A research report submitted to the Faculty of Health Sciences, University of the
Witwatersrand, Johannesburg, in partial fulfilment of the requirements for the degree of MSc Epidemiology (Infectious Disease Epidemiology), Johannesburg, November 2017 / Tuberculosis (TB) is a leading cause of death worldwide, causing more deaths than HIV/AIDS.
A TB patient can have pulmonary or extrapulmonary TB or both. South Africa has a high
incidence rate of TB, recording 834 cases per 100 000 population in 2015, compared to 142
per 100 000 globally. Loss to follow-up (LTFU) rates during TB treatment in South Africa
have ranged from 7% to 30%.
The factors associated with LTFU can be divided into four groups: socioeconomic factors,
patient-related factors, treatment factors, and health system or programmatic factors.
Socioeconomic factors include a lack of support and a low socioeconomic status. Patient
related factors include substance abuse, beliefs and low TB knowledge, while treatment factors
include side effects and a history of LTFU. Among health system or programmatic factors that
contribute to LTFU are a poor relationship with the healthcare workers and large treatment
programmes.
Studies to determine the factors associated with LTFU in HIV-uninfected TB patients are few
as most studies have focused on HIV/TB co-infected patients. Co-infected patients make up
almost 60% of TB patients. The aim of this study was to determine the demographic and
clinical factors associated with LTFU in HIV-uninfected TB patients who registered for TB
treatment in Ekurhuleni North sub-district from 1st January 2011 to 30th June 2012. LTFU was
defined as a lack of a documented treatment outcome among TB patients who should have
completed TB treatment based on TB treatment start date.
The study was a retrospective cohort study involving the secondary analysis of routine TB
treatment data collected from 18 primary care clinics in Ekurhuleni North sub-district. The
participants were described at the beginning of TB treatment using clinical and demographic
data. The treatment duration and outcomes were also described. The burden of LTFU was
determined. Univariate and multivariate logistic regression and Cox proportional hazards
regression were used to determine the factors associated with LTFU. In addition, survival
analysis was conducted to determine if there was a difference in the time to LTFU among HIV
uninfected TB patients based on clinical and demographic factors. Sensitivity analysis of the
multivariate logistic regression and Cox proportional hazards regression was carried out to
compare the results obtained when follow-up was restricted to 8 months to those obtained for
12 months of follow-up. Sensitivity analysis was also conducted around the definition of
LTFU. The impact on the results of multivariate logistic regression after assuming that
participants who had a missing treatment outcome in the primary study were not lost to follow
up was determined.
Five hundred and fifteen participants were included in the analysis. The median age of the
participants was 33 years (IQR: 26-47). Fifty-eight percent of the participants were male.
Pulmonary TB was the most common form of TB among the participants. The rate of treatment
success was 77.67% and that of LTFU was 17.28%. Of those lost to follow-up, 60 had a missing
treatment outcome and 29 had default as an outcome in the primary study. The median length
of treatment was 6.39 months (IQR: 5.67-7.44), and the median time to LTFU was 3.67 months
(IQR: 1.54-6.33). Eighty-two percent of the participants had a documented change of treatment
phase. Clinics with a high patient burden had a similar proportion of poor outcomes (death,
LTFU and treatment failure) to clinics with low patient burdens. Significant differences in
change of treatment phase and length of treatment were observed between those lost to follow
up and those not lost to follow-up.
LTFU took place throughout TB treatment, with a steady increase in the probability of LTFU
over the first 6 months of follow-up. None of the factors investigated had a significant effect
on time to LTFU. Following logistic regression and Cox proportional hazards regression
analyses, none of the factors assessed were significantly associated with LTFU. Sensitivity
analysis showed that censoring the participants at 8 months did not change the results of the
logistic regression analysis. For Cox proportional hazards regression, female participants had
a 5% lower risk of LTFU compared to male participants in the 12-month analysis. In the 8
month analysis, female participants had a 5% higher risk of LTFU. When participants with a
missing treatment outcome were not considered lost to follow-up, sex was found to be
significantly associated with LTFU. Female participants had a 66% lower risk of LTFU
compared to male participants.
A limitation of the use of secondary data in this study was that the study question asked in this
study was different from the question that was asked in the primary study. As a result, the
variables collected in the primary study were different from the variables required in this study.
Information on socioeconomic status, residence type, comorbidities, treatment clinics and
health system factors was not available.
None of the factors investigated in this study were significantly associated with LTFU in HIV
uninfected TB patients in Ekurhuleni North sub-district. The factors influencing LTFU in
Ekurhuleni North may not have been investigated in this study. More studies need to be
conducted with a wide range of variables in Ekurhuleni North to determine the factors that
influence LTFU among HIV-uninfected TB patients. / XL2018
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:wits/oai:wiredspace.wits.ac.za:10539/24667 |
| Date | January 2017 |
| Creators | Moyo, Batanai |
| Source Sets | South African National ETD Portal |
| Language | English |
| Detected Language | English |
| Type | Thesis |
| Format | Online resource (65 leaves), application/pdf |
Page generated in 0.0245 seconds