The purpose of the study is whether the provision of tuberculosis care using volunteer community health workers or self-administered treatment for 7 months is equally effective with the existing 8 months of TB care in public health facilities by health workers. Patient care by volunteer community health workers and 7 months of self-administered treatment are more patient-convenient delivery options than the ongoing TB care in health facility.

Study setting

The study will be conducted in 10 districts which fully implemented the DOTS program in all their health facilities in order to determine whether better patient quality of care and convenience using three TB care options could improve treatment outcome. In the study districts there are 149 tabias and 1163 kushets with 1.2 million residents representing 30% of the total population of the region. All public health facilities (8 hospitals, 11 health centres and 48 health posts) in these districts will be included. District hospitals and health centres provide TB diagnostic services. The study districts comprise 2316 volunteer community health workers (CHWs) who are functional and accountable to the public health system.

Study population

Patients will be recruited from the 10 study districts of all public health institutions which serve both urban and rural population. New patients diagnosed in the study districts (8 hospitals and 10 of the health centres) and those referred for treatment will be included. Patients suspected of pulmonary tuberculosis (PTB) disease will be diagnosed based on the following case definitions:

1. A patient with at least one of the following criteria will be diagnosed as smear positive PTB:

   1. Two initial smear examinations positive by direct microscopy for acid-fast bacilli (AFB) or,
   2. One initial smear examination positive by direct microscopy and culture positive, or
   3. One initial smear examination positive by direct microscopy and x-ray abnormalities suggestive of active TB as determined by the physician.

2. A patient is diagnosed as smear negative PTB when three initial smear examinations were negative by direct microscopy for AFB, and there was failure to respond to a course of broad-spectrum antibiotics followed by there were three negative smear examinations and x-ray abnormalities suggestive of active TB as determined by the treating physician.

Both smear positive (PTB+) and smear negative pulmonary (PTB-) patients will be recruited to increase the generalizability of the study.

Pre-trial measures

Interventions to improve the quality of TB care

Baseline studies were conducted to determine the quality of TB diagnosis, patient care and treatment outcome and to identify alternative options TB care appropriate to local conditions. Six months prior to the trial, the limitations of the ongoing DOTS found by these studies were communicated to district health office heads, hospital directors and district TB experts in order to strengthen the quality of TB care and management of district TB programs. Specific tools that were not developed by the National Tuberculosis Program (NTC) to ensure consistent application of diagnostic criteria, patient counselling/health information, standards on supervisory support, drug supply and district evaluation were prepared in consultation with district health management experts and hospital directors. Subsequently, district TB coordinators, health institution-based TB focal persons and clinicians were trained on patient diagnosis, patient care, community-based tuberculosis care and the study protocol. Based on our baseline assessment of quality of care, the following measures which improve patient-centred care were instituted at health facility level:

1. Quality control measures to improve the quality of diagnosis.
2. Definition of what health information should be conveyed to patients,
3. Patient teaching using a one to one interactive counselling method.
4. Training of health workers in TB management (TB focal person) with particular emphasis on patient management, management of drug side effects and interpretation of treatment outcome, and reporting.
5. Treatment in patients' most convenient health facilities.
6. Provision of daily out patient TB services in health facilities.
7. Continuous support by health workers or CHWs with a workable referral system.

Tuberculosis treatment registers were translated into the local language and were modified to fit with the monitoring and interpretation of treatment outcomes at health facility level for patients in the three treatment groups. Volunteer CHWs were also trained by their nearby health institution TB focal person on patient care, drug handling, indications for referral, managing side effects, and the use of the patient adherence card. Quarterly supervisory support was given by regional experts to district health offices, while health institutions were supervised on a monthly basis by district TB coordinators. The nearby TB focal person provided technical support to volunteer CHWs during their monthly meeting in the respective health facilities. Biannual evaluation between district health staff with HQ staff and district health staff with TB focal persons were instituted. The implementation of the above measures prior to the launching of the trial created 'operational conditions' of components of the strategy as recommended by the NTC and International Union against Tuberculosis and Lung disease (IUTL)/WHO (Wold Health Organisatio) guidelines. Strengthening the health system based on local conditions is assumed to be a basis for determining specific needs, and for creating the conditions for the development of the necessary managerial and technical skills for effective implementation of a community-based TB care.

Selection of delivery options for TB care

Four prominent TB care delivery options were identified from our community based survey: Volunteer CHWs, Public health workers, Self-administered treatment and Family member. Further assessment regarding their reliability and implications for the DOTS strategy was made using the following selection criteria:

1. Uninterrupted supply of anti-TB drugs to patients could be ensured,
2. Quality of care comparable to the prevailing health facility based TB care could be instituted,
3. Accountability in terms of guaranteeing proper use of drugs and monthly reporting of treatment outcomes of patients to the public health system,
4. Feasibility in terms of cost when applied on a larger scale and,
5. Level of acceptance by local communities as determined from our cross-sectional study.

In addition to the prevailing health facility-based TB care, the use of volunteer CHWs and self-administered treatment were found to be appropriate options for pilot testing. Based on our criteria, care by family members had the least operational feasibility because the costs involved to families (labour cost from interruptions of daily activities and transport cost) and the health systems (health worker time and cost of training) could be high. Operational difficulties were also anticipated in setting up an appropriate system for monitoring of anti-TB drugs used, provision of health education, referral, monthly reporting of treatment outcome particularly among family members who could be in remote parts of the region.

The decentralization of TB care at community level using the existing volunteer CHWs program was anticipated to be a feasible option to institute most of the components of the WHO-recommended DOTS strategy. Volunteer CHWs can be trained in patient care on a continuous basis. Accountability on anti-TB drugs used and reporting on treatment outcomes can be obtained using the existing link between CHWs and nearby public health facilities. The services they render are appreciated by respective communities. The self-administered treatment option was considered after critical examination of previous effectiveness studies, patients' convenience to tuberculosis care and, most importantly, taking into account the views expressed by communities and patients. Operational challenges, in which health workers could be forced to allow patients to take treatment by themselves when there are no volunteer CHWs who could offer TB care in remote areas was also considered.

Defining intervention arms

The final step was how optimal quality of TB care with less inconvenience could be reached among the two alternative options from patients' perspective in order to achieve optimum treatment outcome target. These integrated interventions were therefore targeted to test treatment outcome by instituting both a comparable patient quality care and patient convenience to TB care among the three strategies. In order to do this three factors were considered.

1. Epidemiological factors From the baseline study, most patients do not interrupt treatment during the intensive phase of treatment. But, the occurrence of death is high during the first 2 months of treatment. Even though its contribution is unknown in Ethiopia, the high HIV-TB co-morbidity and mortality during anti-TB therapy has been attributed in other sub-Saharan countries. Thus, proper management in health facilities during the first 28 days for patients assigned into volunteer CHW care or self-administered groups would ensure their health care needs.
2. Operational factors The intervention was designed to be implemented under operational conditions of the district health system and the ongoing community-based volunteer CHW programme. CHWs get technical and material support form nearby health facilities. They are required to submit their activity report at these facilities on a monthly basis. This arrangement has been found effective in maintaining the functionality of CHWs while minimizing disruptions in their daily life. Given this operational condition, patients have to be managed in nearby health facilities for the first four weeks until proper selection and introduction with respective CHWs are made. Other operational factors in favour of this arrangement include its feasibility when scaling up the intervention, low time cost of CHWs and the health system. This operational set up was anticipated to enable TB focal persons to organize a monthly evaluation session with CHWs in their catchments.
3. Ethical issues Health interventions should be safe and minimize potential risks leading to patient morbidity and mortality. In Ethiopia a significant proportion of tuberculosis patients are also infected with HIV. The fact that assigning such patients into an intervention whose effectiveness is unknown (volunteer CHWs) or allowing them to die (self-administer group) was also regarded as unethical.

Interventions

Accordingly, all patients should receive daily TB care including observed daily doses of treatment for the first 28 days from the most convenient health facility. Patients have to be properly counselled about their illness and managed accordingly by the TB focal persons. Counselling should be conducted using a standard procedure developed from our baseline studies in order to ensure good quality of tuberculosis care. They should be taught specifically on the causes and prevention of TB, side effects of drugs, treatment adherence and the schedule for follow up examinations. During this time, the TB focal persons need to assist patients to select volunteer CHWs from their village. They should explain their role and created confidence between them. All eligible patients should be treated using anti-TB drugs recommended by the national tuberculosis control of Ethiopia. PTB+ patients should receive weight adjusted Ethambutol (E), Rifampicin (R), Isoniazid (H) and Pyrazinamide (Z) during the first 8 weeks of treatment followed by 6 months of EH, while PTB- should be treated with weight adjusted RHZ for two months, followed by 6 months of EH.

Patients assigned to the health worker-TB care group (HW-TC) should be followed as per the national TB guideline, i.e. daily care including observed treatment by the TB focal persons, in their respective nearby health institutions during the first 8 weeks followed by self-administered treatment for the remaining 6 months. During the continuation phase patients are expected to visit their nearby health facility every month for follow up. Supervisory support will be given to TB focal persons by the respective district TB coordinators on a monthly basis. Anti-TB drugs will be delivered to health institutions on a quarterly basis by district TB coordinators during supervision.

Patients assigned to the volunteer CHW-TB care group (CHW-TC) should have daily TB care by the CHWs in their villages for the second month followed by a monthly visit to the CHWs' home for the remaining 6 months of self-administered treatment. Patients are expected to visit health facility at the end of the second months and the end of treatment for sputum examination. Technical support and anti-TB drugs should be given to CHWs by TB focal persons every fortnight.

Those assigned to the self-administered treatment group (SA-TC) should take their medication at home for seven months. They should be taught by the TB focal person to collect their anti-TB drugs fortnightly and report missed daily doses. Follow up assessment and continued support should be made by TB focal persons on a monthly basis in their nearby health facility. Patients should also be told to report when they encountered severe side effects.

Procedures

A randomized controlled design will be used to compare the existing HW-TC with CHW-TC and self-administered treatment. Eligible PTB+ and PTB- patients should be assigned using random numbers, computer generated by a statistician at regional level. To avoid disproportionate allocations among treatment groups two separate random blocks of 3 for PTB+ and PTB- are arranged. Two sets of sequentially numbered sealed envelopes for each district will be prepared. Two separate boxes containing sealed envelopes for PTB+ and PTB- will be given to the head of each district health office. The district recruiting officers should obtain the allocation arm for each patient from these district health office heads. For each district, allocation codes used should be recorded weekly by telephone. A statistician will monitor the quality of assignment by cross checking the original list in the research HQ with the sequence of patients' attendance for recruitment. Recruiting officers will be trained using a pre-tested guideline prepared to ensure that standard procedures are applied consistently in screening and informing patients. Written consent should be obtained from every patient following proper counselling, with particular emphasis on tuberculosis treatment and the alternative options of care. Patients who refused the randomly assigned method of TB care should be allowed to select either of the two remaining care options.

A request for a haemoglobin test from every eligible patient should be made by the health facility TB focal person in the diagnostic health facilities (hospitals and health centres) in order to collect serum for anonymous HIV testing before patients are sent to the district recruiting officer. Laboratory technicians should send 5 cc of serum labelled by a coding system developed by the Regional Research Laboratory (RRL) in the HQ. HIV test (Determine® HIV-1/2, Abboti Laboratories) results should be made available to the research team at the end of the trial. Baseline data should be collected from each eligible patient before he/she is referred to the most convenient health facility. Patients are only identified by code numbers in order to maintain confidentiality about their assignment. The codes should be revealed to the study team after the preliminary data analysis and draft reporting phases are completed.

Standard treatment outcome indicators developed by the WHO and the NTC will be used to compare the effectiveness of the strategies. Treatment outcome data will be collected by a team of three tuberculosis experts who are not involved in the study. Treatment outcome of PTB+ patients should be assessed by sputum microscopy for the presence of acid fast bacilli while clinical improvements and adherence to treatment will be used to assess the outcome of PTB- patients by clinicians. Cure and treatment completion rates are used to compare treatment groups. A patient is declared cured if confirmed to be sputum smear negative at the end of treatment or one month prior to the completion of treatment and on at least one previous occasion. A patient should be recorded as "treatment completed" if smear results were not available on at least two occasions or has negative pre-treatment results. A patient who has been on treatment for at least 4 weeks and whose treatment was interrupted for more than 8 consecutive weeks or for a cumulative period of more than 12 weeks should be classified as a defaulter and a patient who remains smear-positive at 5 months or later during treatment should be considered to be a failure. A patient who had started treatment for at least 4 weeks and had been transferred to another district should be recorded as transferred out.

Statistical analysis

The comparative effectiveness of health workers-TB care to that of volunteer CHW-TC and SA-TC in randomized control trials has never been reported so far. It is hypothesized that no significant difference in cure and treatment completion rate among the three care groups would be shown, taking the prevailing health facility-based TB care as the gold standard of care. The sample size for PTB+ (179 patients per arm) was planned to detect at least a 10% difference in treatment success rates between any of care groups (assuming 85% treatment success rate in the highest arm and a baseline treatment success rate of 72%) with a power of 80% at a significant level of 5%. With the same assumption (85% treatment completion rate in the highest arm and a baseline treatment completion rate of 69%), all eligible PTB- patients (131 patients per arm) should also be included during the study period. Separate comparisons of effectiveness between the three methods of TB care will be carried out for PTB+ and PTB- patients. Comparison of effectiveness of the care groups will be undertaken according to the principle of intention-to-treat. The strength of association between treatment outcomes and each care group will be determined using the chi-square test. Adjusted effects of patients' characteristics, area of residence, HIV status, time elapsed by each patient for treatment, the type of health facility and method of TB care used on treatment outcomes will be computed using logistic regression analysis.

Inclusion Criteria:

• age 15 years or above
• no previous tuberculosis treatment
• sputum exam positive for tuberculosis (PTB+); or negative sputum results but with consistent clinical and chest x-ray features (PTB-)

Exclusion Criteria:

• age less than 15 years
• previously treated with anti-TB drugs