Does decentralisation of tuberculosis care influence treatment outcomes? The case of Oromia Region, Ethiopia

Abstract

Setting: Oromia Region, Ethiopia.

Objective: To investigate the effect of decentralised care on anti-tuberculosis treatment outcomes and identify factors affecting outcome among new smear-positive tuberculosis (TB) patients.

Design: This was a retrospective cohort study comparing patients treated in the community during the continuation phase with those managed throughout treatment in health facilities. Data were collected from TB registers and patient cards using a pre-tested data capture form.

Results: Of the 2226 new smear-positive TB patients registered from July 2010 to June 2012 who were included in the study, 1599 (72.6%) received treatment in health facilities, and the rest in the community. Overall treatment success was 94.7%. Patients treated in the community had comparable treatment success with those managed in health facilities (aOR 1.7, 95%CI 0.80–3.57). Missing doses (OR 0.22, 95%CI 0.08–0.55), supervision during the continuation phase (OR 2.6, 95%CI 1.34–5.05), positive sputum at month 2 (OR 0.07, 95%CI 0.04–0.13) and human immunodeficiency virus infection (OR 0.25, 95%CI 0.13–0.46) were independent predictors of treatment success.

Conclusion: Overall treatment success is high in new smear-positive TB patients in Oromia. Patients receiving treatment in the community during the continuation phase have treatment success comparable with that of patients managed in health facilities.

The burden of tuberculosis (TB) disproportionately affects developing countries, and most of the estimated 8.6 million incident cases globally occur in Asia and Africa. The African Region has approximately one quarter of the world's TB cases, and the highest rates of cases and deaths relative to population.^1,2 Ethiopia ranks ninth among the world's 22 high TB burden countries and sixteenth among the 27 high multi-drug-resistant TB (MDR-TB) burden countries. TB kills approximately 16 000 people every year in Ethiopia.¹ Oromia, the largest region in Ethiopia, with one third of the national population, reported 37% of the national TB cases in 2012.^3,4

Community involvement in TB care and prevention, promoted by the World Health Organization (WHO) as one of the elements of the Stop TB Strategy, is considered to be important for improving adherence to anti-tuberculosis treatment.⁵ In Ethiopia, patients used to receive daily directly observed treatment (DOT) at the health facility during the first 2 months of treatment, and after conversion moved to self-administered treatment for the continuation phase of treatment. In 2010, Ethiopia shifted from an 8-month regimen to a 6-month regimen containing rifampicin throughout for new TB patients. Following the introduction of the new regimen, all patients are expected to receive daily DOT at the health facility during the intensive phase of treatment; however, DOT services during the continuation phase are decentralised to the health post/community level.

Ethiopia introduced an innovative community-based strategy, the Health Extension Programme (HEP), in 2004 to deliver preventive and promotive services as well as selected high-impact curative interventions at community level. The programme aims to improve the utilisation of health services by bridging the gap between the community and health facilities through the deployment of health extension workers (HEWs). As a preventive programme, the HEP promotes four areas of care: disease prevention and control, family health, hygiene and environmental sanitation, and health education and communication.⁶ With the decentralisation of TB care, anti-tuberculosis treatment follow-up services were integrated into the HEP programme. This allows TB patients to take their daily medicine at home or at the health post under the direct observation of the HEWs and TB treatment supporters. The HEWs are oriented in how to provide DOT and support the patient during the treatment. They are also provided with DOT monitoring tools and are supposed to observe daily drug intake at the health post or patient's home. The TB patient needs to visit the health post daily for medication under the direct observation of the HEW.

For patients who cannot attend the health post, the HEW is expected to visit the TB patient's home daily to observe the drug intake, or prepare the treatment supporter to do this if a daily home visit is not feasible due to distance. The treatment supporter is usually a family member, close friend or relative who is told how to provide DOT and how to support the patient. For patients taking their medications under the DOT supporters, the patient and the treatment supporter are supposed to visit the HEW once a week to collect drugs and report adherence status. The HEWs monitor adherence to treatment during the weekly visits by checking the TB patient treatment card and counting the pills remaining from the patient's weekly drug supply for this group of patients. The HEWs are expected to provide a monthly report to the health facility on the number of patients followed and their adherence to treatment.

It is envisaged that the involvement of the HEWs would improve TB care by reducing the workload of overstretched health facilities, enabling DOT close to home, enhancing treatment adherence and improving the treatment success rate. Under the national community TB care guidelines, all new TB patients are eligible for health post/community level anti-tuberculosis treatment services during the continuation phase, except for human immunodeficiency virus (HIV) infected and seriously ill patients, who need to be managed at health facilities.

Although the scale-up of community-based TB care in Oromia Region started four years ago, the impact of decentralisation on treatment outcomes has not been evaluated. Health care workers have questioned whether the quality of treatment follow-up services is maintained under decentralised care. This concern slowed down the scale-up of community-based TB care, and in 2013 only 45% of the health posts were providing anti-tuberculosis treatment services.³

The objectives of the present study were to determine and compare anti-tuberculosis treatment outcomes among patients receiving treatment at health facilities and at community level, and to identify factors affecting successful outcome among new smear-positive TB patients.

METHODS

The estimated population in Oromia Region was 32 million in 2013. The region has 257 rural and 53 urban districts and is administratively divided into 18 zones and 6 towns. In 2013, there were 1250 functional health facilities (48 hospitals and 1202 health centres) and 6570 health posts in the region. A health post serves a population of 5000 and is staffed by two HEWs.

The study design was a retrospective cohort study that reviewed National TB Programme (NTP) records in the region. All new smear-positive TB clients aged ⩾15 years and registered for treatment from 1 July 2010 to 30 June 2012 in selected public health facilities were included in the study. All these patients received DOT at the health facility during the intensive phase and either at the health facility or at the community level during the continuation phase of treatment.

Sample size was determined by comparing two population proportions, assuming a treatment success rate of respectively 93% and 87% among new TB patients treated at the health post/community and health facility level, based on earlier studies.^4,7 Considering a 95% certainty, 80% power, a design effect of 1.5, one third of new TB patients receiving treatment at community level, and 15% record incompleteness, a minimum sample size of 2146 was calculated using OpenEpi (Rollins School of Public Health, Emory University, Atlanta, GA, USA). Multistage sampling was used to select the health facilities. Based on geographic proximity of the 18 administrative zones in the Oromia Region, these were arranged into four groups: South, East, Central and West Oromia. One zone was randomly selected from each group. From each zone, one hospital and nine health centres were randomly selected; a total of 40 health facilities were included in the study. On average, 25–30 new smear-positive TB patients were estimated to be registered for treatment at a health facility over a 1-year period; this assumption was used to determine the number of health facilities to be included in the study. The number of TB patients to be included per facility was allocated proportionally to the total number of new smear-positive TB patients registered between 1 July 2010 and 30 June 2012 in the selected health facilities. Routine TB programme data were used to obtain the total number of new smear-positive TB patients registered in each of the selected health facilities over the study period, which was used as a denominator to determine the proportion of patients to be included from each facility. New smear-positive TB patients were recruited consecutively as study subjects from 1 July 2010 onwards in each health facility until the allocated sample size was reached.

Data were extracted from the TB Unit registers and TB patient treatment cards using a pre-tested, structured data collection tool. Information was collected on age, sex, residence, treatment follow-up site, missed doses, sputum results at month 2, weight at treatment initiation and during follow-up periods, HIV status, TB-HIV services (antiretroviral therapy and cotrimoxazole preventive therapy) and treatment outcome. During analysis, outcomes were categorised as successful (cured and treatment completed) and unsuccessful. Treatment follow-up sites were categorised into health facility level (health centre and hospital) and community level (community and health post). Each questionnaire was checked for completeness, coded and entered using Epi Info™ version 3.5.4 (Centers for Disease Control and Prevention, Atlanta, GA, USA). Data were cleaned and exported to SPSS version 20 (Statistical Package for the Social Sciences, Chicago, IL, USA) for analysis.

Univariate analysis was performed to describe the demographic and clinical characteristics of the study subjects and determine treatment outcomes in the two groups. Bivariate analysis was used to assess the association between anti-tuberculosis treatment success and independent variables. Crude odds ratios (ORs) with 95% confidence intervals (CIs) were calculated. The χ² test was used to compare proportions. To assess the independent effects of each variable, a multivariate logistic regression was performed and adjusted ORs were calculated.

Ethical clearance was obtained from the Oromia Regional Health Bureau Ethical Review Board, Addis Ababa, Ethiopia.

RESULTS

Demographic and clinical characteristics of study subjects

Of the 2226 new smear-positive TB patients registered for treatment during the study period, 1559 (72.6%) were followed up at health facilities during the continuation phase, while 604 (27.4%) received community TB care. The median age among the study participants was 27 years, with a standard deviation of 12.4 years. Age and sex distribution did not differ between treatment groups. More patients from rural areas were receiving care at community level, while HIV-positive patients received care more frequently at health facility level (Table 1).

TABLE 1.

Demographic and clinical characteristics of new smear-positive tuberculosis patients receiving treatment at community and health facility levels, Oromia Region, July 2010–June 2012

Treatment outcome

Treatment outcome was not known for 69 (3.1%) patients due to recording problems, and these were excluded from the analysis. The majority (89%) of the patients with missing treatment outcome were followed at a health facility. Overall treatment success was 94.7%, which was slightly higher in patients managed at community level compared to those at health facilities (97.8% vs. 93.5%, P < 0.001) (Table 2). The proportion cured was similar between the two groups, while more patients completed treatment at community level. Fewer patients died or defaulted among those followed at community level. Adherence to DOT was respectively 95.3% and 89.4% among patients treated at community and health facility level (Table 3).

TABLE 2.

Treatment outcomes among new smear-positive tuberculosis cases treated at community and health facility levels, Oromia Region, July 2010–June 2012

TABLE 3.

Adherence and follow-up status among new smear-positive tuberculosis patients receiving treatment at community and health facility levels, Oromia Region, July 2010–June 2012

Factors affecting anti-tuberculosis treatment outcome

In multivariate analysis, there was no statistically significant association between TB treatment success and treatment site (OR 1.7, 95%CI 0.80–3.57). Treatment success was lower among HIV-positive patients (OR 0.25, 95%CI 0.13–0.46, P < 0.001), patients who remained sputum smear-positive at the end of month 2 (OR 0.07, 95%CI 0.04–0.13) and among patients who reported missed doses (OR 0.22, 95%CI 0.08–0.55) during the continuation phase of treatment. Treatment success was higher among patients who had received DOT during the continuation phase of treatment than patients whose treatment was self-administered (OR 2.6, 95%CI 1.34–5.05) (Table 4).

TABLE 4.

Factors associated with treatment success among new smear-positive tuberculosis patients in Oromia Region, July 2010–June 2012

DISCUSSION

Overall TB treatment success among new smear-positive TB patients included in this study was 94.7%, which is higher than the nationally reported figure and above the international target of 85%.^3,8 This study showed that patients treated at community level had treatment success at least comparable to that of those treated at the health facility. Treatment follow-up site was not found to be significantly associated with treatment success. Comparable treatment success among patients receiving community-based DOT and health facility-based DOT was also observed in studies conducted in Tanzania^9,10 and Uganda.¹¹ This suggests that, in terms of treatment success, it is not necessary to keep TB patients at the health facility for treatment supervision by health workers during the continuation phase of treatment: a trained HEW and TB treatment supporter can execute this duty at least as effectively at community level. These findings encourage the scale-up of decentralisation of TB treatment services to improve access to TB care for the rural community.

Moreover, decentralisation of anti-tuberculosis treatment services to the community seems to have improved adherence to treatment, as observed by better adherence to DOT during the continuation phase of treatment and fewer missed doses among patients receiving treatment at community level. Patients with a tendency to interrupt treatment at health facility level are often not followed up as efficiently as at community level with an HEW who has more contact with the community through health development army female volunteers. This finding is consistent with a study conducted in Brazil, where patients offered community-based treatment were more likely to accept DOT than those offered clinic-based treatment.¹²

Our results need to be interpreted with caution, however. Although the TB Unit registers provided nearly complete data for most of the outcomes of interest for this study, they did not provide information about whether patients managed by health facilities received their anti-tuberculosis treatment at the health centre or at the hospital throughout the course of treatment or if they were referred to community-based care at a later stage of treatment. In addition, treatment follow-up site for patients receiving their treatment at community level is expected to be documented in the remarks section of the TB register and is not included as one of the variables in the register. In this study, only patients whose treatment follow-up site was documented in the TB Unit register or patient card as ‘health post’ were included in the community group. Some patients who were followed up at community level may therefore have been misclassified under ‘facility-based care’ group if the health care workers failed to record the follow-up site. This may have diluted the health facility group with decentralised patients, and most likely reduced the observed difference in treatment success among the two groups. A modification of the TB Unit register to document treatment adherence support site should be considered. A study to ascertain whether or not community TB care is better and more patient-centred than facility-based care should be conducted.

As reported in other studies, HIV co-infection among new TB patients is independently associated with unsuccessful treatment outcome.^13–15 Consistent with findings from other studies, receiving DOT during the continuation phase of treatment, having instances of missed doses, and sputum conversion at the completion of the intensive phase of treatment were found to be strong predictors of treatment success among new smear-positive TB patients.^16–19

In conclusion, overall treatment success is high among new smear-positive TB patients in the Oromia Region. This study showed that patients receiving their treatment at community level and in health facilities have comparable treatment success. The costs incurred and efforts to visit the health facility daily are worth considering when prioritising good quality community care instead of focusing on facility-based DOT. TB treatment services can be decentralised to the community level to ensure accessible services while maintaining high treatment success in the region.