Decentralisation of care could improve adherence to TB treatment, Ethiopian study shows

Poor treatment adherence in Ethiopian tuberculosis (TB) patients is much more likely if patients have too far to travel to reach a clinic, especially during the continuation phase of TB treatment as they begin to feel better, Ethiopian and Norwegian researchers report in the February edition of PLoS Medicine.

Direct observation of treatment (DOT) by health workers was developed to improve TB treatment. DOT has been adopted in many countries as an integral component of the World Health Organization recommended Directly Observed Treatment-Short Course strategy (DOTS) for the prevention and control of TB. However, there are conflicting findings about the usefulness of DOT in maintaining treatment adherence. Recent studies failed to show that DOT was better than conventional non-observed treatment.

Ethiopia is one of the 22 TB high-burden countries. Before DOTS, treatment non-completion stood at 82% but it significantly declined from 38% to 18% over six years between 1994 and 2000. Despite this positive trend, one in five Ethiopian patients still default from treatment.

One of the United Nations Millenium Development Goals is the attainment of 85% treatment success for the control of TB. In order to achieve this target, it is imperative to identify and address the factors which determine treatment adherence. This is the background of the present study by a team of Ethiopian and Norwergian investigators.

The study was carried out at Hossana Zonal Hospital in Southern Ethiopia. Health facilities register and treat all diagnosed TB patients under the DOTS free of charge. All new TB cases are treated daily with a two-month short-course of chemotherapy comprising of rifampicin, isoniazid, and pyrazinamide (RHZ) plus ethambutol (E) or streptomycin (S) during the initial phase of treatment under DOT. This is followed by a continuation phase of six months on EH by self-treatment.

Retreatment cases (return after default, relapse, or treatment failure cases) are treated with SERHZ for two months, ERHZ for one month, and finally ERH for five months. Most TB patients are treated on an out-patient basis with the exception of the few patients who require hospitalisation. Patients from remote rural areas are referred to the nearest health facility for treatment initiation and follow-up.

A cohort of smear-positive pulmonary TB (PTB+) patients registered for treatment at Hassana Hospital from September 2002 to August 2004 were prospectively enrolled after obtaining informed consent. A structured interview questionnaire was used to record at the beginning of treatment all factors which might predict treatment defaulting. All patients were followed up throughout the treatment period by health care workers and the research team and treatment outcomes were recorded.

Treatment success was defined as cure (smear-negative at treatment completion and on at least one previous occasion) plus treatment completion without confirmation by smear-microscopy. A default case was defined as a PTB+ patient who had been on treatment for at least four weeks and whose treatment was interrupted for more than eight consecutive weeks. All defaulters were traced, interviewed regarding the reasons for defaulting, and counselled to resume treatment.

Of the 404 PTB+ patients enrolled, 82% (331) were from the Hadiya zone, 43% (174) were female, 21% (83) had treatment follow up at Hossana Hospital, and 78% (308) were treated at other health centres and facilities. HIV seropositivity was 12% (25/199). Overall, 81 patients (20%) defaulted from treatment while 310 (77%) successfully completed treatment. Ninety one per cent of the defaulting occurred during the continuation (non-observed) phase of treatment, with about two-thirds taking place during the first two months of this phase.

There was no significant association between defaulting from treatment and gender, average family income, HIV status, treatment center, previous knowledge about TB, and knowledge about treatment duration. The factors which were significantly associated with defaulting were age above 25 years, rural residence, education level, occupation, distance from the nearest treatment centre, and the patient’s condition at the time of treatment initiation.

On multivariate analyses, the most important independent risk factors for defaulting were physical distance from the treatment centre (HR = 2.97; p < 0.001), age over 25 years old (HR = 1.71; p =0.02), and the need to use public transport to get to a treatment centre (HR = 1.59; p = 0.06). Survival analysis demonstrated that a walking distance of more than two hours had a negative impact on the proportion of patients on treatment.

The authors speculated that distance, the strongest predictor of treatment interruption in this rural Ethiopian population, might have acted synergistically with other equally frequent reasons given for defaulting by 74 patients. These included the fact that they felt better and did not need more drugs (25.7%), they thought they had completed treatment (14.9%), and that they were too tired (weak) to walk to the TB clinic (10.8%).

The authors point out that specific measures to increase treatment completion might lie in the decentralisation of treatment follow up to community health posts, TB peer clubs, and the new community-based ‘Health Service Extension Programme’. Under the latter programme, thousands of health extension agents have been trained and the DOTS programme will be decentralized.

The major strength of this study is that it is prospective by nature and therefore does not suffer from the problem of recall bias, an important consideration in retrospective studies. Furthermore, because the predictor variables were measured at the beginning of treatment, it is possible to predict defaulting from the start. However, the authors sounded the caveat that the impact of drug side-effects, care provider-patient interaction, and HIV status was not considered.

In conclusion, physical access to health care facilities was the single most important predictor variable. The important policy implication is that in addition to improving access to treatment centres, other innovative approaches must be adopted to increase adherence to TB chemotherapy during the continuation phase of non-observed chemotherapy. This is crucial if the Millenium Development Goal for TB treatment is to be realized in Ethiopia and other TB high-burden countries.