Giving newly diagnosed tuberculosis (TB) patients more counselling, treatment closer to home and a treatment supporter of their own choice to supervise pill taking more than halved the risk of defaulting on a course of TB treatment when compared with the standard directly observed therapy approach, according to findings from a randomised study carried out in Senegal and published today in the Journal of the American Medical Association.
Cure rates among TB patients are low in many countries, despite attempts to implement directly observed therapy (DOTS) as the standard of care. Incomplete TB treatment may lead to the emergence of multidrug-resistant TB or further TB transmission.
Research in Senegal highlighted a number of reasons for treatment default. Poor communication between healthcare workers and patients was an important reason given in qualitative research carried out by Senegal’s national TB programme. Poorly applied directly observed therapy, inadequate follow-up of defaulters and poor supervision of local treatment centres also contributed, the researchers found.
Dr Moustapha Ndir and colleagues designed a package of interventions to address these problems and tested them in a study which randomised 24 district health centres to adopt either the intervention package or a standard DOTS approach for TB treatment.
In the intervention group, all patients received initial counselling from the TB nurse and then commenced treatment. Participants were then referred to a health post near to home, where the local nurse gave more information about TB and asked patients to identify a treatment supporter who could observe each dose of medication taken.
Treatment supporters received training about all aspects of the treatment process, collected drugs each week during the intensive treatment phase and held supplies of drugs during the less intensive continuation phase of treatment.
Any patients who appeared at risk of default were visited by a community healthcare worker and encouraged to stick with their treatment.
In the control group participants received standard directly observed TB treatment with no community involvement and no follow-up of defaulters.
The study recruited 1,522 patients newly diagnosed with sputum-positive pulmonary TB, aged over 15 years (778 in the intervention group, 744 in the control group). Approximately two-thirds in each group were male and equally distributed across age ranges. Participants in the control group tended to be drawn from treatment centres with a lower previous success rate for TB treatment (median 53% cure rate compared to 68% in the intervention group).
Analysis of participant outcomes after eight months showed that patients in the intervention group were 20% more likely to be cured of TB (adjusted odds ratio 1.20) and 57% less likely to have defaulted on treatment (95% confidence interval 0.21-0.89, AOR 0.43). The intervention also significantly delayed defaulting when compared to the control group (relative risk 0.19, 95% CI 0.07-0.49).
Fifty-nine per cent of the intervention group chose to have their treatment supervised by a family member, 31% by a local nurse and 9% by the community healthcare worker. Those supervised by a family member tended to have a higher cure rate (88%) when compared to those supervised by healthcare workers (77%). Those supervised by a family member also had a lower default rate (3.9% vs 7.9%).
The authors point out that using a community-based approach allowed health workers to trace defaulters, while intensified supervision of district health centres and local health posts maintained a higher quality of follow-up.
The researchers say that one element of the intervention cannot be isolated to explain the success of the strategy, and that coherent public health strategies are often multitargeted.
“We believe that this approach may be generalised with the context of TB control programmes in other resource-poor countries,” the authors conclude.
Thiam S et al. Effectiveness of a strategy to improve adherence to tuberculosis treatment in a resource-poor setting. A cluster randomized controlled trial. JAMA 297: 380-386, 2007.