TB Notes 1, 2005
No. 1, 2005
Factors Associated with Default from Multidrug-Resistant
Tuberculosis Treatment, South Africa, 1999-2001
In the past decade, South Africa has experienced a rapid escalation
of TB incidence, with new cases rising from less than 100 per 100,000
persons in 1990 to 558 per 100,000 in 2003. In the 2002 Report on
Global TB Control, the World Health Organization (WHO) ranked South
Africa third in the world in terms of reported TB incidence for
1999 and seventh in terms of absolute number of persons with active
TB.1 Health services are currently burdened by more than
200,000 new TB patients per year, of whom 60% are estimated to be
infected with HIV.2
At present, the estimated proportion of TB cases in South Africa
identified as multidrug-resistant TB (MDR TB) ranges from 1% to
2% among new cases and 4% to 14% among retreatment cases, depending
on the province. Treatment default has been identified as a significant
problem in these patients. Failure to adhere to MDR TB therapy may
result in primary transmission of MDR TB in the community, amplify
resistance to second-line drugs, and increase the chance of failure
in subsequent treatment attempts.3
CDC technical assistance was requested to assist in conducting
an evaluation of factors associated with treatment default among
patients receiving treatment for MDR TB. This case-control investigation
to identify patient-level and provider-level factors for default
to MDR TB treatment was conducted in 2003 and 2004. These results
will assist the South Africa National TB Control Program in strengthening
existing strategies for improving adherence among all TB patients
and introduce new strategies to maintain adherence among the growing
pool of MDR TB patients being treated under the national standardized
MDR TB treatment guidelines.
We conducted an unmatched, questionnaire-based, case-control study
among adult persons aged 18 years and older diagnosed with MDR TB
between October 1, 1999, and September 30, 2001, and for whom MDR
TB treatment was initiated. Data was collected from sites in five
provinces. We selected all patients from MDR TB treatment register
lists generated by each province’s MDR TB referral hospital. We
defined cases as MDR TB patients 18 years or older who initiated
standardized MDR TB therapy but subsequently defaulted. We defined
controls as MDR TB patients who initiated MDR TB treatment and were
considered to have completed a full treatment course (either cure
or completion or failure). Owing to logistical limitations, we restricted
our interviewers to look for patients living within 200 kilometers
of the hospital.
A questionnaire was developed in English and translated into the
other common South African languages (Xhosa, Zulu, Tswana, and Afrikaans)
and back-translated into English to evaluate the quality of translation.
This study was conducted using a face-to-face, administered questionnaire.
The questionnaire consisted of a mixture of open-ended, multiple
choice, and yes/no questions. Questions were asked to establish
demographic and social characteristics, health service experience,
clinical characteristics, TB/MDR TB knowledge, and self-reported
reasons for defaulting (if applicable).
Our study provided multiple layers of information about patient,
provider, and health system-level factors that may influence patients’
adherence to MDR TB treatment. First, we discovered that 13% of
treatment defaulters (our cases) and 10% of treatment completers
were improperly classified in the MDR TB registers we examined.
The proper recording of treatment outcome is of enormous importance
in a country like South Africa, with an emerging problem of MDR
TB in the midst of a large human immunodeficiency virus (HIV) epidemic.
Misclassification bias, even on the order of 10% seen here, can
markedly skew “good outcomes” (treatment success of 70%-75%) to
“modest” outcomes (treatment success of 60%-65%).
Second, not surprisingly, our investigators uncovered a significant
death rate among defaulters and those completing MDR TB treatment
in our sampling cohort. More than one in four patients who were
classified as defaulters were in fact found to have died after the
last contact with their MDR TB clinic. Twenty percent of those for
whom we found death dates had died within 2 months after stopping
treatment, and were thus “true deaths.”
We were also unable to contact another one third of treatment defaulters,
likely due to internal migration or death. Therefore the “true death
rate” among those undergoing treatment for MDR TB may not be known,
but it is likely higher than the 25% reported in recent MDR TB outcome
studies in South Africa.
We identified four major areas associated with treatment default
in the patients we interviewed. The most significant factors we
found associated with treatment default were related to the quality
of the patient-provider relationship. Our multivariate model showed
that defaulters were nearly ten times more likely to report dissatisfaction
with health care worker attitudes. They were also ten times more
likely to report missing treatment due to health care worker attitudes.
Though it is possible that a small proportion of defaulters went
into treatment with preconceived negative feelings about the health
care services, this bias is unlikely to account for such an overwhelming
disparity of opinion.
Lack of support from family and friends during treatment also appears
to be crucial. Our multivariate model showed that cases were twice
as likely to report feeling ashamed about having MDR TB. Twice as
many cases reported a lack of social support from family or friends
during treatment. This is often reflective of the stigma they feel
is attached to TB, the amount of support they feel in disclosing
their disease to others, and of being seen as a TB patient by friends
Importantly, our model showed that the use of illegal substances
during treatment, such as marijuana or mandrax, was more than ten
times more likely to be reported by cases than controls. The use
of alcohol on an occasional or regular basis was also more commonly
reported by cases. Cases were four times more likely to report having
spent time in prison during MDR TB treatment than controls.
When cases were asked directly why they defaulted treatment, the
most common response was that side effects were too common. It is
possible that patients’ discomfort in having MDR TB, the associated
stigma of having TB, and a poor experience in the health care setting
may translate into a lowered tolerance to side effects from medication.
The outward expression of these difficulties in treatment could
be perceived by patients to be medication side effects.
Based on our findings, we are able to recommend that the National
TB Control Programme consider strengthening the training, supervision,
and support of health care providers of MDR TB patients to avoid
burnout and overwork. We also feel it is necessary for the programme
to provide continuing education for health professionals on the
importance of the patient-provider relationship, and the importance
of the health care provider attitude. Given that support from family
and friends is a crucial component of completing treatment, the
programme should consider supporting patients’ treatment and care
package with family support sessions, treatment counseling, and
substance abuse counseling.
—Submitted by Timothy Holtz, MD, MPH
Div of TB Elimination
- World Health Organization. Global Tuberculosis Control: Surveillance,
Planning, Financing. WHO Report 2003. Geneva, Switzerland: World
Health Organization, 2003; report no. WHO/CDS/TB/2003.316.
- World Health Organization. Tuberculosis control in South Africa:
joint programme review 1996. Geneva: World Health Organization,
- World Health Organization. Anti-tuberculosis drug resistance
in the world. The WHO/IUATLD global project on anti-tuberculosis
drug resistance surveillance. Geneva, World Health Organization,