DOT for patients with limited access to health care facilities in a hill district of Eastern Nepal

SETTING: The hill district in Nepal, where access to health care facilities is difficult. OBJECTIVE: To compare results before and after a decentralised directly obs erved treatment (DOT) intervention. DESIGN: Prospective study of patients registered in Dhankuta district, Nepa l, 1996-1999. Patients received their intensive phase treatment under healt h worker supervision via one of three DOT options: 1) ambulatory from the p eripheral government health facilities; 2) ambulatory from an international non-governmental organisation (INGO) TB clinic in district centre; or 3) r esident in INGO TB hostel in district centre. Historical data from 1995-199 6, with unsupervised short-course chemotherapy, were used for comparison. RESULTS: Of 307 new cases, respectively 126 (41%), 86 (28%) and 95 (31%) to ok their intensive phase treatment via options 1, 2 and 3. Smear conversion (at 2 months) and cure rates in new smear-positive pulmonary tuberculosis cases were respectively 81.6% (vs. 58.8% historical, P = 0.001) and 84.9% ( vs. 76.7% historical, P = 0.03). Overall costs to the INGO provider fell by 7%, mainly as a result of staffing reductions in the INGO services made po ssible by rationalisation with government services during the intervention. CONCLUSION: By offering varied DOT delivery routes, including an in-patient option, satisfactory results are possible with DOT even in areas where acc ess to health care facilities is difficult. Provision, of in-patient care v ia an INGO TB hostel allowed a significant proportion of new cases (31%) to receive their intensive phase treatment who otherwise may have had difficu lty accessing treatment, due either to the distance to the nearest health f acility or to disease severity. Substitution of government hospital beds or local hotel beds for the INGO hostel beds may allow the model to be reprod uced elsewhere in similar geographical conditions in Nepal, but further stu dies should be performed in a non-INGO supported district beforehand.