A cluster-randomised evaluation of a theory-based intervention to help people with Tb disease get the most from their treatment and care: the impact feasibility study

Introduction and Objectives Treatment for tuberculosis (TB) with multiple, potentially toxic, drugs lasts months but will cure >90% with drug sensitive TB, if taken as prescribed. Non-adherence inevitably occurs, and most methods to improve adherence use medication ‘reminders’ or supervised treatment that may not take into account the broader psychosocial and structural issues affecting adherence. We developed a theory-based manualised intervention including an enhanced structured needs assessment and risk management plan, educational videos, and an interactive patient booklet, to support adherence during anti-TB treatment. We report the intervention’s acceptability to patients and its ability to influence perceptions and behaviour when compared to standard care.

Methods Adult TB patients were recruited from four London TB centres within a pilot cluster-randomised controlled trial. Assessments included a process evaluation (to determine acceptability), the Beliefs about Medicines Questionnaire (BMQ) administered at 2 weeks, 3 and 6 months from baseline (measuring patient perception), and treatment monitoring boxes recording medication use (reflecting patient behaviour, measured as the proportion of doses taken at 168 days from treatment start).

Results 80 patients were recruited to the study (37 intervention, 43 control). Ethnicity and gender were evenly split between arms, age ranged from 18–81 years, with the largest group 28–38 years. The intervention had a positive effect on perceptions of treatment necessity compared to controls from 2 weeks to 3 months (p=.05) (medium to large effect size). Patients regarded all elements of the intervention as helpful, with mean helpfulness ratings (min0-max100) of 86 (patient booklet), 82 (needs assessment), 77 (medication diary/plan) and 72 (videos). The unadjusted mean proportion of doses taken over 168 days in the control arm was 0.88 compared to 0.81 in the intervention arm (estimated mean difference of -0.07, p= 0.23, CI -0.41–0.27). Adjusted estimates were similar.

Conclusions The IMPACT intervention is acceptable to patients and may change their perceptions of anti-TB treatment, though this was not reflected in increased total adherence. Measures to sustain the intervention’s effect throughout treatment should be investigated in future work.