A randomized, controlled trial of interventions to improve adherence to isoniazid therapy to prevent tuberculosis in injection drug users

PURPOSE: To determine the effect of several interventions on adherence to tuberculosis preventive therapy. METHODS: We conducted a randomized trial with a factorial design comparing strategies for improving adherence to isoniazid preventive therapy in 300 injection drug users with reactive tuberculin tests and no evidence of active tuberculosis. Patients were assigned to receive directly observed isoniazid preventive therapy twice weekly (Supervised group, n = 99), daily self-administered isoniazid with peer counseling and education (Peer group, n = 101), or routine care (Routine group, n = 100). Patients within each arm were also randomly assigned to receive an immediate or deferred monthly $10 stipend for maintaining adherence. The endpoints of the trial were completing 6 months of treatment, pill-taking as measured by self-report or observation, isoniazid metabolites present in urine, and bottle opening as determined by electronic monitors in a subset of patients. Results: Completion of therapy was 80% for patients in the Supervised group, 78% in the Peer group, and 79% in the Routine group (P = 0.70). Completion was 83% (125 of 150) among patients receiving immediate incentives versus 75% (112 of 150) among patients with deferred incentives (P = 0.09). The proportion of patients who were observed or reported taking at least 80% of their doses was 82% for the Supervised arm of the study, compared with 71% for the Peer arm and 90% for the Routine arm. The proportion of patients who took 100% of doses was 77% for the Supervised arm (by observation), 6% for the Peer arm (by report), and 10% for the Routine arm (by report; P <0.001). Direct observation showed the median proportion of doses taken by the Supervised group was 100%, while electronic monitoring in a subset of patients showed the Peer group (n = 27) took 57% of prescribed doses and the Routine group (n = 32) took 49% (P <0.001). Patients in the Routine arm overreported adherence by twofold when data from electronic monitoring were used as a gold standard. There were no significant differences in electronically monitored adherence by type of incentive. CONCLUSION: Adherence to isoniazid preventive therapy by injection drug users is best with supervised care. Peer counseling improves adherence over routine care, as measured by electronic monitoring of pill caps, and patients receiving peer counseling more accurately reported their adherence. More widespread use of supervised care could contribute to reductions in tuberculosis rates among drug users and possibly other high-risk groups.