokyo, Japan, 3Social Insurance Central General Hospital, IBD Centre, Tokyo, Japan, 4Kannai Suzuki Clinic, Division of Gastroenterology, Yokohama, Japan, 5Yokohama City University Graduate School of Medicine, Department of Gastroenterology, Yokohama, Japan


Non-adherence to aminosalicylates is an important issue in medication for patients with inflammatory bowel disease. Non-adherence ratio in ulcerative colitis (UC) patients has been reported by 30–45%. Theoretical approach is essential for medical professionals to motivate patients to their adequate behaviour about medication. The Health Belief Model (HBM) is one of the most common theory used in understanding the motivation for taking health behavior, including medication. This study aimed to develop an assessment method based on Health Belief Model to identify patient with UC at high-risk of non-adherence to aminosalicylates.


This study was a multicenter, cross-sectional study. Outpatients with UC who had taken aminosalicylates were enrolled. Aminosalicylate adherence was assessed using both a self-administered questionnaire and medical records. We defined “Non-adherence” as: taking less than an 80% of the prescribed dose. Health beliefs for taking aminosalicylates were assessed using self-administered questionnaire, which consisted of 19 items reflecting the elements of HBM: perceived susceptibility, severity, benefits, barriers, cues to action. We applied a logistic regression model to identify related factors for non-adherence. Then we developed a numeric scoring system to identify non-adherent patients and assessed its discrimination efficiency.


Among the 429 eligible patients, non-adherence was found in 129 (29.6%). Multiple logistic regression model showed that lower perception of susceptibility, severity, benefit, fewer cues to action and higher perception of barrier to HBM, which means that they were problems as to perception of health beliefs were identified as related factors to non-adherence. Twenty-two items were included in the assessment model. When the cutoff point was set at 60, the model had a sensitivity of an 85.0%, and a specificity of 69.2%, with the area under the curve (AUC) of 0.84 (95% Confidence Interval = 0.79–0.91). The tool score stratified patients into four groups, demonstrating a graded relation to non-adherence rate (P < 0.01).


The model we developed in this study appeared to be reliable for identifying patients with a high-risk of non-adherence to aminosalicylates based on HBM. This scoring model could provide clinically relevant information to support medication taking behaviour to suppress UC relapse.

Acknowledgements: This study was supported by a Chiyoda-Kenko Foundation grant to Aki Kawakami. Authors have no other sources of funding to report.