https://orcid.org/0000-0002-1329-3864
Figures
Abstract
Background
This study investigated the associations between socio-economic deprivation and tuberculosis (TB) treatment outcomes, alongside well-known TB risk factors. The effects of healthcare expenditures and their growth on trends in TB incidence from 2009 to 2013 were also assessed.
Methods
Secondary data analysis was performed on data obtained from various sources including governmental, non-governmental and research institutions. Indicators for TB treatment outcomes included TB death rate, TB rate among the household contacts of the Index TB cases, TB treatment failure, HIV associated TB death rate, TB defaulter rate, and new TB smear positive cases. Analysis of variance (ANOVA) and Turkey’s tests for post-hoc analysis were used to compare means of variables of interest considering a type I error rate of 0.05. Regression models and canonical discriminant analysis (CDA) were used to explore the associations between trends in TB incidence and independent TB predictors. During CDA, Fischer’s linear functions, Eigen values, and Mahalanobis distances were determined with values of Wilk’s Lambda closer to zero being the evidence for well discriminated patient groups. Data analysis was performed using SPSS® statistical software version 23.0 (Chicago, IL).
Results
In total, 62 400 records of TB notification were analyzed for the period 2009–2013. The average TB incidence rate over a 5-year period was 298 cases per 100,000 inhabitants per year. The incidence of TB was reduced by 79.70% at the end of the evaluation as compared to the baseline data in 2009. Multiple linear regression analysis showed that the Expenditure per patient day equivalent (PDE) and PHC expenditure per capita were significantly and independently associated with the decline of TB incidence (adjusted R2 = 60%; ρ = 0.002) following the equation: Y = (- 209× Expenditure per PDE) + (- 0.191 × PHC expenditure per capita). CDA showed that in the most socio-economically deprived communities (quintile 1), HIV associated TB death rates were significantly more likely to be higher as compared to the least socio-economically deprived group (quintile 5) [Eigen value (12.95), function coefficient (1.49) > (.77); Wilk’s Lambda = .019, p < .0001].
Conclusions
Although TB control programs in OR Tambo district have averted thousands of TB incident cases, their effects on HIV associated TB deaths among the most deprived communities remain insignificant. There is an urgent need for strengthening integration of TB/HIV services in most deprived settings.
Citation: Dlatu N, Longo-Mbenza B, Apalata T (2022) Predictors of tuberculosis incidence and the effects of multiple deprivation indices on tuberculosis management in OR Tambo district over a 5-year period. PLoS ONE 17(3): e0264811. https://doi.org/10.1371/journal.pone.0264811
Editor: Frederick Quinn, The University of Georgia, UNITED STATES
Received: May 20, 2020; Accepted: February 17, 2022; Published: March 10, 2022
Copyright: © 2022 Dlatu et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The datasets used and/or analysed during the current study are available at Figshare (DOI: 10.6084/m9.figshare.14055296).
Funding: No funding was obtained.
Competing interests: The authors have declared that no competing interests exist.
Abbreviations: AFB, acid-fast bacilli; ANC, African National Congress; APP, Annual performance plan; ANOVA, Analysis of variance; CDA, canonical discriminant analysis; DC, District code; ECSECC, Eastern Cape Socio-economic Consultative Council; GDP, gross domestic product; HIV, human immunodeficiency virus; HST, Health System Trust; HSRC, Human Sciences Research Council; MDR-TB, multidrug-resistant tuberculosis; NDoH, National Department of Health; PDE, patient day equivalent; PHC, primary healthcare; SA, South Africa; SAIMD, South African Index of Multiple Deprivation; SEQ, Socio-economic quintile; SPSS, Statistical Package for the Social Sciences; TB, Tuberculosis; WHO, World Health Organization; XDR-TB, extensively-drug resistant tuberculosis
Background
According to the 2020 Global Tuberculosis (TB) report, almost 10 million new cases of TB disease were reported worldwide, and this was equivalent to 130 cases per 100,000 populations in 2019 [1]. Ranking above HIV/AIDS as the leading cause of death due to a single infectious agent, TB is confirmed to be among the top 10 killers globally [1, 2]. South Africa remains among the 30 high TB burden countries accounting for about 86% of all new cases of TB disease in the world [1]. The TB incidence in South Africa is estimated to 615 cases per 100,000 populations, ranging from 427 to 835 cases per 100,000 populations in 2019 [1]. This is despite the fact that TB incidence in South Africa has been drastically decreasing since 2009 [3]. Among the key driving factors, we have high rate of HIV co-infection, poverty and the emergence of multidrug-resistant TB (MDR-TB) and extensively-drug resistant TB (XDR-TB) [1, 3].
Health expenditure, expressed as total percentage (%) of the gross domestic product (GDP) in South Africa was reported at 8.797% in 2014, according to the World Bank [4]. However, its impact on trends of TB incidences in the country is not well documented.
Studies have shown that presently, KwaZulu-Natal, Eastern Cape, and Western Cape have the highest TB incidence rates in South Africa [5, 6]. Eastern Cape remains one of the poorest South African provinces characterized by disconnected and inadequate healthcare services. In recent years, attention has turned to the contributions made by the mismanagement of healthcare expenditures and poor service delivery by local governments (according to unverified reports) and the index of multiple deprivation in fueling both TB and HIV epidemics [5, 6]. Unconfirmed evidence also suggests that in the OR Tambo district of the Eastern Cape SA province, the expansion of TB services for people living with HIV has been slower particularly among communities living under extreme poverty. Although the World Health Organization (WHO) and the South African National Department of Health (NDoH) have issued strategic guidelines for the integration of TB and HIV services, yet these guidelines remain poorly implemented, especially among the most socio-economically deprived areas [7, 8]. Hence, the present study sought to investigate the associations between socio-economic deprivation and TB treatment outcomes, alongside other TB risk factors such as population density, number of people living in poverty, poverty gap, supervision rate in primary healthcare (PHC) facilities, and PHC work load. In addition, the effects of healthcare expenditures (Expenditure per capita, Expenditure per patient day equivalent–PDE, and the local government expenditure) and their growth over the years on trends of TB incidences over a 5-year period were also assessed.
Methods
Description of the study settings
O.R Tambo district (see Map 1) is one of the 7 districts of the Eastern Cape province of South Africa. The seat of O.R Tambo is in Mthatha. The vast majority of its population of about 1 676 463 speak isi-Xhosa [9].
Source: http://isrdp.dplg.gov.za/documents/IDP/ISRDP/OR_Tambo_IDP.pdf. (open access).
The district is made up of 4 health sub-districts: King Sabata Dalindyebo sub-district with a population of 442318, Mhlontlo sub-district and its population is 221827, Nyandeni sub-district and its population is 436813, and Qaukeni sub-district with a total population of 659431 [9]. OR Tambo district has been reported to bear the following basic indicators: 64.6% of people are living under poverty with an estimated unemployment rate of 65.5% and the literacy rate of 42.2%. The average annual income of a Black resident is R15,762. Health services are delivered by one central hospital, 1 regional hospital, 12 district hospitals; 11 community health centers, 49 clinics, 52 health posts and 15 mobile health services [9].
Study design, data sources and description of variables of interest
Secondary data review was performed on the OR Tambo TB data obtained between 2009 and 2013 from various sources including governmental, non-governmental and research institutions. Those included data from the 2009–2014 District Health Barometer by the Health System Trust (HST); the 2017 Global Monitoring Report by the WHO and the World Bank; the 2015 WHO End TB Strategy; the 2014–2015 Eastern Cape Department of Health Annual Reports and TB health records from healthcare facilities in the OR Tambo district; the 2015 Re-engineering of Primary Healthcare Strategy by the EC DoH; the 2017 Eastern Cape Socio-economic Consultative Council (ECSECC); the 2012 South African National HIV Prevalence, Incidence and Behaviour Survey by the Human Sciences Research Council (HSRC); the 2015 Guide to Monitoring and Evaluation for collaborative TB/HIV Activities by the WHO and the Global Funds; the 2015–2016 Adequacy and Efficiency in PHC Financing by the SA National Treasury and the National Treasury Report 2015/2016; and the 2017/2018 to 2019/2020 Annual Performance Plan (APP) by the SA NDoH.
The following variables of interest were extracted from the above data sources:
Data analysis and statistical methods
Continuous indicators of TB management were presented as mean ± standard deviation (SD) for bivariate analysis. Pearson’s correlation coefficient was used to measure the degree of linear association between variables in general and between TB indicators and deprivation-concentration in particular, ranging in magnitude on interval scale from -1 to +1. Analysis of variance (ANOVA) and Turkey’s post-hoc tests were used to compare means of variables of interest considering a type I error rate of 0.05. Regression models were used to explore the associations between trends in TB incidence and independent TB predictors. The regression coefficient represented the increase or decrease in the absolute magnitude of the independent variable for each unit of increase in the dependent variables using the equation: Y = constant (slope) ± ax. The slope index of inequality represented the linear regression coefficient that showed the relationship between the level of implementation, the burden of TB and deprivation. Canonical discriminant analysis (CDA) was used to identify the variable of interest that significantly discriminated different levels of socio-economic deprivation from 2009 to 2013. The major underlying assumptions of CDA were: (i) each predictor variable is normally distributed using histograms and P-P plots; (ii) there must be homogeneity of covariance across the groups–the Box’s M test of equality of covariance matrices was used to check the assumption of homogeneity of covariance across the groups using p < .001 as a criterion; (iii) there must be at least two groups or categories, with each study case belonging to only one group so that the groups are mutually exclusive and collectively exhaustive; (iv) the groups or categories should be defined before collecting the data; (v) the predictor variable(s) used to separate the groups should discriminate quite clearly between the groups so that group or category overlap is clearly non-existent or minimal; (vi) group sizes of the dependents should not be grossly different and should be at least five times the number of independent variables.
In order to check for the assumptions (iii), (iv), (v) and (vi), we used Mahalanobis distances in order to support the classification of canonical variates into distinct groups and for comparing divergence among populations’ group centroids so that we can determine the degree of segregation with values of Wilk’s Lambda closer to zero being the evidence for well-separated groups. Each variable that is more than 1.96 Mahalanobis distance units away from a specific centroid group has less than 5% chance of belonging to that group. In addition, during CDA, Fischer’s linear functions and Eigen values were determined considering values of Wilk’s Lambda. Data analysis was performed using SPSS® statistical software version 23.0 (SPSS Inc; Chicago, IL).
Ethics approval and consent to participate
The Research Ethics and Biosafety Committee of the Walter Sisulu University approved the study (ethical clearance No. 29/2014) and permission to conduct the study was obtained from the Eastern Cape Department of Health. This was a secondary data analysis, hence no consents from patients were obtained.
Results
In total, 62 400 records for TB notification were analyzed for the period 2009–2013. The following findings were obtained:
1. Predictors of TB incidence in O.R. Tambo district
Table 1 and Fig 1 display trends of TB incidence per 100,000 populations over a 5-year period as well as the actual TB cases diagnosed annually from 2009 to 2013. The incidence curve of TB trends dropped abruptly and significantly from 2009 until 2013 (ρ < 0.00001) (Table 1 and Fig 1) as a result of a performance efficiency demonstrated by the reduction of 79.70% of TB incidence at the end of 2013 as shown in Fig 2.
Data on population density and related poverty indicators as well as indicators for healthcare expenditures and their growth over a 5-year period are displayed in Tables 2 and 3, respectively. Correlation analyses between TB incidence rates and potential TB predicting factors in Tables 2 and 3 showed a bivariate positive correlation between population density (r = 0.812; ρ < 0.0001), poverty gap index (r = 0.210; ρ < 0.01), and TB incidence (new cases of TB per 100,000 populations). However, there was a significant negative correlation between supervision rate (r = - 0.173; ρ = 0.030), PHC professional nurse clinical workload (r = -0.164; ρ = 0.021), Expenditure per patient day equivalent (r = -0.282; ρ = 0.015), PHC Expenditure per capita (r = -0.159; ρ = 0.034), local government expenditure on PHC (r = -0.244; ρ = 0.022), and TB incidence.
However, using a multiple linear regression analysis, only Expenditure per patient day equivalent (PDE) and PHC Expenditure per capita were identified as the most important, significant and independent predictors (adjusted R2 = 60%; ρ = 0.002) in terms of declining TB incidence following the equation: Y = (- 209× Expenditure per PDE) + (- 0.191 × PHC expenditure per capita).
2. Association between socio-economic deprivation and TB care indicators in O.R. Tambo district
Table 4 shows bivariate associations between TB care indicators and levels of socio-economic deprivation over a 5-year period. Following Turkey’s tests for post-hoc analysis, significantly higher TB death rates were observed in the most socio-economically deprived group (Quintile 1) as compared to Quintile 3, Quintile 4, and the least socio-economically deprived group (Quintile 5) (P < 0.0001) as shown in Fig 3A. Although a higher HIV-associated TB death rate was also observed among the most socio-economically deprived population (Quintile 1) than other socio-economically deprived groups, this difference was significant only when comparing Quintile 1 and Quintile 5 (P = 0.02) (Fig 3B) during a post-hoc multiple comparisons. TB incidence rates among the household contacts of the index TB cases were significantly more likely to be higher in the most socio-economically deprived group (Quintile 1) as compared to Quintile 3 (P = 0.007), Quintile 4 (P = 0.005), and the least socio-economically deprived group (Quintile 5) (P < 0.0001) (Fig 3C).
A. Correlation between TB death rate and deprivation index (P < 0.0001). B. Correlation between HIV associated TB death rate and deprivation index (P = 0.049). C. Correlation between TB rate among household contacts of index TB cases and deprivation index (P < 0.0001). E. Correlation between TB treatment failure rate and deprivation index (P < 0.0001). D. Correlation between TB defaulter rates and deprivation index (P < 0.0001). F. Correlation between new TB smear positive cases and deprivation index (P = 0.001).
However, significantly higher TB defaulter rates were observed among individuals in Quintile 3 (P = 0.015), Quintile 4 (P<0.0001) and the least socio-economically deprived group (Quintile 5) (P<0.0001) as compared with the most socio-economically deprived group (Quintile 1) (Fig 3D). Similarly, higher TB treatment failure rates were significantly observed among the least deprived groups, particularly Quintile 4 and Quintile 5 groups (P<0.0001) as compared with the most socio-economically deprived group (Quintile 1) (Fig 3E). Finally, there was a significantly high number of new TB smear positive cases observed in the least socio-economically deprived group (Quintile 5) as compared to Quintile 3 (P = 0.001) and Quintile 1 groups (P = 0.002) as shown in Fig 3F.
Using canonical discriminant analyses (CDA) on normally distributed variables that showed significant univariate associations, three canonical functions were identified as able to discriminate among the socio-economically deprived groups. The obtained 3 discriminant functions had Eigen values that are summarized in Table 5. The larger the Eigen value represents more shared variance in the linear combination of variables.
Of the 3 identified canonical functions, the first two functions contributed a total of 98.5% of the total variance, which easily stratifies into the 5 study groups (from quintile 1 to quintile 5) (Box’s M (3.77), p (0.47) > α (0.001)). The Wilks’ Lambda value was significant: 0.019; χ2 = 137.62; ρ < 0.0001 (Table 6).
Functions’ coefficients were calculated and used to decide which variables predicted group membership. Comparing the values between groups, the higher coefficient means the predictor variable attributes more for that group (Table 7).
Table 8 demonstrated Fisher’s linear discriminant functions according to classification function coefficients. After adjusting for confounding factors (TB rate among the household contacts of the Index TB cases, TB defaulter rate, new smear positive TB cases, TB death rate, and TB treatment failure), only HIV-associated TB death rate was identified as the most important, significant, and independent indicator able to discriminate the most deprived communities with a deprivation index far from other deprivation-concentration-dispersion groups, being positioned at 2.985 Mahalanobis distance units away from quintiles 2–5 (p <0.0001) (Table 8 and Fig 4).
Standardized canonical discriminant function coefficients were used for predicting functions defining group centroids (Fig 4). The canonical group means (also called group centroids) are the mean for each group’s canonical observation scores. The larger the difference between the canonical groups implies the better the predictive power of the canonical discriminant function in classifying observations.
In conclusion, CDA showed that TB incidences in the most socio-economically deprived communities (quintile 1) were significantly more likely to have high HIV–associated TB death rates as compared to the least socio-economically deprived group (quintile 5) [Eigen value (12.95), function coefficient (1.49) > (.77); Wilk’s Lambda = .019, p < .0001].
Discussion
There are two main findings of this study. First, data have shown substantial decrease of TB incidence from 2009 to 2013 due to increased healthcare expenditures, particularly expenditure per patient day equivalent (PDE) and PHC expenditure in OR Tambo district. Whilst PHC expenditure measures the total amount of money spent annually by each district divided by the total population in the district, expenditure PDE is a composite process indicator that connects financial data with service-related data from the healthcare facilities. This indicator measures how resources available to the healthcare facility are being spent, making expenditure PDE the most relevant marker of healthcare efficiency. Therefore, our findings suggest that coordinated efforts in TB control and prevention in the OR Tambo district yielded a remarkable number of new TB cases being averted.
Previous studies have shown that when healthcare expenditures are efficiently managed, TB cases will significantly decrease because health care providers will be trained to improve the ability to diagnose and treat persons with TB disease; there will be improvements in laboratory diagnostic methods with early testing and management of TB patients; implementation of appropriate infection prevention and control precautions in health care facilities and other congregate settings in order to reduce TB transmission rates; as well as strengthening of local health facilities and TB control programs in order to undertake monitoring and evaluation of TB patients on therapy [15–17].
Second, whilst milestones in reduction in TB cases are significantly progressing towards their achievements, findings from this study have unfortunately revealed that HIV associated TB deaths particularly among the most socio-economically deprived communities in OR Tambo district remain high, hence a cause for concern. The present study demonstrated that highest level of deprivation, concentration, and dispersion index was associated with the highest numbers of death among TB patients who are co-infected with HIV. Similar findings have been previously reported elsewhere [17–20].
In explaining this disparity, it is possible that either health resources are not evenly distributed between the most and the least socio-economically deprived communities to combat both TB and HIV diseases or TB and HIV health services are not fully integrated in settings where individuals are the most deprived. Either way, it is clear that TB disproportionately affects the socially and economically marginalized, with a recognized role for HIV co-infection in addition to traditional factors such as poverty and overcrowding, characterized by high population density [12, 18–21]. Since the South African provincial and local government health expenditure per headcount does provide insight into equity in justice and resources distribution [7, 10, 11], there is therefore an urgent need for addressing TB and HIV health services integration in most deprived geographical areas in the OR Tambo district in order to curb HIV associated TB deaths. Previous studies have shown that strengthening of HIV and TB services integration plays a positive role in reducing both TB/HIV incidence and mortality among individuals living under conditions of highest deprivation-concentration-dispersion [7, 16, 17, 21].
South Africa has more people living with HIV and the vast majority are on ART [9]. Despite the fact that the study by Williams demonstrated that the successful roll-out of ART has been associated with a 72% reduction in the incidence of HIV among adults from 1996 to 2016 and a 74% reduction in AIDS-related mortality from 2006 to 2016 [22], integrating TB and HIV health services is thought to actually play a major role, and might substantially reduce TB/HIV mortality in South Africa by 2030 irrespective of other traditional risk factors. Unfortunately, high rates of TB and HIV co-infection, poverty, inadequate training and supervision of the health professionals, absence of staff and material resources, and lack of infection prevention and control policies might be barriers to the implementation of HIV/TB integration, particularly among the most socio-economically deprived communities [23] such as in the O.R Tambo District as demonstrated by this present study.
Conclusions and recommendations
The study found out that the health burden due to tuberculosis is closely correlated to expenditure per patient day equivalent, a strong indicator of healthcare efficiency. On the other hand, HIV associated TB deaths particularly among the most socio-economically deprived communities in OR Tambo district remains high. This health outcome provides a more encompassing view on the devastating effect of TB and HIV co-infection among the poorest communities. These findings highlight the importance of expanding coverage of TB and HIV integrated services to ensure prompt diagnosis and initiation of treatment among the most deprived people in resource-limited settings. Public and private partnership should be strengthening in order to develop comprehensive, collaborative and holistic approaches toward TB elimination by bringing together key stakeholders among policy makers, researchers, health professionals, civil society, non-governmental organizations (NGOs) and private sector. It is evident that that combined and coordinated efforts are the only way to halt the dual epidemics of TB and HIV by integrating healthcare services. However, further studies are warranted to address barriers and opportunities for integrating TB and HIV services among the most socio-economically deprived communities of OR Tambo district.
Acknowledgments
We thank the staff and colleagues in the Department of Public Health for their support during the collection of data used in this study.
References
- 1.
World Health Organisation. The Global tuberculosis report 2020. Geneva: World Health Organization; 2020. Licence: CC BY-NC-SA 3.0 IGO. Available: https://apps.who.int/iris/bitstream/handle/10665/336069/9789240013131-eng.pdf [31 January 2021].
- 2. Loveday M, Mzobe YN, Pillay Y, Barron P. Figures of the dead: A decade of tuberculosis mortality registrations in South Africa, SAMJ 2019; 109(10): 728–732. pmid:31635566
- 3. Manda SO, Masenyetse LJ, Lancaster JL, van der Walt ML. Risk of Death among HIV co-infected Multidrug Resistant Tuberculosis Patients, Compared to Mortality in the General Population of South Africa. J AIDS Clin Res. 2013 Jul 2; Suppl 3:7. pmid:24455448
- 4.
World Bank. Overcoming the Legacy of Exclusion in South Africa. World Bank 2018; Report No: 125838-ZA.
- 5.
World Health Organisation. The Global Tuberculosis Control. WHO, Geneva, 2018. Accessed at: www.who.int/tb/publications/global_report/en/ [20 December 2019].
- 6. Duarte R, Lönnroth K, Carvalho C, Lima F, Carvalho ACC, Muñoz-Torrico M, et al. Tuberculoses, social determinants and co-morbidities (including HIV). Pulmonology. 2018 Mar; 24(2): 115–9. pmid:29275968
- 7. Naidoo K, Gengiah S, Yende-Zuma N, Padayatchi N, Barker P, Nuun A, et al. Addressing challenges in scaling up TB and HIV treatment integration in rural primary healthcare clinics in South Africa (SUTHI): a cluster randomized controlled trial protocol. Implementation Sci 12, 129 (2017) pmid:29132380
- 8.
SA National Department of Health. South Africa’s National Strategic Plan for HIV, TB and STIs 2017–2022. SA NDoH 2017. Accessed at: https://www.gov.za/sites/default/files/gcis_document/201705/nsp-hiv-tb-stia.pdf [20 December 2019].
- 9.
Statistics South Africa. Census 2011 Statistical release–P0301.4. Statistical SA 2012.
- 10.
Noble M, Zembe W, Wright G, Avenell D. Multiple Deprivation and Income Poverty at Small Area Level in South Africa in 2011. Cape Town: SASPRI 2013. Available at http://www.saspri.org/wp‐content/uploads/SAIMD%202011%20at%20ward%20level%20National%20Report%20for%20Web.pdf [20 December 2019].
- 11. Massyn N, Day C, Peer N, Padarath A, Barron P, English R, editors. District Health Barometer 2013/14. Durban: Health Systems Trust; October 2014.
- 12.
World Health Organization. Intensified tuberculosis case-finding and isoniazid preventive therapy for people living with HIV in resource-constrained settings. WHO Guidelines 2011, ISBN: 978924150070 8.
- 13. Scott VE, Sanders D. Evaluation of how integrated HIV and TB programs are implemented in South Africa and the implications for rural-urban equity. Rural and Remote Health Journal 2013; Volume 13, Issue 2, p.2165. pmid:23713881
- 14. Floyd K, Glaziou P, Zumla A, Raviglione M. The global tuberculosis epidemic and progress in care, prevention, and research: an overview in year 3 of the End TB era. Lancet Respir Med. 2018 Apr; 6 (4):299–314. pmid:29595511
- 15. Uyei J, Coetzee D, Macinko J, Weinberg SL, Guttmacher S. Measuring the degree of integrated tuberculosis and HIV service delivery in Cape Town, South Africa. Health Policy Plan. 2014 Jan;29(1):42–55. pmid:23274439
- 16. Dlwati LV, Mavundla TR, Mbengo F. Facilitators for and Barriers to the Implementation of National Tuberculosis Management Guidelines. Africa Journal of Nursing and Midwifery 2018; 19(3):13
- 17. Obaromi D, Ndege J, Yongsong Q. Disease mapping of tuberculosis prevalence in Eastern Cape Province, South Africa. Journal of Public Health volume 2018; 27, pp: 241–248.
- 18. Manosuthi W, Wiboonchutikul S, Sungkanuparph S. Integrated therapy for HIV and tuberculosis. AIDS Research and Therapy volume 2016; 13 (22). s pmid:27182275
- 19. Secretary of Sahyoga JS, Lauxa TS, Patila S. Predictors of tuberculosis treatment outcomes among a retrospective cohort in rural, Central India. J Clin Tuberc Other Mycobact Dis 2018; 12: 41–47. pmid:31720398
- 20. Chingonzoh R, Manesen MR, Madlavu MJ, Sopiseka N, Nokwe M, Emwerem M, et al. Risk factors for mortality among adults registered on the routine drug resistant tuberculosis reporting database in the Eastern Cape Province, South Africa, 2011 to 2013. PLoS One 2018; 13(8): e0202469. pmid:30133504
- 21. Castro KG, Marks SM, Chen MP, Hill AN, Becerra JE, Miramontes R, et al. Estimating tuberculosis cases and their economic costs averted in the United States over the past two decades. Int J Tuberc Lung Dis. 2016 Jul; 20(7): 926–933. pmid:27287646
- 22. Williams GB, Gupta S, Wollmers M, Granich R. Progress and prospects for the control of HIV and tuberculosis in South Africa: a dynamical modelling study. The Lancet Public Health 2017; 2 (5): e223–e230. pmid:29253488
- 23. Uwimana J, Jackson D, Hausler H, Zarowsky C. Health system barriers to implementation of collaborative TB and HIV activities including prevention of mother to child transmission in South Africa. Tropical Medicine and International Health 2012; 17(5): 658–66. pmid:22394016