Impact of a novel molecular TB diagnostic system in patients at high risk of TB mortality in rural South Africa (Uchwepheshe): study protocol for a cluster randomised trial.
ABSTRACT BACKGROUND: Tuberculosis control in sub-Saharan Africa has long been hampered by poor diagnostics and weak health systems. New molecular diagnostics, such as the Xpert(R) MTB/RIF assay, have the potential to improve patient outcomes. We present a cluster randomised trial designed to evaluate whether the positioning of this diagnostic system within the health system has an impact on important patient-level outcomes. METHODS: This pragmatic cluster randomised clinical trial compared two positioning strategies for the Xpert MTB/RIF system: centralised laboratory versus primary health care clinic. The cluster (unit of randomisation) is a 2-week time block at the trial clinic. Adult pulmonary tuberculosis suspects with confirmed human immunodeficiency virus infection and/or at high risk of multidrug-resistant tuberculosis are enrolled from the primary health care clinic. The primary outcome measure is the proportion of culture-confirmed pulmonary tuberculosis cases initiated on appropriate treatment within 30 days of initial clinic visit. Univariate logistic regression will be performed as the primary analysis using generalised estimating equations with a binomial distribution function and a logit link. CONCLUSION: Diagnostic research tends to focus only on performance of diagnostic tests rather than on patient-important outcomes. This trial has been designed to improve the quality of evidence around diagnostic strategies and to inform the scale-up of new tuberculosis diagnostics within public health systems in high-burden settings.Trial registration: Current Controlled Trials ISRCTN18642314; South African National Clinical Trials Registry DOH-27-0711-3568.
- SourceAvailable from: Elke Wynberg[Show abstract] [Hide abstract]
ABSTRACT: Introduction: Point-of-care testing for CD4 cell count is considered a promising way of reducing the time to eligibility assessment for antiretroviral therapy (ART) and of increasing retention in care prior to treatment initiation. In this review, we assess the available evidence on the patient and programme impact of point-of-care CD4 testing. Methods: We searched nine databases and two conference sites (up until 26 October 2013) for studies reporting patient and programme outcomes following the introduction of point-of-care CD4 testing. Where appropriate, results were pooled using random-effects methods. Results: Fifteen studies, mainly from sub-Saharan Africa, were included for review, providing evidence for adults, adolescents, children and pregnant women. Compared to conventional laboratory-based testing, point-of-care CD4 testing increased the likelihood of having CD4 measured [odds ratio (OR) 4.1, 95% CI 3.5-4.9, n=2] and receiving a CD4 result (OR 2.8, 95% CI 1.5-5.6, n=6). Time to being tested was significantly reduced, by a median of nine days; time from CD4 testing to receiving the result was reduced by as much as 17 days. Evidence for increased treatment initiation was mixed. Discussion: The results of this review suggest that point-of-care CD4 testing can increase retention in care prior to starting treatment and can also reduce time to eligibility assessment, which may result in more eligible patients being initiated on ART.Journal of the International AIDS Society 01/2014; 17(1):18809. · 4.21 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: The landscape of diagnostic testing for tuberculosis (TB) is changing rapidly, and stakeholders need urgent guidance on how to develop, deploy and optimize TB diagnostics in a way that maximizes impact and makes best use of available resources. When decisions must be made with only incomplete or preliminary data available, modelling is a useful tool for providing such guidance. Following a meeting of modelers and other key stakeholders organized by the TB Modelling and Analysis Consortium, we propose a conceptual framework for positioning models of TB diagnostics. We use that framework to describe modelling priorities in four key areas: Xpert(®) MTB/RIF scale-up, target product profiles for novel assays, drug susceptibility testing to support new drug regimens, and the improvement of future TB diagnostic models. If we are to maximize the impact and cost-effectiveness of TB diagnostics, these modelling priorities should figure prominently as targets for future research.The International Journal of Tuberculosis and Lung Disease 09/2014; 18(9):1012-8. · 2.76 Impact Factor
STUDY PROTOCOLOpen Access
Impact of a novel molecular TB diagnostic system
in patients at high risk of TB mortality in rural
South Africa (Uchwepheshe): study protocol for a
cluster randomised trial
Richard J Lessells1,2*, Graham S Cooke2,3, Nuala McGrath2,4, Mark P Nicol5, Marie-Louise Newell2,6
and Peter Godfrey-Faussett1
Background: Tuberculosis control in sub-Saharan Africa has long been hampered by poor diagnostics and weak
health systems. New molecular diagnostics, such as the Xpert® MTB/RIF assay, have the potential to improve patient
outcomes. We present a cluster randomised trial designed to evaluate whether the positioning of this diagnostic
system within the health system has an impact on important patient-level outcomes.
Methods/Design: This pragmatic cluster randomised clinical trial compared two positioning strategies for the Xpert
MTB/RIF system: centralised laboratory versus primary health care clinic. The cluster (unit of randomisation) is a 2-
week time block at the trial clinic. Adult pulmonary tuberculosis suspects with confirmed human immunodeficiency
virus infection and/or at high risk of multidrug-resistant tuberculosis are enrolled from the primary health care clinic.
The primary outcome measure is the proportion of culture-confirmed pulmonary tuberculosis cases initiated on
appropriate treatment within 30 days of initial clinic visit. Univariate logistic regression will be performed as the
primary analysis using generalised estimating equations with a binomial distribution function and a logit link.
Conclusion: Diagnostic research tends to focus only on performance of diagnostic tests rather than on patient-
important outcomes. This trial has been designed to improve the quality of evidence around diagnostic strategies
and to inform the scale-up of new tuberculosis diagnostics within public health systems in high-burden settings.
Trial registration: Current Controlled Trials ISRCTN18642314; South African National Clinical Trials Registry DOH-27-
Keywords: Tuberculosis, Multidrug-resistant tuberculosis, HIV, Molecular diagnostics, Point-of-care systems,
Control of the tuberculosis (TB) epidemic in sub-
Saharan Africa is a major public health challenge [1,2].
The epidemic has been exacerbated by the co-existent
explosive human immunodeficiency virus (HIV) epi-
demic and the emergence of drug-resistant Mycobacter-
ium tuberculosis strains leading to high mortality rates
[2,3]. Enshrined in Millennium Development Goal 6
and the Stop TB Partnership Global Plan 2006–2015
are the targets to reduce TB prevalence and TB mortal-
ity rates by 50% (compared to 1990) by 2015 and to
eliminate TB as a public health problem by 2050 [4,5].
At current rates of progress these targets will not be
achieved in sub-Saharan Africa. New interventions and
improved strategies for delivery of interventions are ur-
TB control at present relies primarily on the diagnosis
and treatment of individuals with active TB disease. Early
case detection and initiation of appropriate antituberculous
* Correspondence: firstname.lastname@example.org
1Department of Clinical Research, London School of Hygiene and Tropical
Medicine, Keppel Street, London WC1E 7HT, UK
2Africa Centre for Health and Population Studies, University of KwaZulu-Natal,
Mtubatuba, South Africa
Full list of author information is available at the end of the article
© 2013 Lessells et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative
Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and
reproduction in any medium, provided the original work is properly cited.
Lessells et al. Trials 2013, 14:170
therapy is necessary not only to reduce mortality but
also to interrupt transmission. TB microscopy (still the
most common diagnostic method in use worldwide) is
poorly equipped to control the current TB epidemic in
sub-Saharan Africa given its poor sensitivity, particu-
larly in HIV co-infection, and inability to detect drug re-
sistance . Additionally, the placement of diagnostics
in centralised facilities distant from where patients seek
care contributes to significant delays [7,8] and default
[9-13] before initiation of treatment. The impact of this
is illustrated most starkly in multidrug-resistant TB
(MDR-TB), where delays in culture and drug suscepti-
bility testing (DST) techniques mean that 50% of pa-
tients have died by the time their culture/DST result is
The development of novel molecular tools, in particu-
lar the Xpert® MTB/RIF assay, offers new opportunities
to tackle these problems. This is a fully automated,
closed cartridge diagnostic system that utilises hemi-
nested polymerase chain reaction (PCR) and molecular
beacon technology to detect the presence of Mycobacter-
ium tuberculosis and rifampicin-resistant mutations dir-
ectly from clinical samples in less than 2 h [16-18].
The World Health Organization (WHO) recommended
the system be implemented in high-burden settings on the
basis of initial data from validation and demonstration
studies [19-21]. Many countries are now moving ahead
with implementation and there is a need for research to
address key questions in the early phase of implementa-
tion so as to inform future scale-up . One critical
question relates to the optimal positioning of the diagnos-
tic system within different health systems, and this is the
focus of the research study.
The primary objective is to test the hypothesis that
timely initiation of appropriate TB treatment when the
diagnostic system is positioned at the primary health
care clinic (point of care) is different from when the
diagnostic system is positioned centrally at the district
hospital laboratory. Secondary objectives are:
? To evaluate the impact of Xpert MTB/RIF
positioning on additional clinical outcomes
(mortality, hospital admission, time to initiation of
? To explore the cost-effectiveness of Xpert MTB/RIF
implementation at primary health care clinic level
? To compare the operational feasibility of Xpert
MTB/RIF placement at the primary health care
clinic level and district hospital laboratory level.
The trial is being conducted in Hlabisa health sub-
district, uMkhanyakude district, northern KwaZulu-
Natal, South Africa (Figure 1). This area has an ex-
tremely high dual burden of TB and HIV: the TB notifi-
cation rate for the sub-district in 2010 was 1,130/
100,000; HIV seroprevalence in the adult population
(≥15 years) within the Africa Centre surveillance area
was 24.1% in 2010; in 2008, 76% of TB cases were asso-
ciated with HIV infection . In the years 2000–2006
HIV and TB accounted for 71.5% of deaths in young
adults (25–49 years) in the Africa Centre surveillance
area . HIV and TB treatment and care are deliv-
ered at 17 primary health care (PHC) clinics through
decentralised collaborative programmes. Participants
are recruited from the largest PHC clinic that is situ-
ated within a small urban township in the south of the
sub-district, approximately 60 km by road from the dis-
The study is a pragmatic cluster randomised clinical trial
comparing two positioning strategies for the Xpert
MTB/RIF system: positioning at centralised laboratory
level (district hospital laboratory) versus positioning at
primary health care clinic level (point of care). The clus-
ter (unit of randomisation) is a 2-week time block at the
primary health care clinic (clinic blocks), and clusters
are randomly assigned to the district hospital laboratory
strategy or point-of-care strategy. The trial schema is
shown in Figure 2.
Adult (≥18 years) pulmonary TB suspects with con-
firmed HIV infection and/or at high risk of MDR-TB are
included after giving informed consent. These criteria
were defined because of the high risk for mortality in
these groups and prioritisation for Xpert MTB/RIF test-
ing, in line with the WHO recommendations . A TB
suspect is defined for the purposes of the trial as an indi-
vidual with a current cough (of any duration) with or
without other symptoms. High risk of MDR-TB is de-
fined according to national and international guidelines
and incorporates the following categories: failure of the
standard treatment regimen (2HRZE/4HR), failure of the
re-treatment regimen (2HRZES/1HRZE/5HRE), acid-fast
bacilli (AFB) smear non-conversion at month 2 or 3 of
the standard or re-treatment regimen, relapse or return
after default, any other previous TB (4 or more weeks of
TB treatment), household contact with a known MDR-
TB case, prison inmate within the last 12 months and
health care worker [24,25]. Participants are excluded
if they report a previous diagnosis of MDR-TB or exten-
sively drug-resistant TB (XDR-TB), are severely unwell
requiring admission to hospital, or are unable to give in-
formed consent. Participants are recruited between the
times of 0800 and 1630, Monday to Friday.
Lessells et al. Trials 2013, 14:170
Page 2 of 10
All participants provide two spontaneously expectorated
sputum specimens on the day of enrolment (spot speci-
mens). The first sputum specimen is submitted for Xpert
MTB/RIF testing. The second specimen is submitted for
Mycobacterial Growth Indicator Tube (MGIT) culture,
line probe assay (LPA) ± phenotypic drug susceptibility
testing (DST). In both strategies, the specimen for cul-
ture/LPA/DST is transported daily (in the afternoon) to
the National Health Laboratory Service (NHLS) labora-
tory at the district hospital and then onwards to the pro-
vincial NHLS referral laboratory. The results of this are
used to define TB cases and to define the primary out-
District hospital laboratory strategy
Sputum specimens are transported on a daily basis to the
National Health Laboratory Service (NHLS) laboratory at
the district hospital using the routine sample transport
system. Xpert MTB/RIF testing is performed by a trained
laboratory technician at the earliest convenience (within
24 h of the specimen being received in the laboratory) and
printed results are returned to the clinic using the same
routine transport system. Under this strategy, participants
are requested to return for results after 72 h.
Point-of-care (POC) strategy
The diagnostic system is located at the primary health care
in a dedicated room close to the TB clinic (Figure 3).
Figure 1 Maps showing location of (a) the study site and (b) primary health care clinic (trial clinic) and district hospital.
Lessells et al. Trials 2013, 14:170
Page 3 of 10
Figure 2 Trial schema.
Figure 3 Professional nurse operating the Xpert MTB/RIF system at the primary health care clinic.
Lessells et al. Trials 2013, 14:170
Page 4 of 10
Xpert MTB/RIF testing is performed immediately by the
study nurse, on the same day where possible. Participants
are invited to wait for the result (approximately 2 h) or, if
they are unable to wait or it is towards the end of the
working day, they are advised to return the following day.
The observational unit for all analyses is the individual
participant. The primary outcome for the study is the
proportion of culture-confirmed pulmonary TB cases
initiated on appropriate TB treatment within 30 days of
according to results of LPA ± phenotypic DST on the
Secondary outcomes at an individual level are the fol-
lowing, with all time-to-event analyses using the initial
clinic visit as time zero:
? All-cause mortality in TB suspects and MDR-TB
suspects at 60 days
? Time to initiation of appropriate TB treatment
(days) for culture-confirmed pulmonary TB cases
? Time to initiation of MDR-TB treatment (for MDR-
TB cases confirmed by culture/LPA/DST)
? Proportion of TB suspects and MDR-TB suspects
with at least one hospital attendance within 60 days
? Time to initiation of antiretroviral therapy (ART) for
HIV-infected TB suspects and MDR-TB suspects
not yet receiving but eligible for ART
? Sensitivity and specificity of Xpert MTB/RIF
▪ for M. tuberculosis detection (compared to
reference standard of single MGIT culture)
▪ for detection of rifampicin resistance (compared
to reference standard of phenotypic DST ± LPA)
The study was designed to detect an increase from 85%
to 95% in the proportion of culture-confirmed pulmon-
ary TB cases initiated on appropriate treatment within
30 days. Sample size was calculated with the equation of
Hayes and Bennett, using the coefficient of variation (κ)
. With κ = 0.05 and a cluster size of 12 culture-
positive cases, we would need 16 clusters and 188
culture-positive TB cases in each arm to detect this dif-
ference with 95% confidence and 80% power. We as-
sumed 10% of individual participants would be lost to
follow-up at 60 days, so we would need 208 culture-
positive TB cases in each arm. Based on the assumption
that 25% of TB suspects would have a positive MGIT
culture, we would require enrolment of 1,664 TB sus-
pects. The total sample size will therefore be 32 clusters
and 1,664 individual participants.
The coefficient of variation (κ) is small, but as the clus-
ters are clinic time blocks rather than geographic areas or
health care facilities, minimal variation is expected be-
tween clusters. This value of κ corresponds to a range of
proportions appropriately treated within 30 days in the
district hospital laboratory arm of 77-94%.
For the secondary endpoint of all-cause mortality
within 60 days, the analysis will incorporate all partici-
pants (all suspects), regardless of presence or absence of
TB disease. The sample size of 32 clusters and 60 partic-
ipants per cluster gives approximately 80% power to de-
tect a 33% reduction in mortality from 12% in the
district hospital laboratory arm to 8% in the point-of-
care arm, with 95% confidence.
The allocation schedule for random assignment of 2-
week time blocks was computer generated, using ran-
dom permuted blocks of eight. Allocation for each clinic
block was placed into sealed envelopes by the statisti-
cian; the principal investigator opens the envelope on
the Friday before the start of a new 2-week block and
communicates the allocation for the next 2 weeks to
Health care workers at the primary health care clinic
identify potential participants. All individuals reporting
cough are referred to the study nurse. Eligibility criteria
are checked by the nurse, and subjects meeting the in-
clusion criteria are provided spoken and written infor-
mation about the study in isiZulu and/or English; those
willing to participate are taken through the informed
consent process and are asked to provide a signature or
thumbprint on the consent form.
A baseline assessment is performed by the study
nurse. Demographic information, current symptoms,
previous TB history, risk factors for drug resistance,
HIV status, and history of ART use are documented on
a case report form.
With both strategies, clinical decisions are made by
the study nurse on the basis of the Xpert MTB/RIF re-
sult and according to pre-defined algorithms. TB pa-
tients without resistance to rifampicin are commenced
on standard anti-TB therapy (4HRZE/2HR) by the study
nurse. All patients with rifampicin-resistant TB are
reported to the trial physician on the same day and are
subsequently referred to the specialist drug-resistant TB
treatment centre in Durban. Management of suspects
with a negative Xpert MTB/RIF follows existing proto-
cols for smear-negative TB suspects: oral antibiotics are
prescribed and patients are advised to return if symp-
toms do not improve after 14 days. Patients who remain
symptomatic following this course of antibiotics can be
referred to the district hospital for chest X-ray and phys-
ician review. Results from MGIT culture and DST are
Lessells et al. Trials 2013, 14:170
Page 5 of 10
returned through the routine laboratory system and are
used to guide clinical management where appropriate.
To ascertain the primary and secondary outcomes, at
enrolment all participants are allocated a review date 2
months from the enrolment visit. Participants are invited
to attend clinic for review but are also invited to consent
to telephonic follow-up and/or home visit in case clinic
visit is not possible. Additional contact details are pro-
vided for at least one other family member (or other per-
son designated by participant) at enrolment, wherever
possible. Participants are told that, when attending
the clinic for the follow-up visit, they will be reim-
bursed with a ZAR 50 grocery voucher (approximately
equivalent to USD 6). Outcome data pertaining to TB
treatment initiation, additional investigations, hospital
attendances and admissions, and ART initiation (where
appropriate) are collected on a case report form by the
study nurse. In the event that no contact is made with
patient or with named contact persons, follow-up in-
formation is collected from the clinic TB registers and
the operational HIV programme database – permission
to use these data is also included in the informed con-
Analysis of baseline characteristics will be performed to
characterise the study population and to identify base-
line imbalances between the study arms in order to de-
cide whether any covariates need to be adjusted for in
the final analyses. The baseline data will include: age,
sex, body mass index (BMI), history of previous TB, HIV
infection status, CD4+ cell count, and use of antiretro-
viral therapy and isoniazid preventive therapy.
All final analyses will be intention-to-treat analyses
performed at the individual level taking account of
within-cluster correlation. The definition of TB cases for
the primary outcome will be based on MGIT culture
positivity. The proportion of TB cases initiated on ap-
propriate TB treatment within 30 days will be based
on whether the appropriate treatment regimen was
commenced within 30 days of the initial clinic visit—
appropriate regimens are defined according to drug
susceptibility pattern and with reference to national
guidelines (Table 1) . The primary analysis will in-
clude only TB cases not on TB treatment at the time of
enrolment, i.e. excluding smear non-converters or fail-
ures still on treatment. The primary outcome is a binary
variable (initiation of appropriate treatment or not) so uni-
variate logistic regression will be performed as the primary
analysis using generalised estimating equations (GEE) with
a binomial distribution function and a logit link . The
odds ratio will be reported with 95% confidence intervals
and a p-value from the Wald test. This method will allow
for the correlation between observations (within clusters)
without needing to specify a distributional assumption for
the correlations. In addition, important individual-level
characteristics that are unbalanced between arms will be
included in the model as covariates. For the secondary
outcomes with binary variables, GEE models will also be
fitted with a binomial distribution function and a logit
link. For the secondary outcomes with time-to-event mea-
sures, Cox proportional hazard models will be fitted with
the shared frailty option to account for the cluster ran-
domisation . Hazard ratios will be presented with 95%
The diagnostic performance of Xpert MTB/RIF will be
compared between the two arms. Estimation of sensitiv-
ity and specificity of Xpert MTB/RIF for the detection of
M. tuberculosis against the reference standard of single
MGIT culture will be based on complete case analysis
(participants with paired valid Xpert MTB/RIF and
MGIT culture results) and will only include individuals
not on TB treatment at the time of enrolment. Estima-
tion of sensitivity and specificity of Xpert MTB/RIF for
the detection of rifampicin resistance against the refer-
ence standard of genotypic ± phenotypic DST on the
culture isolate will be based on participants with M.
tuberculosis detected by Xpert MTB/RIF and with a
positive MGIT culture and valid drug susceptibility test
(LPA ± phenotypic DST) results. This will include indi-
viduals on TB treatment at the time of enrolment (e.g.
participants with AFB smear non-conversion or treat-
In addition to evaluating the effectiveness of point-
of-care positioning of Xpert MTB/RIF, data from the
trial will be combined with those from a costing ana-
lysis to explore the cost-effectiveness of point-of-care
placement. Health system costs will be obtained
through monitoring of study expenditure and inter-
views with health service management. Collection of
data relating to patient and household costs will be
nested within the trial—this will involve additional
data collected from a subset of patients at baseline
and at the 2-month follow-up to determine direct
and indirect costs incurred during the diagnostic
presented in Table 2. The health system costs and
patient costs will be combined with the outcome
data to generate an average incremental cost per TB
case appropriately treated.
The study will also compare the operational feasibility
of Xpert MTB/RIF implementation at the hospital
Lessells et al. Trials 2013, 14:170
Page 6 of 10
laboratory and at the primary health care clinic. This en-
compasses an assessment of the performance and ro-
bustness of the system, as well as evaluation of the
practicality of operating the system at the laboratory and
at the clinic. The key indicators to be assessed are
displayed in Table 3. Data on these indicators will be
collected throughout the trial.
The study has been approved by the Biomedical Re-
search Ethics Committee of the University of KwaZulu-
Natal (BF033/11), the Ethics Committee of the London
School of Hygiene and Tropical Medicine (5926), and
the Health Research Committee of the KwaZulu-Natal
Department of Health (HRKM084/11). Permission for
the study was granted by Hlabisa Hospital and by the
Community Advisory Board of the Africa Centre for
Health and Population Studies.
Given that the units of randomisation are time blocks,
it is not possible for individuals to consent to random-
isation. Individual consent for participation remains im-
portant given that the intervention is delivered to
individuals and that individual-level data are collected at
enrolment and at follow-up.
There are TB suspects who are not eligible for this
study and therefore will not have access to Xpert MTB/
RIF testing within the study (suspects who are neither
HIV-infected nor at high risk for MDR-TB). The justifi-
cation for this is that these groups were a lower priority
for this intervention given the much lower mortality
rates and these suspects continue to receive diagnostic
evaluation including sputum microscopy ± culture/LPA/
DST according to national guidelines. At the time of
study design it was predicted that, were Xpert MTB/RIF
to be implemented in South Africa, the WHO recom-
mendations would be followed (use in HIV-infected and
those at high risk of MDR-TB). Although the national
roll-out plan went beyond this in incorporating its use
for all TB suspects, Xpert MTB/RIF has not yet been
installed in Hlabisa sub-district and is therefore not yet
available in the sub-district outside the trial.
Evaluation of diagnostic tools provides different chal-
lenges than those of therapeutic interventions. Diagnos-
tic accuracy studies are usually the starting point for
evaluation of new technologies, yet to inform public
health policies and implementation it is crucial to evalu-
ate patient-important outcomes . The ultimate
Table 2 Components of cost analysis
Health service costsPatient and household costs
Direct costs associated
with diagnostic services
Building space Transport to/from clinic
(patient ± carer)
UtilitiesTransport to/from hospital
(patient ± carer)
Internal/external QC OPD attendance
Xpert MTB/RIF testsTraditional healer consultation
Consumables (gloves, N95 masks)
Staff work time (based on time analysis)
Direct costs associated
with medical services
First-line TB therapyLost time (salary) at work
(patient ± carer)
Table 1 Definitions of appropriate initial anti-TB drug regimen for primary outcome measurement
Case definition* Appropriate initial anti-TB drug regimen
M. tuberculosis susceptible to rifampicin and isoniazidIsoniazid + rifampicin + pyrazinamide + ethambutol
M. tuberculosis with mono-resistance to isoniazidIsoniazid + rifampicin + pyrazinamide + ethambutol
M. tuberculosis with mono-resistance to rifampicin Standardised second-line regimen§ with isoniazid
Multidrug-resistant M. tuberculosis (MDR-TB)†Standardised second-line regimen§
Standardised XDR-TB regimenǁ
Extensively drug-resistant M. tuberculosis (XDR-TB)†
* Case definition based on results of MGIT culture + line probe assay + phenotypic DST.
† MDR-TB defined as resistance to rifampicin and isoniazid.
† XDR-TB defined as MDR plus resistance to ofloxacin and kanamycin.
§Standardised regimen according to national guidelines (kanamycin/amikacin + fluoroquinolone + ethionamide + cycloserine/terizidone ± pyrazinamide ±
ǁStandardised regimen according to national guidelines (capreomycin + fluoroquinolone + ethionamide + cycloserine/terizidone + PAS + clofazimine) .
Lessells et al. Trials 2013, 14:170
Page 7 of 10
impact of a new diagnostic test should be measured by
its capacityto generate
randomised trial is the most rigorous design to evaluate
clinical outcomes from a diagnostic intervention. Indi-
vidual randomisation was considered logistically challen-
ging and potentially disruptive in the context of a busy
clinic and laboratory system, although this would have
been the most efficient statistical design . Cluster
randomisation by health care facility was not possible
given the limited resources. Therefore cluster random-
isation by time block was the preferred study design to
maximise internal validity and to minimise disruption to
clinic services. Other trial designs were considered (non-
randomised controlled trial with allocation by day/week/
month or controlled before and after study) but were felt
to be inferior in addressing the research hypothesis,
mainly because of the potential for bias and therefore
loss of internal validity. There have been few published
trials where the unit of randomisation is a time block ra-
ther than a geographical or organisational cluster. The
trials that have been published have not used consistent
methods for sample size calculation—some have ad-
justed appropriately for cluster variation [31-33] whereas
others have arbitrarily inflated the sample size from that
for an individual RCT  and others have based the
sample size on the numbers available to participate .
Blinding of patients or of health care workers is not
feasible in this pragmatic diagnostic trial because alloca-
tion to trial arms involves different actions by the patient
and the clinical staff. As a result of this, the outcomes
are as objective as possible to limit potential bias from
differential ascertainment of outcomes in the two arms.
The possibility of differential recruitment into the trial
arms exists but will be minimised by standardised refer-
ral criteria for the clinic health care workers; recruitment
will be monitored by reviewing the clinic records to as-
certain what proportion of patients with cough were re-
ferred to the study during each 2-week block. This will
be reported if there is a major imbalance in recruitment
to the two trial arms. There is a further risk of selection
bias if there is differential non-participation. The propor-
tions of eligible subjects consenting by trial arm will be
monitored and will be reported accurately at the con-
clusion of the trial. There is some risk of contamination
between the arms if, over time with point-of-care test-
ing, the health workers see the importance and the ef-
fect of receiving the test result in a timely fashion and
this then improves their ability to encourage all sus-
pects to return and receive their result. This would
tend to bias the findings towards the null hypothesis.
We will explore this by assessment of the variability in
the proportion returning for their test result by cluster
and by time period.
The evaluation of diagnostic accuracy is not the pri-
mary focus of the trial and the reference standard of a
single culture could be considered an imperfect gold
standard. In the initial Xpert MTB/RIF clinical validity
studies, the reference standard used results of liquid and
solid culture on two specimens (four culture results in
total) . Conversely, in the later demonstration stud-
ies, the reference standard varied between study sites
and in some sites included results of only a single cul-
ture . Observational data from the district in 2007
suggested that 5% of all culture-positive cases were
multidrug-resistant . Given that we will preferentially
include suspects with a high risk of MDR-TB, we expect
the overall proportion with MDR-TB to be at least 10%.
It is possible that the impact of Xpert MTB/RIF posi-
tioning may be different for the drug-susceptible and
drug-resistant cases. If this is the case then a higher or
lower than expected proportion with MDR-TB could
modify the effect of point-of-care placement, and this
will be explored in secondary analyses.
There are a number of trials evaluating the impact of
Xpert MTB/RIF in different settings and with different
research hypotheses. Information about research pro-
jects is collated by the TREAT TB Xpert Research
Mapping Project . Several studies are examining
point-of-care implementation but, to our knowledge,
this is the only study directly comparing point-of-care
use to centralised laboratory use. There is already some
evidence from South Africa of the feasibility of imple-
mentation at the primary health care level, although
several operational challenges were experienced when
Table 3 Indicators for operational feasibility evaluation
IndicatorMethod of measurement
Power supplyTime log for power cuts/
Operating temperature for
Storage temperature for
Xpert MTB/RIF kits
Hands-on user timeActivity log
Indeterminate resultsGeneXpert software
Data errors (incomplete
Maintenance needsRequirement for supplier support
Training requirementsRecording of initial and follow-up
Supervision requirementsLog of assistance from other
Waste managementRecording of problems with disposal
of used cartridges
User appraisalRegular appraisal by laboratory staff
and study staff
User performanceRegular independent observation of
Lessells et al. Trials 2013, 14:170
Page 8 of 10
implemented within a very large urban clinic [38,39].
This study should provide direct evidence of any bene-
fits of point-of-care positioning as well as further infor-
mation about the costs and logistical challenges of
such strategies. This can also be considered as a proof-
of-principle study that will help to understand the ben-
efits of bringing diagnostics closer to patients, and this
will have broader relevance as we continue to develop
and evaluate diagnostic technologies suitable for point-
of-care use [40,41].
The study received final ethical approval in June 2011.
Enrolment commenced on 22 August 2011. Enrolment
is scheduled to complete in March 2013 and follow-up
will be complete in May 2013.
ART: Antiretroviral therapy; DST: Drug susceptibility testing; GEE: Generalised
estimating equation; HIV: Human immunodeficiency virus; MDR-
TB: Multidrug-resistant TB; MGIT: Mycobacterial growth indicator tube;
NHLS: National Health Laboratory Service; PCR: Polymerase chain reaction;
PHC: Primary health care; TB: Tuberculosis; WHO: World Health Organization;
XDR-TB: Extensively drug-resistant TB.
The authors declare that they have no competing interests.
RJL and PGF conceived and designed the trial, with additional input from
GSC, NM, MPN and MLN. NM helped with the statistical design and analysis
plan. RJL wrote the first draft of the manuscript. All authors contributed to
revision of the manuscript and approved the final version.
The authors would like to acknowledge the independent members of the
Trial Steering Committee (David Moore, Katherine Fielding and Yunus
Moosa) who contributed to discussions about trial design, conduct and
analysis. The authors would also like to thank the study team and all staff
working at the clinic.
Source of funding
This work is supported by the Wellcome Trust (grant no. 090999/Z/09/Z). NM
is supported by a Wellcome Trust fellowship grant (083495MA). The funder
had no role in the study design, data collection and analysis, decision to
publish or preparation of the manuscript.
1Department of Clinical Research, London School of Hygiene and Tropical
Medicine, Keppel Street, London WC1E 7HT, UK.2Africa Centre for Health
and Population Studies, University of KwaZulu-Natal, Mtubatuba, South
Africa.3Department of Infectious Disease, Imperial College, London, UK.
4Academic Unit of Primary Care and Population Sciences and Academic Unit
of Social Sciences, University of Southampton, Southampton, UK.5Division of
Medical Microbiology, University of Cape Town, Cape Town, South Africa.
6UCL Institute of Child Health, London, UK.
Received: 22 February 2013 Accepted: 28 May 2013
Published: 12 June 2013
1. Chaisson RE, Martinson NA: Tuberculosis in Africa–combating an HIV-
driven crisis. N Engl J Med 2008, 358:1089–1092.
2.Abdool Karim SS, Churchyard GJ, Abdool Karim Q, Lawn SD: HIV infection
and tuberculosis in South Africa: an urgent need to escalate the public
health response. Lancet 2009, 374:921–933.
3.Gandhi NR, Nunn P, Dheda K, Schaaf HS, Zignol M, van Soolingen D, Jensen
P, Bayona J: Multidrug-resistant and extensively drug-resistant
tuberculosis: a threat to global control of tuberculosis. Lancet 2010,
Stop TB Partnership, World Health Organization: The global plan to stop TB,
2006–2015. Geneva: World Health Organization; 2006.
United Nations General Assembly: Road map toward the implementation of
the United Nations Millennium Declaration. http://www.un.org/
Parsons LM, Somoskovi A, Gutierrez C, Lee E, Paramasivan CN, Abimiku A,
Spector S, Roscigno G, Nkengasong J: Laboratory diagnosis of tuberculosis
in resource-poor countries: challenges and opportunities. Clin Microbiol
Rev 2011, 24:314–350.
Sreeramareddy CT, Panduru KV, Menten J, Van den Ende J: Time delays in
diagnosis of pulmonary tuberculosis: a systematic review of literature.
BMC Infect Dis 2009, 9:91.
Storla DG, Yimer S, Bjune GA: A systematic review of delay in the
diagnosis and treatment of tuberculosis. BMC Publ Health 2008, 8:15.
Botha E, den Boon S, Lawrence KA, Reuter H, Verver S, Lombard CJ, Dye C,
Enarson DA, Beyers N: From suspect to patient: tuberculosis diagnosis
and treatment initiation in health facilities in South Africa. Int J Tuberc
Lung Dis 2008, 12:936–941.
Botha E, Den Boon S, Verver S, Dunbar R, Lawrence KA, Bosman M, Enarson
DA, Toms I, Beyers N: Initial default from tuberculosis treatment: how
often does it happen and what are the reasons? Int J Tuberc Lung Dis
Creek TL, Lockman S, Kenyon TA, Makhoa M, Chimidza N, Moeti T, Sarpong
BB, Binkin NJ, Tappero JW: Completeness and timeliness of treatment
initiation after laboratory diagnosis of tuberculosis in Gaborone,
Botswana. Int J Tuberc Lung Dis 2000, 4:956–961.
Nyirenda T, Harries AD, Banerjee A, Salaniponi FM: Registration and
treatment of patients with smear-positive pulmonary tuberculosis. Int J
Tuberc Lung Dis 1998, 2:944–945.
Squire SB, Belaye AK, Kashoti A, Salaniponi FM, Mundy CJ, Theobald S, Kemp
J: ‘Lost’ smear-positive pulmonary tuberculosis cases: where are they and
why did we lose them? Int J Tuberc Lung Dis 2005, 9:25–31.
Gandhi NR, Shah NS, Andrews JR, Vella V, Moll AP, Scott M, Weissman D,
Marra C, Lalloo UG, Friedland GH: HIV coinfection in multidrug- and
extensively drug-resistant tuberculosis results in high early mortality. Am
J Respir Crit Care Med 2010, 181:80–86.
Heller T, Lessells RJ, Wallrauch CG, Barnighausen T, Cooke GS, Mhlongo L,
Master I, Newell ML: Community-based treatment for multidrug-resistant
tuberculosis in rural KwaZulu-Natal, South Africa. Int J Tuberc Lung Dis
Helb D, Jones M, Story E, Boehme C, Wallace E, Ho K, Kop J, Owens MR,
Rodgers R, Banada P, et al: Rapid detection of Mycobacterium
tuberculosis and rifampin resistance by use of on-demand, near-patient
technology. J Clin Microbiol 2010, 48:229–237.
Boehme CC, Nabeta P, Hillemann D, Nicol MP, Shenai S, Krapp F, Allen J,
Tahirli R, Blakemore R, Rustomjee R, et al: Rapid molecular detection of
tuberculosis and rifampin resistance. N Engl J Med 2010, 363:1005–1015.
Boehme CC, Nicol MP, Nabeta P, Michael JS, Gotuzzo E, Tahirli R, Gler MT,
Blakemore R, Worodria W, Gray C, et al: Feasibility, diagnostic accuracy,
and effectiveness of decentralised use of the Xpert MTB/RIF test for
diagnosis of tuberculosis and multidrug resistance: a multicentre
implementation study. Lancet 2011, 377:1495–1505.
World Health Organization: Policy statement: automated real-time nucleic
acid amplification technology for rapid and simultaneous detection of
tuberculosis and rifampicin resistance: Xpert MTB/RIF system. Geneva: World
Health Organization; 2011.
World Health Organization: Rapid implementation of the Xpert MTB/RIF
diagnostic test: technical and operational ‘How-to’; practical considerations.
Geneva: World Health Organization; 2011.
Weyer K, Mirzayev F, Migliori G, Van Gemert W, D’Ambrosio L, Zignol M,
Floyd K, Centis R, Cirillo D, Tortoli E, et al: Rapid molecular TB diagnosis:
evidence, policy-making and global implementation of Xpert(R)MTB/RIF.
Eur Respir J 2012 [Epub ahead of print].
Wallrauch C, Heller T, Lessells R, Kekana M, Barnighausen T, Newell ML: High
uptake of HIV testing for tuberculosis patients in an integrated primary
health care HIV/TB programme in rural KwaZulu-Natal. S Afr Med J 2010,
Lessells et al. Trials 2013, 14:170
Page 9 of 10
23. Herbst AJ, Cooke GS, Barnighausen T, KanyKany A, Tanser F, Newell ML:
Adult mortality and antiretroviral treatment roll-out in rural KwaZulu-
Natal, South Africa. Bull World Health Organ 2009, 87:754–762.
World Health Organization: Guidelines for the programmatic management of
drug-resistant tuberculosis: emergency update 2008. Geneva: World Health
Department of Health RoSA: National Tuberculosis Management Guidelines.
Pretoria: Department of Health; 2009.
Hayes RJ, Bennett S: Simple sample size calculation for cluster-
randomized trials. Int J Epidemiol 1999, 28:319–326.
Department of Health RoSA: Management of drug-resistant tuberculosis:
Policy guidelines. Pretoria: Department of Health; 2011.
Hayes RJ, Moulton LH: Cluster randomised trials. Boca Raton, Florida:
Chapman & Hall/CRC; 2009.
Glidden DV, Vittinghoff E: Modelling clustered survival data from
multicentre clinical trials. Stat Med 2004, 23:369–388.
Lessells RJ, Cooke GS, Newell ML, Godfrey-Faussett P: Evaluation of
tuberculosis diagnostics: establishing an evidence base around the
public health impact. J Infect Dis 2011, 204(Suppl 4):S1187–S1195.
Bagchi AD, Dale S, Verbitsky-Savitz N, Andrecheck S, Zavotsky K, Eisenstein
R: Examining effectiveness of medical interpreters in emergency
departments for Spanish-speaking patients with limited English
proficiency: results of a randomized controlled trial. Ann Emerg Med 2011,
Loten C, Attia J, Hullick C, Marley J, McElduff P: Validation of a point of
care troponin assay in real life Emergency Department conditions. Emerg
Med Australas 2009, 21:286–292.
Zwarenstein MF, Dainty KN, Quan S, Kiss A, Adhikari NK: A cluster
randomized trial evaluating electronic prescribing in an ambulatory care
setting. Trials 2007, 8:28.
Hall S, Reid E, Ukoumunne OC, Weinman J, Marteau TM: Brief smoking
cessation advice from practice nurses during routine cervical smear tests
appointments: a cluster randomised controlled trial assessing feasibility,
acceptability and potential effectiveness. Br J Cancer 2007, 96:1057–1061.
Kunkler IH, Prescott RJ, Lee RJ, Brebner JA, Cairns JA, Fielding RG, Bowman
A, Neades G, Walls AD, Chetty U, et al: TELEMAM: a cluster randomised
trial to assess the use of telemedicine in multi-disciplinary breast cancer
decision making. Eur J Cancer 2007, 43:2506–2514.
Wallengren K, Scano F, Nunn P, Margot B, Buthelezi SS, Williams B, Pym A,
Samuel EY, Mirzayev F, Nkhoma W, et al: Drug-resistant tuberculosis,
KwaZulu-Natal, South Africa, 2001–2007. Emerg Infect Dis 2011,
Xpert Research Mapping Project: [http://xrmt.treattb.org/]
Clouse K, Page-Shipp L, Dansey H, Moatlhodi B, Scott L, Bassett J, Stevens
W, Sanne I, Van Rie A: Implementation of Xpert MTB/RIF for routine point-
of-care diagnosis of tuberculosis at the primary care level. S Afr Med J
Van Rie A, Page-Shipp L, Hanrahan CF, Schnippel K, Dansey H, Bassett J,
Clouse K, Scott L, Stevens W, Sanne I: Point-of-care Xpert MTB/RIF for
smear-negative tuberculosis suspects at a primary care clinic in South
Africa. Int J Tuberc Lung Dis 2013, 17:368–372.
Batz HG, Cooke GS, Reid SD: Towards lab-free tuberculosis diagnosis. http://
Dheda K, Ruhwald M, Theron G, Peter J, Yam WC: Point-of-care diagnosis
of tuberculosis: past, present and future. Respirology 2013, 18:217–232.
Cite this article as: Lessells et al.: Impact of a novel molecular TB
diagnostic system in patients at high risk of TB mortality in rural South
Africa (Uchwepheshe): study protocol for a cluster randomised trial.
Trials 2013 14:170.
Submit your next manuscript to BioMed Central
and take full advantage of:
• Convenient online submission
• Thorough peer review
• No space constraints or color figure charges
• Immediate publication on acceptance
• Inclusion in PubMed, CAS, Scopus and Google Scholar
• Research which is freely available for redistribution
Submit your manuscript at
Lessells et al. Trials 2013, 14:170
Page 10 of 10