A community-based isoniazid preventive therapy for the prevention of childhood tuberculosis in Ethiopia

BACKGROUND: Although children in contact with adults with tuberculosis (TB) should receive isoniazid (INH) preventive therapy (IPT), this is rarely implemented. OBJECTIVE: To assess whether a community-based approach to provide IPT at the household level improves uptake and adherence in Ethiopia. METHODS: Contacts of adults with smear-positive pulmonary TB (PTB+) were visited at home and examined by health extension workers (HEWs). Asymptomatic children aged <5 years were offered IPT and followed monthly. RESULTS: Of 6161 PTB+ cases identified by HEWs in the community, 5345 (87%) were visited, identifying 24 267 contacts, 7226 (29.8%) of whom were children aged <15 years and 3102 (12.7%) were aged <5 years; 2949 contacts had symptoms of TB and 1336 submitted sputum for examination. Ninety-two (6.9%) were PTB+ and 169 had TB all forms. Of 3027 asymptomatic children, only 1761 were offered (and accepted) IPT due to INH shortage. Of these, 1615 (91.7%) completed the 6-month course. The most frequent reason for discontinuing IPT was INH shortage. CONCLUSION: Contact tracing contributed to the detection of additional TB cases and provision of IPT in young children. IPT delivery in the community alongside community-based TB interventions resulted in better acceptance and improved treatment outcome.

aged ,15 years and 3102 (12.7%) were aged ,5 years; 2949 contacts had symptoms of TB and 1336 submitted sputum for examination. Ninety-two (6.9%) were PTBþ and 169 had TB all forms. Of 3027 asymptomatic children, only 1761 were offered (and accepted) IPT due to INH shortage. Of these, 1615 (91.7%) completed the 6-month course. The most frequent reason for discontinuing IPT was INH shortage. C O N C L U S I O N : Contact tracing contributed to the detection of additional TB cases and provision of IPT in young children. IPT delivery in the community alongside community-based TB interventions resulted in better acceptance and improved treatment outcome. K E Y W O R D S : tuberculosis; children; preventive therapy; health extension workers; Ethiopia CHILDREN IN CONTACT with adults with smearpositive pulmonary tuberculosis (PTBþ) have a high risk of infection and disease progression; contact investigation is therefore critical for diagnosing additional cases and preventing vulnerable individuals from progressing from infection to overt disease. Nearly all National TB Control Programmes (NTPs) recognise this risk and recommend screening contacts, especially children, for the presence of symptoms. According to international recommendations, symptomatic contacts should be investigated for active TB, whereas asymptomatic children aged ,5 years should receive isoniazid (INH) preventive therapy (IPT). 1 Although IPT reduces the risk of disease progression, 1 very few NTPs in low-and middle-income countries (LMICs) implement this recommendation. 2,3 This is because INH is often unavailable, staff have the perception that the risk of disease progression is low, that monotherapy promotes the risk of drug resistance if active TB is not excluded and due to reluctance to take on additional work. 4,5 Parents are also reluctant to give pills to their asymptomatic children and adhere poorly to the 6month course. 3,6,7 In 2010, we implemented a novel community-based approach to enhance TB case finding and treatment outcome in southern Ethiopia. This project trains female health extension workers (HEWs) to identify individuals with symptoms of TB, collect sputum specimens and prepare and fix smears for examination in 'kebeles' (equivalent to a village, with an average population of 5000 people). 8 HEWs are supported by supervisors who transport the smears to the nearest diagnostic laboratory and bring the drugs to initiate treatment at home or at the kebele health post. Contacts of adult TB cases are required to be screened for TB. However, as it was unlikely contacts would attend the health facilities or bring their children for screening, 8 we initiated contact tracing and the provision of IPT in the community through HEWs. We report the acceptability of and adherence to IPT among children in contact with adults with PTBþ identified and managed in the community by HEWs.
who had been diagnosed with PTBþ in the Sidama Zone, in the Southern Nations Nationalities and Peoples' Regional State (SNNPR) of Ethiopia. The project covered the entire zone, which has a population of 3.2 million residing in 19 rural districts and two town administrations. In this zone, three hospitals and 104 health centres provide health services to 524 rural and 39 urban kebeles. 9 The project was implemented under the Ethiopian Health Extension Programme (HEP), which has been providing health services at the village level since 2003. The HEP trains women who have completed secondary education for 1 year; HEWs are salaried cadres responsible for routinely conducting household visits to implement 16 basic health packages. 10 Adults with PTBþ reported here were diagnosed between May 2011 and March 2013, as previously described. Briefly, individuals with cough of 72 weeks were identified by HEWs during routine household visits. Symptomatic individuals were requested to provide two sputum specimens over 2 consecutive days. Smears were prepared and fixed by the HEWs and transported by supervisors to the nearest laboratory. If the smear was positive, the supervisor brought the anti-tuberculosis drugs to the village and initiated treatment at home or at the health post, depending on the patient's preference. 8 All household members were considered contacts of the index case. The index cases were asked to list all household members and their ages at the time the supervisor or HEW visited the household to disclose the diagnosis of PTBþ and initiate treatment. This list was used to ask about the presence of symptoms suggestive of TB among contacts. Houses of index cases reporting symptomatic household contacts not present at the time of the first visit were revisited to examine the contacts. However, asymptomatic contacts who were absent from the household at the time the list was prepared were not systematically revisited.
Presumptive cases among contacts were defined as individuals who reported having cough of 72 weeks with or without chest pain, shortness of breath, fever, weight loss, failure to thrive or night sweats. Presumptive cases able to expectorate were requested to submit sputum specimens. Children unable to produce sputum were referred to the nearest health facilities for further clinical examination and chest Xray (CXR). Transport subsidies were provided as needed for household members who could not afford to travel to ensure they were able to access the services. TB cases diagnosed among symptomatic contacts were provided treatment in the same manner as the index cases. Children considered to have other infections were given broad-spectrum antibiotics and were followed up by health facilities.
Asymptomatic children aged ,5 years were offered 6 months of INH at a daily dose of 5 mg/kg. Children were not tested with the tuberculin skin test or interferon-gamma release assays, as these are not locally available and are not indicated in the Ethiopian TB guidelines. 10 Parents and guardians of the children were instructed to break the tablets into two (or four) pieces to provide the closest approximate dose possible. Parents of children aged ,2 years or unable to swallow tablets were advised to crush the pieces into a powder. Parents were advised on the symptoms of TB, the importance, purpose and side effects of IPT and to seek the HEWs for advice if the child developed symptoms or side effects. Adult contacts received health education about TB, its transmission, symptoms and what to do if symptoms developed in the future.
Index cases receiving anti-tuberculosis drugs and children receiving IPT were followed by HEWs and volunteer TB treatment supporters. Follow up comprised monthly home visits or visits at the health post at the same time as the index case was receiving antituberculosis treatment. During each visit, parents/ guardians were asked by the HEWs about the presence of symptoms, adverse effects and adherence. Parents of children with minor side effects were advised to continue IPT and to immediately report any changes. Children with major IPT side effects or TB symptoms were told to discontinue IPT and were referred to the nearest hospital, where they were investigated and monitored. 'Refusal to accept IPT' was defined as a parent refusing to initiate IPT. 'IPT discontinuation' was defined as a child initiating IPT but stopping the medication for .2 months continuously in the absence of side effects or on medical advice.
Parents or guardians were informed about the importance of IPT and adherence, and children took their medication under their supervision. Parents or guardians visited their local health posts for refills and discussed any concerns about the medication, side effects and TB-related symptoms with HEWs or supervisors. Monthly meetings between HEWs, supervisors and health centre staff were held to discuss progress in the implementation of IPT in the community. Children receiving IPT were registered in health post IPT registers, and data were updated monthly and at the time of drug refills.
Semi-structured questionnaires were used for symptom screening among household contacts. TB and IPT registers were used to collect data related to diagnosis for active TB and IPT outcomes. Questionnaires were checked for completeness and consistency; data were entered by a data officer into Excele (MicroSoft, Redmond, WA, USA) and exported to SPSS for Windowse 20 (Statistical Package for the Social Sciences, Armonk, NY, USA). The main outcomes of the study were the number of children who completed 6-month IPT and the number who discontinued or were lost to follow-up. Secondary outcomes were the number of children initiating IPT who developed minor and major side effects and the number of children diagnosed with TB or who died during the 6-month follow-up. We conducted a univariate analysis for categorical variables. P , 0.05 was considered statistically significant. Children aged ,5 years were categorised by age group to identify risk factors for IPT discontinuation.
The Federal Ministry of Health of Ethiopia, Addis Ababa, and the Southern Regional Health Bureau, Hawassa, Ethiopia, provided written support for the implementation of the study. The study protocol was submitted for consideration to the Liverpool School of Tropical Medicine Ethics Committee, Liverpool, UK. The Committee waived the need for ethics approval as it considered the project was implementing internationally accepted treatment. Parents were informed that the service was part of national TB guidelines and that data monitoring was required to document the proportion of children who completed the treatment. Only 472 (93.4%) of the 523 symptomatic children aged ,15 years who were examined had symptoms recorded, compared with 100% of the 864 symptomatic adults who were examined. The most common symptoms identified in these children were cough (66.3%), fever (63.8%), shortness of breath (55.7%) and constitutional symptoms, including night sweats, loss of appetite and weight loss. Only a few children had haemoptysis. These frequencies were similar to those recorded among symptomatic adults, except for cough, which was more common among adults (P , 0.01, Table 2).

RESULTS
Of the 3027 asymptomatic children aged ,5 years who were eligible for IPT, 1761 (58%) were offered IPT (males 888, 50.4%; females 867, 49.4%; sex data were missing in 6). The remaining 1266 (42% of the 3027 eligible) were not offered IPT due to INH shortage. All parents who were offered IPT accepted the therapy.
In total, 1574 (89.3%) completed the 6-month course of IPT (Table 3). Infants appeared to be more likely to discontinue IPT than those aged 1-2 years and 3-5 years (89.0% vs. 90.8% and 92.1%, respectively), and to do so within 1 month of prophylaxis initiation. However, the frequencies were small and did not reach statistical significance. The proportion of children adhering to IPT at monthly intervals by age is shown in the Figure (age was

DISCUSSION
Despite the universal recognition of the value of the systematic screening and provision of IPT among children in contact with adults with TB, 11-14 very few NTPs in LMICs implement this service at the health facility level, 12,15,16 and even fewer NTPs are able to provide the service at the community level in poor, remote and rural communities. Reasons for not implementing contact investigation and IPT range from lack of prioritisation by NTPs, logistical problems in securing INH supplies, the time required to initiate and monitor IPT, staff perception that the risk of disease progression was minimal and the risk of developing drug resistance. 12,[17][18][19][20] We report our experience in implementing IPT at the community level among children within the framework of the Ethiopian Health Extension Programme. Children and adults exposed to adults with pulmonary TB were visited at home and questioned about the presence of symptoms; symptomatic individuals were asked to provide sputum samples. Of the 7226 children aged ,15 years, 523 were reported to be symptomatic and 35 had TB. All parents with children aged ,5 years visited by HEWs who were offered IPT accepted the therapy and attained high adherence, with .90% of the children completing the 6-month course. Despite the acceptability of the approach, an important shortcoming of the intervention was the major logistical problems faced in procuring INH despite a close partnership with the NTP. Before the intervention, the NTP had only implemented IPT at the health facility level and on a small scale, and the relatively large scale of our project led to major national procurement problems. The government also requested proxy registrations of the children at the nearest health centre before releasing the INH. This resulted in bottlenecks and a protracted process to release the drugs to the community, which took several months to resolve.
Our study had methodological limitations that must be explained to facilitate interpretation. The intervention was conducted under operational conditions, with the purpose of developing a system to provide TB services close to the community. As such, we were limited by the amount of information that could be collected to avoid overburdening the system. We were unable to obtain sputum from about 55% of the contacts who initially reported having symptoms. This was due to contacts being unable to expectorate, having no cough or because the symptoms resolved before the HEW had examined them. Furthermore, although children who could not expectorate were referred for CXR, not all parents were able to attend the clinic and some TB cases were likely missed. A further limitation is the underrepresentation of Figure Proportion of children who adhered to IPT by age group and month of treatment. IPT ¼ isoniazid preventive therapy.
selected age groups, as only 29.8% of the contacts identified were children. This is in contrast to the national population statistics, which estimate that 44% of the population is aged ,15 years. This deficit may reflect the working pattern of the HEWs, as household visits were conducted during working hours, and children may have been missed due to behaviour patterns, such as herding and school activities. Finally, child-friendly formulations, such as those that have been recently developed, would have simplified the delivery of prophylaxis, but these were not available at the time of the study. 21 The INH dose used (5 mg/kg) might also have been too low, as the World Health Organization increased the recommended dose to 7-10 mg/kg after the study had been implemented for 2 years. Despite these limitations, the approach presented here suggests that partnership with HEWs to provide community-based screening of contacts and IPT can achieve a high degree of acceptance and adherence. To our knowledge, this is the highest adherence level ever reported from this continent, 7,22 suggesting that a large component of the lack of acceptability and adherence reported by most studies may be the poor accessibility of the service.
Although these results are encouraging, a major precondition for the provision of the service without incurring major health service expenses was the availability of community-based services, with the deployment of multipurpose HEWs at the village level. As these cadres are selected from and by the community, they are likely to remain in the village and have trusting relationships with the community. Several evaluations of the programme have demonstrated their impact in other health areas, 23,24 and the acceptability of TB packages. 25 A further constraint is the inherent ineffectiveness of contact investigation as a method of early detection of TB cases, as secondary TB cases typically develop after a lag time of several weeks in the first 2 years after infection. Asymptomatic contacts could thus develop TB at a later time, and fewer secondary cases will be detected if contacts are only interviewed for symptoms immediately after the diagnosis of the index case. Furthermore, symptombased screening can miss many TB cases, as it is increasingly recognised that this method has poor sensitivity. Children receiving IPT may therefore be reexposed to TB when older members of the household develop TB, and parents need to be aware of the need to maintain vigilance for symptoms for a long time.
We had reported earlier that very few children in contact with adult TB cases initiating IPT completed the 6-month prophylaxis in the SNNPR, 7 and that adherence was poor, with very few parents continuing beyond the first few weeks. Contact investigations at health facilities are often conducted days or weeks after initiation of treatment of the index case and once the treatment has been established. This delay separates the processes of treating the patient and preventing further disease in the family, potentially contributing to lower IPT uptake and adherence. In the present study, we observed that initiating antituberculosis treatment and IPT together creates a better mind imprint of the importance of adherence, and that it is convenient for the parents to collect the drugs at the same time as receiving treatment. This temporal association could be one of the reasons for the high adherence rates attained, in addition to the client-friendly characteristics of the communitybased approach implemented, but we were unable to separate these factors.
C O N C L U S I O N : La recherche des contacts a contribuéà la détection de cas supplémentaires de TB età la fourniture de l'IPT aux jeunes enfants. La fourniture de l'IPT dans la communauté parallèlementà des interventions communautaires relativesà la TB a aboutià une meilleure acceptation du traitement età un meilleur résultat.