Objective: To develop an adherence rating score (ARS) system specific for tuberculosis (TB) patients.
Methods: A cross-sectional survey of 124 TB patients was conducted to figure out risk factors for adherence to treatment. The step-wise logistic regression models were used for selecting adherence-related variables. ARS was developed based on the weighting scores of the parameters of all the predicted variables in the logistic model. The reliability and responsibility of ARS was evaluated by using external data from an open label randomized controlled trial on 574 TB patients. The patients were grouped as adherence group (247 patients) and non-adherence group (327 patients) based on the predicted ARS. And the non-adherence group was randomized divided into a trail group (146 patients) and a control group (181 patients). The intervention for the trail group was custom health educational material aimed to reduce ARS, while the intervention for control groups was general TB education material, which was routinely used in the current local TB control settings. The cumulative non-adherence rates of the three groups were compared with each other after six-month follow-up period of treatment.
Results: The ARS system had 7 items which covered the following domains: disease status, psychology, patients' KAP (knowledge, attitude, and practice), regularly life-style and social supports. The score of ARS was 2.38+/-0.18 (mean+/-SD) for adherence patients, and 4.69 +/-0.20 (mean+/-SD) for non-adherence patients (t=8.52, P<0.01). In the randomized controlled trial, the six months cumulative non-adherence rates ware 24.7% for the trail group and it was 41.4% for the control group(P<0.01); while the six months cumulative non-adherence rates were not statistical significant difference between trail group and adherence group (P>0.05).
Conclusion: The ARS system was reliability and validity for evaluating the adherence of TB treatment in the stop TB settings in China.