Predictors of Treatment Outcomes of Childhood Tuberculosis among Patients Attending the TB Clinic at Mbagathi County Hospital, Nairobi County, Kenya

Abstract

Tuberculosis (TB) is a serious worldwide health problem that wreaks havoc on health systems, communities, and economies worldwide. The World Health Organization (WHO) ranked tuberculosis (TB) as the thirteenth biggest cause of death and the world’s second most infectious killer after corona virus disease (COVID-19) (Chilyabanyama et al., 2024). Kenya ranks among the global 30 high TB burden countries which accounted for 87% of the world’s cases. Unlike studies on TB in adults, less is published on the predictors of treatment outcomes in children. The aim of TB treatment policy is to cure patients and therefore alleviate suffering and prevent death from the disease. It’s also aimed at preventing long-term complications arising from the disease and prevent relapse. Treatment is also aimed at preventing the transmission of the infection and development of drug resistance. Benefits of TB treatment is attributed to both the individual patient, family and the community as a whole. Outcomes of treatment is a good indicator of performance of the TB program. The broad objective of the study was to determine predictors of treatment outcomes amongst children registered for TB treatment between 1st January, 2018 and 31st December, 2020 at Mbagathi County Hospital. A cross-sectional study design was used that utilised secondary data of children registered for treatment of TB during the three-year period under review from the facility TB register. Data from all 126 children aged < 15 years registered in the years under review was analysed. A structured pre-coded data abstraction tool was used to record patient and clinical variables such as age, sex, weight, nutritional status, type of TB, TB sub-type, Genexpert test, HIV test and treatment outcomes. Data was analysed using Statistical Package for Social Sciences (SPSS) version 27 tool. Descriptive variables were analysed for frequencies. Fischer’s exact test was carried out to determine the association and significance between the predictor and outcome variables. Bivariate analysis was carried out to identify the strength of the relationship of the independent variables and the treatment outcome. Statistical significance was considered at p-value <0.05. The findings of the study showed that the proportion of males and females was comparable at 62 (49.2%) and 64 (50.8%), respectively. Of these children, 64 (50.8%) were aged below one year and among 114 assessed for nutritional status, 47 (41.2%) had severe acute malnutrition. Among all study subjects, 84 (66.7%) had pulmonary tuberculosis, while 39 (31.0%) had extra-pulmonary tuberculosis; patients with miliary tuberculosis were classified as pulmonary TB cases. HIV testing was conducted in 115 (91.2%) children, with an HIV positivity rate of 28 (22.2%). Genexpert testing was performed in 54 (42.9%) children, of whom 30 (55.6%) had Mycobacterium tuberculosis detected. Good treatment outcomes were observed in 68 (53.9%) of the children. Among those aged between one year and less than five years, 13 out of 19 (68.4%) achieved good treatment outcomes, the highest proportion across age groups. Of the 47 children with severe acute malnutrition, 29 (61.7%) had good treatment outcomes. Among 116 children whose DOT supporter was a household member, 63 (54.4%) achieved good treatment outcomes. Children with pulmonary tuberculosis had better outcomes, with 50 (61.0%) achieving favourable results compared to 18 (43.9%) of the 41 children with extra-pulmonary tuberculosis. Binary logistic regression was used to determine whether the patient and clinical level predictors were associated with the likelihood of having favourable/good TB treatment outcome. Out of all variables, type of TB was the only one that significantly contributed to the model. Children below one year of age contributed the highest TB burden and malnutrition being a very important factor associated with TB disease. Pulmonary TB remained the predominant type of the disease and HIV positivity rate was double the national scale. Most children were self-referrals and those whose treatment outcomes were not evaluated contributed a high proportion of poor treatment outcomes. It is recommended that a good referral framework is implemented so as capture data particularly for treatment outcomes. The more representative multi-centre research is also recommended.