Advertisement

Factors associated with unfavorable treatment outcomes among pediatric tuberculosis cases in Harare, Zimbabwe during 2013–2017

Open AccessPublished:September 02, 2020DOI:https://doi.org/10.1016/j.ijid.2020.08.079

      Highlights

      • Factors contributing to unfavorable TB treatment outcomes are often unknown in high-burden areas.
      • Determinants of pediatric TB treatment outcome were assessed in Harare, Zimbabwe.
      • Only 57% of the study patients had a favorable treatment outcome.
      • Treatment outcome differed significantly according to the socioeconomics of the healthcare setting.
      • Positive pretreatment sputum smear was associated with unfavorable TB outcome.

      Abstract

      Background

      Historical neglect of pediatric tuberculosis (TB), compounding the absence of a universally effective vaccine, highlights the importance of successful treatment in combating the global epidemic. Furthermore, compliance with international standards of pediatric TB treatment remains unknown in many high-burden, resource-limited settings.

      Methods

      In this cross-sectional study, using TB surveillance data, we assessed the treatment outcomes among 853 pediatric TB cases (<15 years old), a study sample that represented all the pediatric TB cases with treatment outcome records in Harare, Zimbabwe during 2013–2017. We also identified factors associated with treatment outcome by multivariate logistic regression.

      Results

      Of these 853 analyzed cases, 57% were either cured or had completed treatment. In a model accounting for confounding variables, hospital center and pretreatment sputum smear were associated with unfavorable treatment outcome. Cases from Beatrice Road Infectious Disease Hospital were four times as likely to have an unfavorable outcome compared with those from Wilkins Infectious Disease Hospital (adjusted odds ration [aOR]: 4.0; 95% CI 2.9–5.5). Children whose pretreatment sputum smear was positive were 2.4 times as likely to have an unfavorable outcome as those who were negative (aOR: 2.4; 95% CI 1.7–3.6).

      Conclusion

      Pediatric TB case management needs to be improved, especially among those with a positive pretreatment sputum smear. Efforts to address TB treatment outcome disparities between clinical settings in high-burden settings, such as Harare, Zimbabwe, are essential in improving global TB control.

      Keywords

      Introduction

      Even with a heightened global approach to combating the tuberculosis (TB) epidemic, the number of TB incident cases in 2018 was estimated to be 10.0 million (
      • World Health Organization
      Global tuberculosis report 2019.
      ). There were over 1.5 million deaths among cases of TB, making it the leading cause of mortality among infectious agents (
      • World Health Organization
      Global tuberculosis report 2019.
      ). Of these deaths, 14% occurred among those under the age of 15 (
      • World Health Organization
      Global tuberculosis report 2019.
      ). Furthermore, it has been estimated that only 33% of the 1.12 million new cases of pediatric tuberculosis were actually diagnosed (
      • World Health Organization
      Global tuberculosis report 2019.
      ,
      • Jenkins H.E.
      Global burden of childhood tuberculosis.
      ). A staggering 87% of incident cases of childhood TB are confined to 30 high-burden countries, many of which are located in Sub-Saharan Africa and classified as lower- or middle-income countries (LMICs) (
      • Jenkins H.E.
      Global burden of childhood tuberculosis.
      ,
      • Starke J.R.
      • Cruz A.T.
      The global nature of childhood tuberculosis.
      ). In addition, pediatric TB has suffered from historical neglect in all facets of public health, including research, reporting, and surveillance (
      • World Health Organization
      ,
      • Graham S.M.
      • Sismanidis C.
      • Menzies H.J.
      • Marais B.J.
      • Detjen A.K.
      • Black R.E.
      Importance of tuberculosis control to address child survival.
      ). The paucibacillary nature of pediatric TB has led to misperceptions about its role in the global TB epidemic, and causes unique challenges to diagnosing pediatric TB (
      • Marais B.J.
      • Schaaf H.S.
      Childhood tuberculosis: an emerging and previously neglected problem.
      ,
      • Newton S.M.
      • Brent A.J.
      • Anderson S.
      • Whittaker E.
      • Kampmann B.
      Paediatric tuberculosis.
      ). Nevertheless, the world is now paying more attention than ever to childhood TB, a fact highlighted by both the World Health Organization (WHO)’s ambitious END TB Strategy and the UN’s Sustainable Development Goals (SDGs), specifically 3.2.1 (
      • Falzon D.
      • Schünemann H.J.
      • Harausz E.
      • Gonzalez-Angulo L.
      • Lienhardt C.
      • Jaramillo E.
      • et al.
      World Health Organization treatment guidelines for drug-resistant tuberculosis, 2016 update.
      ,
      • UN General Assembly
      Transforming our world: the 2030 agenda for sustainable development.
      ). In order to successfully end the global TB epidemic, more emphasis needs to be placed on treatment success of vulnerable populations, like children, and evaluation of treatment and reporting adherence in high-burden settings, especially in the absence of a universally effective vaccine (
      • Jenkins H.E.
      Global burden of childhood tuberculosis.
      ,
      • Starke J.R.
      • Cruz A.T.
      The global nature of childhood tuberculosis.
      ,
      • World Health Organization
      ,
      • World Health Organization
      Global tuberculosis report 2017.
      ,
      • World Health Organization
      Global strategy and targets for tuberculosis prevention, care and control after 2015 report by the secretariat.
      ,
      • Oeltmann J.E.
      • Chengeta B.
      • Mboya J.J.
      • Wells C.D.
      • Kilmarx P.H.
      • Samandari T.
      • et al.
      Reported childhood tuberculosis treatment outcomes, Gaborone and Francistown, Botswana, 1998–2002.
      ,
      • Adejumo O.A.
      • Daniel O.J.
      • Adebayo B.I.
      • Adejumo E.N.
      • Jaiyesimi E.O.
      • Akang G.
      • et al.
      Treatment outcomes of childhood TB in Lagos, Nigeria.
      ).
      Treatment of children is essential as they are often the first to become ill with active TB and experience more severe symptoms and negative outcome of disease than adults, if untreated (
      • Starke J.R.
      • Cruz A.T.
      The global nature of childhood tuberculosis.
      ,
      • Hailu D.
      • Abegaz W.E.
      • Belay M.
      Childhood tuberculosis and its treatment outcomes in Addis Ababa: a 5-years retrospective study.
      ). The risk of mortality that accompanies non-adherence is higher in pediatric cases than in adults and largely understated, especially for those with HIV-coinfection (
      • Jenkins H.E.
      Global burden of childhood tuberculosis.
      ,
      • Starke J.R.
      • Cruz A.T.
      The global nature of childhood tuberculosis.
      ,
      • van Zyl S.
      • Marais B.J.
      • Hesseling A.C.
      • Gie R.P.
      • Beyers N.
      • Schaaf H.S.
      Adherence to anti-tuberculosis chemoprophylaxis and treatment in children.
      ). Furthermore, failure to adhere to treatment results in persistent transmission and may give rise to drug resistance (
      • Namukwaya E.
      • Nakwagala F.N.
      • Mulekya F.
      • Mayanja-Kizza H.
      • Mugerwa R.
      Predictors of treatment failure among pulmonary tuberculosis patients in Mulago Hospital, Uganda.
      ). However, few studies have been conducted to identify clinical and demographic predictors of favorable outcomes of pediatric TB (
      • Carlucci J.G.
      • Blevins Peratikos M.
      • Kipp A.M.
      • Lindengren M.L.
      • Du Q.T.
      • Renner L.
      • et al.
      Tuberculosis treatment outcomes among HIV/TB-coinfected children in the International Epidemiology Databases to evaluate AIDS (IeDEA) network.
      ).
      The WHO has published detailed recommendations for treatment initiation for any probabilistic cases of TB in LMICs, in an effort to achieve targets laid out by the SDGs and WHO’s END TB Roadmap (
      • World Health Organization
      Global tuberculosis report 2019.
      ,
      • World Health Organization
      ). The primary goals of treatment are curing the patient and stopping further transmission. It is also highly recommended that treatment commences immediately with fixed-dose combinations using standardized regimens until treatment completion. While treatment of TB in children is similar to that for adults, the level of adherence to international guidelines, such as clinical and laboratory monitoring as well as documentation for every pediatric case from diagnosis through to treatment completion, is largely absent from the literature, especially in LMICs (
      • Carlucci J.G.
      • Blevins Peratikos M.
      • Kipp A.M.
      • Lindengren M.L.
      • Du Q.T.
      • Renner L.
      • et al.
      Tuberculosis treatment outcomes among HIV/TB-coinfected children in the International Epidemiology Databases to evaluate AIDS (IeDEA) network.
      ,
      • Zimbabwe Ministry of Health and Child Care
      National tuberculosis control program guidelines.
      ).
      To address these important knowledge gaps in pediatric TB treatment, we conducted a cross-sectional study to determine the treatment success rate among pediatric TB cases and to identify demographic and clinical characteristics associated with unfavorable pediatric TB treatment outcome in Harare, Zimbabwe during 2013–2017, using TB surveillance data.
      Zimbabwe is one of the 30 high TB/HIV high-burden countries, with an estimated 13.2% of all TB cases occurring in those under the age of 15 (
      • World Health Organization
      Global tuberculosis report 2019.
      ,
      • World Health Organization
      Global tuberculosis report 2017.
      ). Furthermore, it was reported by Earley et al., that over half (58%) of children with TB in the capital city, Harare, tested positive for HIV (
      • Earley M.
      • Chirenda J.
      • Highet A.
      • Mujuru H.A.
      • Yang Z.
      Characterizing pediatric tuberculosis with and without human immunodeficiency virus coinfection in Harare, Zimbabwe.
      ). However, treatment outcome was not investigated in Earley’s study. The study setting, therefore, provides a unique opportunity to address the knowledge gaps in pediatric TB treatment.

      Methods

      Study sample and data collection

      The study sample included a total of 853 cases of pediatric TB. This represented 58% of the total number (1460) of incident cases of TB diagnosed among children under the age of 15 years residing in Harare, Zimbabwe during 2013–2017, for which treatment outcome records were available. Diagnosis of TB was based on the history of contact with suspected TB cases, physical examination looking for signs and symptoms suggestive of TB, chest X-ray, tuberculin skin test, and confirmatory tests including sputum smear microscopy, in line with Zimbabwe National Guidelines (
      • Zimbabwe Ministry of Health
      National guidelines on management of tuberculosis in children.
      ). Inclusion criteria required residency in Harare, having a diagnosis of TB according to Zimbabwe National Guidelines, and a record of treatment outcome.
      Data for the study patients were derived from the national TB surveillance dataset, which was collected mainly through the two major infectious disease hospitals in Harare. Beatrice Road Infectious Disease Hospital (BRIDH) and Wilkins Infectious Disease Hospital (WIDH) diagnose over 90% of TB cases in Harare (
      • Earley M.
      • Chirenda J.
      • Highet A.
      • Mujuru H.A.
      • Yang Z.
      Characterizing pediatric tuberculosis with and without human immunodeficiency virus coinfection in Harare, Zimbabwe.
      ). Only 4% of TB cases were diagnosed and treated at satellite clinics throughout Harare (59/460). Data used for the study included basic patient demographic characteristics, such as age and sex, and clinical information, including treatment center, location of disease, and pretreatment sputum smear microscopy (SSM) result, HIV-serological test result, and treatment outcome. This study was approved by the Institutional Review Board of the University of Michigan for the Health and Behavior Sciences and the Joint Review Ethics Committee of the University of Zimbabwe.

      Patient classification

      The study patients were classified into favorable and unfavorable outcome groups based on their treatment outcome, according to WHO Guidelines (
      • World Health Organization
      Definitions and reporting framework for tuberculosis – 2013 revision – updated December 2014 and January 2020.
      ). Patients defined as having favorable outcomes included those cases who had been cured or had completed the recommended standard treatment. Cases with recorded outcomes of either lost to follow-up, not evaluated, experienced treatment failure, or died were classified as unfavorable outcomes (
      • Carlucci J.G.
      • Blevins Peratikos M.
      • Kipp A.M.
      • Lindengren M.L.
      • Du Q.T.
      • Renner L.
      • et al.
      Tuberculosis treatment outcomes among HIV/TB-coinfected children in the International Epidemiology Databases to evaluate AIDS (IeDEA) network.
      ,
      • World Health Organization
      Definitions and reporting framework for tuberculosis – 2013 revision – updated December 2014 and January 2020.
      ). In addition, the study patients were grouped into the three age categories that WHO used in its Global Tuberculosis Report (
      • World Health Organization
      Global tuberculosis report 2019.
      ). Based on the site of disease, the patients were classified as having pulmonary TB (PTB) if involving the lungs and extrapulmonary tuberculosis (EPTB) if involving sites other than the lungs (
      • World Health Organization
      Global tuberculosis report 2019.
      ).

      Statistical analysis

      We first computed descriptive statistics for demographic and clinical attributes of the study sample (n = 853). In an effort to assess potential sampling bias, we compared demographic and clinical characteristics between the 853 (58.4%) cases with a reported outcome and the remaining (41.6%) of cases without a reported outcome (Table 1). In order to identify patient characteristics associated with treatment outcome, we used the 853 cases for which treatment outcome records were available and calculated TB treatment outcome rate stratified according to various patient demographic and clinical characteristics. We tested the associations between treatment outcome and each of the patient characteristics using Pearson’s chi-square test and bivariate logistic regression. To control for potential confounders, we next performed multivariate logistic regression to obtain adjusted odds ratios (aORs). An alpha level of 0.05 was used to assess significance. To improve the efficiency of our multivariate logistic regression, multiple imputations were used for the missing covariate data. The rate of favorable treatment outcome was determined for the 853 cases with a reported outcome. To assess the potential effect of missing outcome data on the calculation of favorable treatment outcome rate, we also used multiple imputations for the missing outcome data. Ten multiple imputed datasets were generated to fill in missing covariates and outcome values using a sequence of regression models (
      • Raghunathan
      • Lepkowski
      • Van Hoewyk
      • Solenberger
      A multivariate technique for multiply imputing missing values using a sequence of regression models.
      ). Rubin’s rule was then utilized to pool estimates and confidence intervals (
      • Rubin D.B.
      Multiple imputation for nonresponse in surveys.
      ;
      • Little R.J.A.
      • Rubin D.B.
      Statistical analysis with missing data.
      ). Multiple imputation was conducted using IVEware 0.3 (University of Michigan Institute for Social Research), while all other analyses were carried out using SAS 9.4 (SAS Institute, Inc., Cary, NC).
      Table 1Comparison of patient characteristic distributions between cases with and without a recorded outcome among 1460 incident cases of pediatric TB diagnosed in Harare, Zimbabwe, during 2013–2017.
      CharacteristicCases with a recorded outcomeCases without a recorded outcomep-Value
      Pearson’s chi-square test.
      n = 853n = 607
      n (%)n (%)
      Sex0.68
       Male459 (53.81)320 (52.72)
       Female394 (46.19)287 (47.28)
      Age0.12
       0–4525 (61.55)374 (61.61)
       5–9143 (16.76)122 (20.10)
       10–14185 (21.69)111 (18.29)
      HIV status0.25
       Negative328 (38.45)253 (41.68)
       Positive405 (47.45)274 (45.14)
       Missing120 (14.10)80 (13.18)
      Case type0.27
       New824 (96.60)593 (97.69)
       Previous TB28 (3.29)14 (2.31)
       Missing1 (0.11)
      Diagnosis0.12
       PTB714 (83.70)526 (86.66)
       EPTB139 (16.30)81 (13.34)
      Sputum<0.01
       Negative336 (39.39)174 (28.67)
       Positive70 (8.21)85 (14.00)
       Missing447 (52.40)348 (57.33)
      Centre<0.01
       BRIDH500 (58.62)560 (92.26)
       WIDH340 (39.86)1 (0.16)
       Clinics13 (1.52)46 (7.58)
      a Pearson’s chi-square test.

      Results

      Study sample characteristics

      Among the 853 study cases, the majority (525/853, 62%) were under the age of 5, with the remaining 338 (38%) cases being equally distributed between the two older age groups (Table 1). Fifty four percent (459/853) of cases were male. Additionally, 120 cases did not have a documented HIV diagnosis, while HIV-coinfection was reported for 405 (55%) of the 733 cases for which HIV seropositivity data were available (Table 1).
      Notably, the diagnosis and treatment of TB cases were not evenly distributed throughout Harare. BRIDH diagnosed and treated 59% of incident cases during the study period, while cases diagnosed and treated by WIDH accounted for 40% of the total pediatric TB patient population during the study time period; the remaining 1% of cases were diagnosed at local clinics throughout Harare. Sputum smear microscopy was reported for just under half of the study sample (406/853), of which 17% (70) were reported as positive (Table 1).

      Treatment outcome

      The rate of favorable treatment outcome was 57% (485/853) among the patients with a reported outcome and 50% (731/1460) among all the patients based on imputed outcome data. Chi-square test results showed that there were no differences in treatment outcome in relation to age, sex, HIV status, history of TB, or site of disease. The group with unfavorable treatment outcome was overrepresented by cases from BRIDH compared with WIDH and other clinics (74% vs 23% and 3%, respectively; p < 0.001).In addition, the proportion of sputum smear positive cases was higher among patients with unfavorable outcomes compared with those with favorable outcomes (10% vs 7%; p = 0.08) (Table 2).
      Table 2Comparison of distribution of patient characteristics between cases with an unfavorable treatment outcome and cases with a favorable outcome among 1460 incident cases of pediatric TB diagnosed in Harare, Zimbabwe, during 2013–2017.
      CharacteristicUnfavorable outcomeFavorable outcomep-Value
      Pearson’s chi-square test, carried out prior to multiple imputation.
      n = 368n = 485
      n (%)n (%)
      Sex0.49
       Male203 (55)256 (53)
       Female165 (45)229 (47)
      Age category0.71
       0–4222 (60)303 (62)
       5–966 (18)77 (16)
       10–1480 (22)105 (22)
      HIV status0.95
       Positive141 (38)187 (39)
       Negative175 (48)230 (47)
       Missing52 (14)68 (14)
      History of TB0.13
       New352 (96)472 (97)
       Previous TB16 (4)12 (3)
       Unknown1 (0)
      Diagnosis0.26
       PTB314 (85)400 (82)
       EPTB54 (15)85 (18)
      Sputum0.08
       Negative152 (41)184 (38)
       Positive37 (10)33 (7)
       Missing179 (49)268 (55)
      Centre<0.01
       BRIDH272 (74)228 (47)
       WIDH83 (23)257 (53)
       Clinics13 (3)
      a Pearson’s chi-square test, carried out prior to multiple imputation.

      Characteristics associated with unfavorable treatment outcome

      After adjusting for confounding variables, multivariate logistic regression revealed that hospital location and pretreatment sputum smear positivity were significantly associated with treatment outcome (Table 3). The odds of having an unfavorable treatment outcome for cases diagnosed at BRIDH was significantly higher than that for cases diagnosed at WIDH (aOR: 4.0; 95% CI 2.9–5.4). In addition, the odds of unfavorable treatment outcome for sputum smear positive (SSP) cases was significantly higher than that for cases who were sputum smear negative (SSN) (aOR: 2.4.; 95% CI: 1.7–3.6). No other clinical or demographic factors were found to be significantly associated with unfavorable treatment outcome.
      Table 3Clinical and demographic factors associated with unfavorable treatment outcomes among pediatric TB cases in Harare, Zimbabwe during 2013–2017 (n = 853
      Only cases with recorded treatment outcomes were included in analyses. All analyses used the multiple imputations method.
      ).
      CharacteristicUnadjusted odds ratio (95% CI)p-Value
      Bivariate logistic regression.
      Adjusted odds ratio (95% CI)p-Value
      Multivariate logistic regression.
      Sex
       MaleRefRef
       Female0.9 (0.7–1.2)0.490.9 (0.7–1.2)0.49
      Age category
       0–4RefRef
       5–91.2 (0.8–1.7)0.411.2 (0.8–1.8)0.35
       10–141.0 (0.7–1.5)0.821.3 (0.9–2.0)0.19
      HIV status
       NegativeRefRef
       Positive1.0 (0.8–1.4)0.870.7 (0.4–1.2)0.20
      History of TB
       NewRefRef
       Previous TB1.8 (0.8–3.8)0.131.4 (0.6–3.3)0.43
      Diagnosis
       PTBRefRef
       EPTB0.8 (0.6–1.2)0.260.9 (0.6–1.3)0.59
      Sputum
       NegativeRefRef
       Positive2.5 (1.8–3.6)<0.0012.4 (1.7–3.6)<0.001
      Centre
       WIDHRefRef
       BRIDH3.7 (2.7–5.0)<0.0014.0 (2.9–5.4)<0.001
       Clinics
      Eliminated from logistic regression.
      * Only cases with recorded treatment outcomes were included in analyses. All analyses used the multiple imputations method.
      a Bivariate logistic regression.
      b Multivariate logistic regression.
      c Eliminated from logistic regression.

      Discussion

      While there has been a recent study discussing trends in pediatric TB and treatment outcomes (
      • Nzombe P.
      • Satyanarayana S.
      • Tweya H.
      • Timire C.
      Declining trends in childhood tb notifications and profile of notified patients in the city of Harare, Zimbabwe, from 2009 to 2018.
      ), our study provides the first assessment of factors associated with having an unfavorable treatment outcome among pediatric cases of TB in the high burden setting of Harare, Zimbabwe. It also highlights the drivers of the aforementioned trends, including the continued struggle to comply with international treatment and data management standards in TB endemic settings with limited resources. The findings of our study represent an opportunity to address the global burden of TB by focusing on pediatric epidemiology specific to Harare, Zimbabwe (
      • Houben R.M.G.J.
      • Menzies N.A.
      • Sumner T.
      • Huynh G.H.
      • Arinaminpathy N.
      • Goldhaber-Fiebert J.D.
      • et al.
      Feasibility of achieving the 2025 WHO global tuberculosis targets in South Africa, China, and India: a combined analysis of 11 mathematical models.
      ). As mentioned by Houben et al., understanding of the local burden and its contribution to overall disease is an essential tool for achieving global reduction (
      • Houben R.M.G.J.
      • Menzies N.A.
      • Sumner T.
      • Huynh G.H.
      • Arinaminpathy N.
      • Goldhaber-Fiebert J.D.
      • et al.
      Feasibility of achieving the 2025 WHO global tuberculosis targets in South Africa, China, and India: a combined analysis of 11 mathematical models.
      ).
      Our study resulted in several crucial findings surrounding pediatric TB treatment outcome. Two important variables were found to be statistically significantly associated with unfavorable treatment outcome, after controlling for confounders. These included the hospital setting where the study cases of TB were diagnosed and the baseline sputum smear positivity.
      In this study, being diagnosed at WIDH was a significant predictor for having a favorable treatment outcome; this was in contrast to being diagnosed at BRIDH. The relationship between hospital setting and treatment outcome may reflect the differences in workload and patient populations served between healthcare settings. It is indeed very true that BRIDH has a much higher workload than WIDH. Historically, BRIDH has served a more economically disadvantaged subpopulation of Harare than WIDH. These trends are further reflected by the fact that the majority of cases seen at BRIDH did not have an outcome recorded, whereas only one case from WIDH did not have a recorded outcome. BRIDH also had a much higher rate of missing data for the clinical and demographic characteristics collected. Furthermore, the BRIDH registries that contribute to the national TB surveillance had been sporadically updated for various reasons, among which were high patient traffic and low staff numbers. These play an important role in the failure to properly update data and lead to losses in follow-up and further treatment compliance issues (
      • World Health Organization
      ,
      • Zimbabwe Ministry of Health and Child Care
      National tuberculosis control program guidelines.
      ). This may also explain the unusually high proportion of children with no record of HIV status.
      This study used hospital location as a proxy for patient demographics in Harare and concurred with previous studies that have shown location of clinic and/or hospital to play a role in treatment outcome across high-burden settings (
      • Namukwaya E.
      • Nakwagala F.N.
      • Mulekya F.
      • Mayanja-Kizza H.
      • Mugerwa R.
      Predictors of treatment failure among pulmonary tuberculosis patients in Mulago Hospital, Uganda.
      ,
      • Shargie E.B.
      • Lindtjørn B.
      Determinants of treatment adherence among smear-positive pulmonary tuberculosis patients in southern Ethiopia.
      ). It also highlighted the fact that the majority of hospital-diagnosed patients in the population are unable to follow up on treatment, providing a potential angle for improving TB control. TB control programs should therefore focus effort on increasing follow-up of treatment and care at hospitals, through improved record collection and management.
      We also found that the presence of positive SSM resulted in a 2.4-fold increase in the odds of an unfavorable outcome compared with having a negative sputum smear upon diagnosis. This is in contrast to the findings of Namukwaya et al., but in agreement with the those of Singla et al., and Hailu et al., who found that pretreatment bacillary load is indicative of poor treatment outcomes (
      • Hailu D.
      • Abegaz W.E.
      • Belay M.
      Childhood tuberculosis and its treatment outcomes in Addis Ababa: a 5-years retrospective study.
      ,
      • Namukwaya E.
      • Nakwagala F.N.
      • Mulekya F.
      • Mayanja-Kizza H.
      • Mugerwa R.
      Predictors of treatment failure among pulmonary tuberculosis patients in Mulago Hospital, Uganda.
      ,
      • Singla R.
      • Singla N.
      • Sarin R.
      • Arora V.K.
      • et al.
      Influence of pre-treatment bacillary load on treatment outcome of pulmonary tuberculosis patients receiving DOTS under revised national tuberculosis control programme.
      ). These findings are important because SSP cases are known to be more contagious and transmissible than SSN (
      • Jenkins H.E.
      Global burden of childhood tuberculosis.
      ,
      • Walters E.
      • Cotton M.F.
      • Rabie H.
      • Schaaf H.S.
      • Walters L.O.
      • Marais B.L.
      Clinical presentation and outcome of tuberculosis in human immunodeficiency virus infected children on anti-retroviral therapy.
      ,
      • CDC
      Core curriculum on tuberculosis: what the clinician should know.
      ). Since these cases are more likely to spread disease, proper control requires timely and quality patient care, which can be inconsistent in high-burden settings, as we have shown. The association between SSP cases and unfavorable outcomes is concerning because it suggests that difficulties in reducing transmission of pediatric TB may vary from setting to setting. The WHO’s End TB Strategy as well as the Center for Disease Control (CDC) protocols recommend immediate contact tracing and identification of networks as one of the most effective ways of stopping transmission of infectious pediatric TB, especially in developed countries like the USA (
      • World Health Organization
      Definitions and reporting framework for tuberculosis – 2013 revision – updated December 2014 and January 2020.
      ,
      • CDC
      Core curriculum on tuberculosis: what the clinician should know.
      ). However, in high-burden settings with a lower quality of patient care, breaking the chain of transmission requires a shift from identifying networks of infection to a focus on successfully treating highly contagious sources of infection, such as SSP pediatric cases of TB. This finding highlights the value of intersecting approaches between public health and medicine in addressing pediatric TB control.
      The national guidelines on pediatric TB management in Zimbabwe adopt WHO strategies, while stressing the importance of treatment completion, proper data management, and reporting to the National TB Program (
      • Zimbabwe Ministry of Health
      National guidelines on management of tuberculosis in children.
      ). The treatment success rate in Zimbabwe has been reported as 81%, which is below the desired rate of 85% (
      • Jenkins H.E.
      Global burden of childhood tuberculosis.
      ,
      • U.S. Agency for International Development
      Global health | Zimbabwe.
      ,
      • Ncube R.T.
      • Takarinda K.C.
      • Zishiri C.
      • van der Boogaard W.
      • Mlilo N.
      • Chiteve C.
      • et al.
      Age-stratified tuberculosis treatment outcomes in Zimbabwe: are we paying attention to the most vulnerable?.
      ). Previous work by Ncube et al., in Bulawayo, Zimbabwe, found that pediatric cases of TB under the age of 10 with unfavorable outcomes only made up 8% of all unfavorable outcomes. However, the authors acknowledged that their findings were in contradiction to findings from other studies conducted in the surrounding regions, which had reported much lower success rates. They attributed their higher success rate to better follow-up in their region compared with surrounding areas (
      • Ncube R.T.
      • Takarinda K.C.
      • Zishiri C.
      • van der Boogaard W.
      • Mlilo N.
      • Chiteve C.
      • et al.
      Age-stratified tuberculosis treatment outcomes in Zimbabwe: are we paying attention to the most vulnerable?.
      ). In our study, we found an overall treatment success rate of 57% based on analysis of the actual data. In comparison with the nationally reported TB treatment success rate of 81%, our study therefore identified a significantly lower favorable treatment outcome rate (
      • Ncube R.T.
      • Takarinda K.C.
      • Zishiri C.
      • van der Boogaard W.
      • Mlilo N.
      • Chiteve C.
      • et al.
      Age-stratified tuberculosis treatment outcomes in Zimbabwe: are we paying attention to the most vulnerable?.
      ). This is an important finding because it highlights that in order to reach the desired goal of 85%, more attention needs to be given to treating pediatric TB, especially at the local and community level (
      • Jenkins H.E.
      Global burden of childhood tuberculosis.
      ).
      While we acknowledge the limitations of imputing approximately two fifths (42%) of the outcome data, the comparison of actual treatment success rate versus the imputed treatment outcome success rate indicates the possibility that the treatment success rate for Harare is possibly an over-estimate. The majority of cases with a missing outcome that were imputed came from BRIDH, indicating that BRIDH and other low-resource areas have shouldered a disproportionate amount of the pediatric TB burden. Future improvements in data management and treatment compliance for pediatric TB, focusing on low-resource hospitals and clinics, is the key to the improvement of the overall success rate of pediatric TB treatment in high TB burden settings, like Harare, Zimbabwe.
      An important strength of our study is that it is the first in Harare, Zimbabwe to use a large dataset to address issues surrounding pediatric TB and its treatment outcome. A major limitation of this study is the large amount of missing data relating to treatment outcome. We found no significant differences in the distribution of patient demographic characteristics and most clinical characteristics of the patients between the study sample and excluded cases. However, statistically significant differences relating to treatment center and sputum smear positivity were observed between the study sample and the excluded cases, suggesting a potential sampling bias in the study (Table 1). If the absent mechanism for the outcome variable was missing at random (MAR), our conclusion based on the multivariate logistic regression should be unbiased. Under a missing not at random (MNAR) scenario, the estimated odds ratio between the two hospitals could have been overestimated only if favorable outcomes were more likely to be missing. Given that there was no evidence that more favorable outcomes were more likely to be missing, we believe our results to be sound. Additionally, we acknowledge that an important variable that was missing was antiretroviral therapy (ART) status. It was not feasible to include ART status due to the widespread inconsistencies in record keeping for cases that were TB/HIV coinfected. This emphasizes the importance of building stronger recordkeeping practices to include such aspects as ART status, timing of ART initiation, and interactions between ART and anti-TB drugs (
      • World Health Organization
      ,
      • Walters E.
      • Cotton M.F.
      • Rabie H.
      • Schaaf H.S.
      • Walters L.O.
      • Marais B.L.
      Clinical presentation and outcome of tuberculosis in human immunodeficiency virus infected children on anti-retroviral therapy.
      ).
      In conclusion, our study showed, for the first time, that healthcare setting and pretreatment SSM status were associated with treatment outcome for pediatric TB cases in the city of Harare. These findings suggest that confronting the imbalance in healthcare resources and quality should be an important focus of TB control, and that improved treatment follow-up for cases with positive SSM also offers a potential roadmap for use in other high-burden settings. While positive SSM status was not significantly different between those with and without a treatment outcome, it was associated with an increase in unfavorable treatment outcome through multivariate logistic regression. Through future research focusing on treatment quality and surveillance in areas with more infectious pediatric TB, it should be possible to effectively combat the decades of neglect that have contributed to the pediatric epidemic (
      • Jenkins H.E.
      Global burden of childhood tuberculosis.
      ,
      • Starke J.R.
      • Cruz A.T.
      The global nature of childhood tuberculosis.
      ).

      Funding support

      This study used existing TB surveillance data. The retrieving of the study data was supported by funds from the Department of Epidemiology and Office of Global Public Health, School of Public Health, University of Michigan .

      Declaration of interests

      The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

      Ethical approval

      This study was approved by the Institutional Review Board of the University of Michigan for the Health and Behavior Sciences and the Joint Review Ethics Committee of the University of Zimbabwe.

      Acknowledgements

      We thank the Department of Community Medicine at the University of Zimbabwe, College of Health Sciences and the Department of Epidemiology, School of Public Health at the University of Michigan for facilitating the research. Further thanks go to the matrons, data collectors, and nurses at Beatrice Road Infectious Diseases Hospital and Wilkins Infectious Diseases Hospital for their contribution to the collection of patient data.

      References

        • Adejumo O.A.
        • Daniel O.J.
        • Adebayo B.I.
        • Adejumo E.N.
        • Jaiyesimi E.O.
        • Akang G.
        • et al.
        Treatment outcomes of childhood TB in Lagos, Nigeria.
        J Trop Pediatr. 2016; 62: 131-138
        • Carlucci J.G.
        • Blevins Peratikos M.
        • Kipp A.M.
        • Lindengren M.L.
        • Du Q.T.
        • Renner L.
        • et al.
        Tuberculosis treatment outcomes among HIV/TB-coinfected children in the International Epidemiology Databases to evaluate AIDS (IeDEA) network.
        J Acquir Immune Defic Syndr. 2017; 75: 156-163https://doi.org/10.1097/QAI.0000000000001335
        • CDC
        Core curriculum on tuberculosis: what the clinician should know.
        (Available at: www.cdc.gov/tb,)2003
        • Earley M.
        • Chirenda J.
        • Highet A.
        • Mujuru H.A.
        • Yang Z.
        Characterizing pediatric tuberculosis with and without human immunodeficiency virus coinfection in Harare, Zimbabwe.
        Am J Trop Med Hyg. 2018; : 601-607https://doi.org/10.4269/ajtmh.18-0025
        • Falzon D.
        • Schünemann H.J.
        • Harausz E.
        • Gonzalez-Angulo L.
        • Lienhardt C.
        • Jaramillo E.
        • et al.
        World Health Organization treatment guidelines for drug-resistant tuberculosis, 2016 update.
        Eur Respir J. 2017; 491602308https://doi.org/10.1183/13993003.02308-2016
        • Graham S.M.
        • Sismanidis C.
        • Menzies H.J.
        • Marais B.J.
        • Detjen A.K.
        • Black R.E.
        Importance of tuberculosis control to address child survival.
        Lancet. 2014; 383: 1605-1607https://doi.org/10.1016/s0140-6736(14)60420-7
        • Hailu D.
        • Abegaz W.E.
        • Belay M.
        Childhood tuberculosis and its treatment outcomes in Addis Ababa: a 5-years retrospective study.
        BMC Pediatr. 2014; 14: 61
        • Houben R.M.G.J.
        • Menzies N.A.
        • Sumner T.
        • Huynh G.H.
        • Arinaminpathy N.
        • Goldhaber-Fiebert J.D.
        • et al.
        Feasibility of achieving the 2025 WHO global tuberculosis targets in South Africa, China, and India: a combined analysis of 11 mathematical models.
        Lancet Glob Health. 2016; 4: PE806-PE815https://doi.org/10.1016/S2214-109X(16)30199-1
        • Jenkins H.E.
        Global burden of childhood tuberculosis.
        Pneumonia. 2018; 8: 24https://doi.org/10.1186/s41479-016-0018-6
        • Little R.J.A.
        • Rubin D.B.
        Statistical analysis with missing data.
        2nd ed. John Wiley, New York2002
        • Marais B.J.
        • Schaaf H.S.
        Childhood tuberculosis: an emerging and previously neglected problem.
        Infect Dis Clin North Am. 2013; 24: 727-749https://doi.org/10.1016/j.idc.2010.04.004
        • Namukwaya E.
        • Nakwagala F.N.
        • Mulekya F.
        • Mayanja-Kizza H.
        • Mugerwa R.
        Predictors of treatment failure among pulmonary tuberculosis patients in Mulago Hospital, Uganda.
        Afr Health Sci. 2011; 11 Suppl 1: S105-S111
        • Ncube R.T.
        • Takarinda K.C.
        • Zishiri C.
        • van der Boogaard W.
        • Mlilo N.
        • Chiteve C.
        • et al.
        Age-stratified tuberculosis treatment outcomes in Zimbabwe: are we paying attention to the most vulnerable?.
        Public Health Action. 2017; 7: 212-217https://doi.org/10.5588/pha.17.0024
        • Newton S.M.
        • Brent A.J.
        • Anderson S.
        • Whittaker E.
        • Kampmann B.
        Paediatric tuberculosis.
        Lancet Infect Dis. 2008; 8: 498-510https://doi.org/10.1016/S1473-3099(08)70182-8
        • Nzombe P.
        • Satyanarayana S.
        • Tweya H.
        • Timire C.
        Declining trends in childhood tb notifications and profile of notified patients in the city of Harare, Zimbabwe, from 2009 to 2018.
        J Trop Med. 2020; https://doi.org/10.1155/2020/4761051
        • Oeltmann J.E.
        • Chengeta B.
        • Mboya J.J.
        • Wells C.D.
        • Kilmarx P.H.
        • Samandari T.
        • et al.
        Reported childhood tuberculosis treatment outcomes, Gaborone and Francistown, Botswana, 1998–2002.
        Int J Tuberc Lung Dis. 2008; 12: 186-192
        • Raghunathan
        • Lepkowski
        • Van Hoewyk
        • Solenberger
        A multivariate technique for multiply imputing missing values using a sequence of regression models.
        Surv Methodol. 2001; 27: 85-95
        • Rubin D.B.
        Multiple imputation for nonresponse in surveys.
        John Wiley & Sons, New York1987
        • Shargie E.B.
        • Lindtjørn B.
        Determinants of treatment adherence among smear-positive pulmonary tuberculosis patients in southern Ethiopia.
        PLoS Med. 2007; 4: e37https://doi.org/10.1371/journal.pmed
        • Singla R.
        • Singla N.
        • Sarin R.
        • Arora V.K.
        • et al.
        Influence of pre-treatment bacillary load on treatment outcome of pulmonary tuberculosis patients receiving DOTS under revised national tuberculosis control programme.
        Indian J Chest Dis Allied Sci. 2005; 47: 19-23
        • Starke J.R.
        • Cruz A.T.
        The global nature of childhood tuberculosis.
        Pediatrics. 2014; 133: e725-e727https://doi.org/10.1542/peds.2013-4139
        • U.S. Agency for International Development
        Global health | Zimbabwe.
        (Available at: https://www.usaid.gov/zimbabwe/global-health,)2018
        • UN General Assembly
        Transforming our world: the 2030 agenda for sustainable development.
        (A/RES/70/1. Available at:)2015
        • van Zyl S.
        • Marais B.J.
        • Hesseling A.C.
        • Gie R.P.
        • Beyers N.
        • Schaaf H.S.
        Adherence to anti-tuberculosis chemoprophylaxis and treatment in children.
        Int J Tuberc Lung Dis. 2006; 10: 13-18
        • Walters E.
        • Cotton M.F.
        • Rabie H.
        • Schaaf H.S.
        • Walters L.O.
        • Marais B.L.
        Clinical presentation and outcome of tuberculosis in human immunodeficiency virus infected children on anti-retroviral therapy.
        BMC Pediatr. 2008; 8: 1https://doi.org/10.1186/1471-2431-8-1
        • World Health Organization
        Roadmap for childhood tuberculosis: towards zero deaths. 44. 2013 (Available at: http://apps.who.int/iris/bitstream/handle/10665/89506/9789241506137_eng.pdf?sequence=1,)
        • World Health Organization
        Global strategy and targets for tuberculosis prevention, care and control after 2015 report by the secretariat.
        (Available at: https://apps.who.int/iris/handle/10665/172828,)2013
        • World Health Organization
        Global tuberculosis report 2017.
        (http://apps.who.int/iris/bitstream/handle/10665/259366/9789241565516eng.pdf;jsessionid=3E90801A51D3878CFCE2FAC5F839096?sequence=1Available at:,)2017
        • World Health Organization
        Global tuberculosis report 2019.
        (Available at: https://www.who.int/tb/publications/global_report/en/,)2019
        • World Health Organization
        Definitions and reporting framework for tuberculosis – 2013 revision – updated December 2014 and January 2020.
        (Available at: https://www.who.int/tb/publications/definitions/en/,)2020
        • Zimbabwe Ministry of Health
        National guidelines on management of tuberculosis in children.
        (Retrieved from: http://apps.who.int/medicinedocs/documents/s21974en/s21974en.pdf,)2013
        • Zimbabwe Ministry of Health and Child Care
        National tuberculosis control program guidelines.
        2015