XTEND Study2018-09-16T19:40:50+00:00

Xpert MTB/RIF for diagnosis of tuberculosis: evaluating impact and cost-effectiveness in the routine roll-out in South Africa

“Xpert for TB: Evaluating a New Diagnostic” (XTEND) trial


South Africa is rolling out Xpert MTB/RIF, a new diagnostic test with improved sensitivity for detection of active tuberculosis combined with simultaneous detection of rifampicin resistance. The test has a greatly reduced turn-around time compared with mycobacterial culture, but a cost per test substantially greater than the smear microscopy it is intended to replace. It is not yet clear whether this will improve patient-relevant outcomes, and how it will contribute to tuberculosis control (both drug susceptible and drug resistant) in South Africa and elsewhere.


To evaluate the effectiveness and cost effectiveness of Xpert MTB/RIF in the investigation of TB and TB drug resistance, and its impact on patient and programme outcomes and transmission at a population level.

Specific Objectives

  • To measure the effectiveness of Xpert MTB/RIF in improving patient and programme outcomes
  • To determine the likely population level impact on TB transmission of using Xpert MTB/RIF in the investigation of TB and TB drug resistance, using mathematical modelling
  • To estimate the cost-effectiveness of Xpert MTB/RIF from a patient and health system perspective


Study Design

A cluster randomised pragmatic trial (CRT) will be conducted, where 20 laboratories in South Africa, in high burden TB districts will be randomised to receive Xpert MTB/Rif technology, or not.

A cluster randomised pragmatic trial (CRT) will be conducted, where 20 laboratories in South Africa, in high burden TB districts will be randomised to receive Xpert MTB/Rif technology, or not.

  • Intervention laboratories will receive Xpert MTB/RIF.
    • TB suspects who attend facilities that use these laboratories will have one Xpert MTB/RIF instead of two smears at diagnosis.
    • Patients that are sputum Xpert MTB positive will have a second sputum specimen collected for smear microscopy (with additional culture and DST for those detected with rifampicin resistant TB), in line with South African national guidelines.
  • Control laboratories will use standard of care diagnostic tests for TB, as per South African national guidelines:
    • two smears using fluoresecent staining (Auramine O) will be done for all TB suspects,
    • Culture will be  done if any of the following criteria are met:
      • Suspects are smear negative, and symptomatic, including those who are HIV positive;
      • high MDR-TB risk;
      • a prior history of TB.

Selection and randomisation of laboratories and relationship between this study and the National Rollout of Xpert MTB/Rif by the NHLS

The NHLS is currently rolling out Xpert MTB/Rif in a phased way. The laboratories in high burden TB districts have all already received Xpert. The 20 laboratories in medium to high districts are all due to receive Xpert MTB/Rif as part of this study. In other words, this study is ‘nested’ within the national roll out of Xpert. The laboratories that are participating in this study were identified in early 2011 by the NDoH and role-players. For planning needs, the random allocation of laboratories in this study to receive Xpert diagnostic tests early (Dec/Jan 2011) or late (Aug/Sept 2012) was done earlier in 2011.

Province District Intervention Laboratories Control Laboratories
Eastern Cape Chris Hani All Saints Cofimvaba
Chris Hani Queenstown
Alfred Nzo Mary Theresa
Alfred Nzo Mt Ayliff
O R Tambo St Patrick Taylor Bequest
Nelson Mandela Bay Metro Uitenhage
Free State Thabo Mofutsanyana Bethlehem Manapo
Gauteng Ekurhuleni Metro Boksburg Benoni Natalspruit
Ekurhuleni Metro Tembisa
City Of Johannesburg Helen Joseph
City Of Tshwane Jubilee
City Of Tshwane Pretoria West Mamelodi
West Rand Carltonville
Mpumalanga Nkangala Witbank
Ehlanzeni Nelspruit

Selection of PHCs for participation

Prior to initiation of the study, two community or primary health care facilities will be chosen for each participating laboratory. The sub-district and district health managers will select the clinics, based on the volume of TB suspects seen per quarter, the availability of space within the facility to accommodate two researchers, and other factors (including presence of other research projects at that facility, or other active NGO support for facilities that may interfere with the study.

Enrolment and follow-up of participants

At each PHC, TB suspects will be offered an opportunity to participate, until 120 suspects have been enrolled at each site. Informed consent will be taken according to approved protocols. Participants will be followed up for 6 months. Amongst the participants, some will develop TB disease and be started on TB treatment. These participants will be followed up for 6 months after the start of initiation of TB treatment. During the course of the study, research nurses will collect information on demographic characteristics, TB and HIV history, clinic usage, economic indicators, diagnostic tests done, and results, and outcome.

Evaluation and analysis

On completion of the study, patient outcomes will be measured on TB suspects and TB patients attending clinics being serviced by these laboratories, and will include six month mortality amongst TB suspects as the primary outcome. We will determine whether the advent of Xpert MTB/RIF alters provider behaviour with respect to investigating TB suspects, and to estimate costs from the patient’s perspective. Comprehensive economic costs (including costs to the health system) will also be measured, together with the parameters required for the modelling of population impact.

Significance of this study

These data will enable us to estimate effectiveness, cost and cost-effectiveness of implementation of Xpert MTB/RIF in the context of a national roll-out. The data generated will populate mathematical and economic models which will explore the impact of roll-out of Xpert MTB/RIF on TB control, in South Africa and elsewhere. In addition, the data will allow us to model the effect of varying test algorithms on cost-effectiveness and future resource requirements; and will guide the development of further work to test how Xpert MTB/RIF can best be used within the health system to improve patient outcomes and TB control.