Globally, tuberculosis remains a major cause of mortality, with an estimated 1.3 million deaths per annum. The Xpert MTB/RIF assay is used as the initial diagnostic test in the tuberculosis diagnostic algorithm. To extend the national tuberculosis testing programme in South Africa, mobile units fitted with the GeneXpert equipment were introduced to high-burden peri-mining communities.
This study sought to assess the cost of mobile testing compared to traditional laboratory-based testing in a peri-mining community setting.
Actual cost data for mobile and laboratory-based Xpert MTB/RIF testing from 2018 were analysed using a bottom-up ingredients-based approach to establish the annual equivalent cost and the cost per result. Historical cost data were obtained from supplier quotations and the local enterprise resource planning system. Costs were obtained in rand and reported in United States dollars (USD).
The mobile units performed 4866 tests with an overall cost per result of $49.16. Staffing accounted for 30.7% of this cost, while reagents and laboratory equipment accounted for 20.7% and 20.8%. The cost per result of traditional laboratory-based testing was $15.44 US dollars (USD). The cost for identifying a tuberculosis-positive result using mobile testing was $439.58 USD per case, compared to $164.95 USD with laboratory-based testing.
Mobile testing is substantially more expensive than traditional laboratory services but offers benefits for rapid tuberculosis case detection and same-day antiretroviral therapy initiation. Mobile tuberculosis testing should however be reserved for high-burden communities with limited access to laboratory testing where immediate intervention can benefit patient outcomes.
Globally, tuberculosis is one of the top 10 causes of mortality.
Clinically, a patient is suspected of having tuberculosis based on the following symptoms: persistent cough of 2 weeks or more, persistent cough of any duration for HIV-positive individuals, fever for over 2 weeks, night sweats, and unexplained weight loss (≥ 1.5 kg within 1 month).
Tuberculosis incidence rates globally are especially high in the mining sector. In gold mines around the world, an estimated 3000 per 100 000 population are infected.
Due to the higher burden of disease among miners, a framework to address tuberculosis in the mining sector was developed for the Southern African Development Community in 2014.
Various studies have demonstrated that mobile testing is feasible, improves access to diagnostics, and may improve linkage to care and decrease time to treatment.
There is limited local data on the cost to provide mobile Xpert MTB/RIF testing in high-burden communities. Only one local study reported that the cost to detect one tuberculosis case was $1117.00 United States dollars (USD)based on 1385 patients enrolled.
The objective of this study was to determine the cost per result and cost per positive result of mobile Xpert MTB/RIF testing and to compare it to the cost of traditional laboratory-based testing.
Ethics clearance was obtained from the University of the Witwatersrand (reference number: M160978). Our study did not contain any patient identifiers. No patient consent was required.
The National Health Laboratory Service implemented mobile testing in three high-tuberculosis-burden districts in South Africa (Kenneth Kaunda, North West, Waterberg, Limpopo, and Sekhukhune, Limpopo). Traditional laboratory-based testing was conducted at the Potchefstroom laboratory, a clinical pathology district laboratory offering a basic repertoire of testing, including tuberculosis testing, in the Kenneth Kaunda district.
The costing analysis was undertaken using Microsoft Excel (Redmond, Washington, United States).
For the calculation of staff costs, we determined the full-time equivalent hours (the number of hours worked by an employee divided by the number of hours worked by a full-time employee) based on the amount of time employees were assigned to mobile testing and multiplied this by the annual cost to company salary scales to determine the AEC. Reagent and test consumable costs were obtained from quotations received from the Oracle enterprise resource planning system used by the National Health Laboratory Service, and the AEC was determined using annual test volumes.
The costs for the initial start-up of the mobile service were determined and included the costs for the purchase of the vehicles, modifications made to the mobile units (benches, air conditioning), and purchase and placement of equipment on the mobile units. The mobile units were equipped with GeneXpert platform instruments (Cepheid, Sunnyvale, California, United States). This is an automated real-time polymerase chain reaction test for the simultaneous detection of tuberculosis and rifampicin resistance.
The test volumes and number of positive results for each mobile unit were reported using bar charts, with the total cost per result presented as a line chart on the secondary
As a comparator, the cost per result was determined for traditional laboratory-based Xpert MTB/RIF testing. Initial laboratory setup included the installation of the four GeneXpert systems (Cepheid, Sunnyvale, California, United States) (capacity of 64 samples per day), an air conditioner, a level two biosafety hood and a vortex mixer. Operational costs included costs to procure reagents, consumables, specimen collection materials, quality control materials (internal and external), printer cartridges and paper. The assumptions for these operational costs were similar to those for mobile testing.
The staff complement required to perform mobile testing included a medical technologist and a laboratory manager, who provided minimal supervision. The technologists performed other testing in addition to Xpert MTB/RIF. The costs of the business management unit (coordinator costs) in the North West province were determined and included the following personnel: business manager, secretary, quality assurance coordinator, human resources officers, training staff, and other support staff. To determine the coordinator costs per result, the AEC was divided by the annual test volume for the province. For the courier costs, the annual expenditure for the laboratory was used.
The three mobile units performed 4866 tuberculosis tests, of which the majority were performed by mobile unit 1 (68.7%). The mobile units covered a total distance of 64 605 km, with mobile units 3 and 1 contributing 73.7% of all travel. A total of 258 healthcare facilities were visited, evenly distributed between the three units. There were 544 tuberculosis-positive samples reported, with an overall tuberculosis positivity of 11.2%. The tuberculosis positivity was 9.6% for mobile unit 1, 16.6% for mobile unit 2, and 10.7% for mobile unit 3. For the period reported, 11 603 tests were done at the Potchefstroom laboratory, of which 1086 were positive (9.4%).
The overall cost per result for mobile testing was $49.16 USD with an AEC of $239 130.00 USD (
Number of tuberculosis tests performed (dark blue bars) by mobile Xpert MTB/RIF testing units in high-burden peri-mining communities in South Africa, 2018. Positive results (red bars) are reported on the primary
Comparison of cost per result between mobile Xpert MTB/RIF testing in high-burden peri-mining communities and traditional laboratory-based Xpert MTB/RIF testing offered at a laboratory in the Kenneth Kaunda district in South Africa, 2018.
Cost category | Mobile testing | Traditional testing | ||||||
---|---|---|---|---|---|---|---|---|
Cost per result (USD) | % | AEC (USD) | Cost per result (USD) | % | AEC (USD) | |||
Reagents | 10.16 | - | 20.7 | 49 461.20 | 10.16 | - | 65.8 | 117 940.47 |
Staffing: Medical technologist | 11.69 | - | 23.8 | 56 866.12 | 1.62 | - | 10.5 | 18 801.30 |
Staffing: Driver | 3.41 | - | 6.9 | 16 600.98 | 0.00 | - | 0.0 | 0.00 |
Specimen collection materials | 0.17 | - | 0.3 | 825.01 | 0.34 | - | 2.2 | 3955.03 |
Test consumables | 0.27 | - | 0.5 | 1297.16 | 1.61 | - | 10.4 | 18 637.87 |
External quality assurance | 0.10 | - | 0.2 | 472.53 | 0.02 | - | 0.1 | 201.40 |
Vehicle purchase? | 7.81 | - | 15.9 | 37 992.92 | 0.00 | - | 0.0 | 0.00 |
Vehicle operation costs | 1.76 | - | 3.6 | 8540.21 | 0.00 | - | 0.0 | 0.00 |
Laboratory equipment? | 10.24 | - | 20.8 | 49 811.96 | 1.37 | - | 8.9 | 15 873.64 |
Courier costs | 0.00 | - | 0.0 | 0.00 | 0.26 | - | 1.7 | 2993.39 |
Coordinator costs | 3.55 | - | 7.2 | 17 262.69 | 0.06 | - | 0.4 | 728.98 |
Total cost per result | 49.16 | - | 100.0 | 239 130.78 | 15.44 | 100.0 | 179 132.08 | |
Less start-up costs | 31.11 | - | - | - | - | - | - | - |
Number of tests performed | - | 4866 | - | - | - | 11 603 | - | - |
TB+ results | - | 544 | - | - | - | 1086 | 9.4 | - |
Cost per result for TB+ results | 439.58 | - | 11.2 | - | - | 164.95 | - | - |
On TB treatment | - | 300 | 55.1 | - | No data | - | - | - |
Cost per result for TB+ patient on treatment | 797.10 | - | - | - | - | - | - | - |
USD, United States dollars; TB+, Xpert MTB/RIF positive; TB, tuberculosis; AEC, annual equivalent cost.
?, Start-up costs.
The three mobile units covered distances of 21 766 km, 16 985 km and 25 854 km. The estimated overall cost per kilometre was $2.34 USD, with mobile unit 2 accounting for the highest cost per kilometre ($8.91 USD). The number of health clinics visited by the mobile units ranged from 79 to 90 clinics. The correlation between the cost per result and distance travelled was not statistically significant (
The AEC for offering mobile testing was $239 130.78 USD to produce 4866 results. There were 544 positive results (11.2%), with 300 patients documented as having received tuberculosis treatment (55.1%). The cost to find one positive tuberculosis case using mobile testing was $439.58 USD and the cost of initiating a positive patient on treatment was $797.10 USD (
The overall cost per result for laboratory-based Xpert MTB/RIF testing was $15.44 USD (
The AEC for laboratory-based testing was $179 132.08 USD to produce 11 603 results. The cost to find one positive tuberculosis case was $164.95 USD. Unfortunately, the number of patients with a laboratory test result who received tuberculosis treatment was not available.
Mobile diagnostics for high burden diseases such as tuberculosis can provide significant public health and epidemiological value in regions where individuals do not have easy access to laboratory facilities. Overall, the average cost per result for all three mobile units was $49.16 USD. However, the cost per result ranged from $30.22 USD to $154.31 USD, highlighting differences in how and where mobile testing was offered. The biggest contributors to cost differences were test volumes and distance travelled. For example, mobile unit 1 performed the most testing with short travel distances and reported the lowest cost per result. In contrast, mobile unit 3 served a very remote area with longer travel times and had the highest cost per result.
Staff, reagents, laboratory equipment and vehicle purchase contributed 88.1% of the total cost per result. This indicates that the majority of costs associated with mobile testing are not flexible, and suggests that the cost of mobile testing could only be reduced by increasing test volumes, reducing input costs or widening the test repertoire. Test volumes could be increased by identifying clinical settings with higher test volumes that would maximise the use of mobile testing. Test volumes are however limited by the daily throughput of the testing platform and space on the mobile units for multiple units of the test platforms. Negotiations with suppliers could result in lower reagent and consumable pricing. In addition, by adding mobile testing to the existing traditional laboratory national tenders, the placement agreement for reagents and analysers could be extended to mobile testing. The higher test volumes would lower the unit costs of the traditional laboratory supply chain management agreements and, by extension, benefit mobile testing. Various point-of-care platforms with a very small footprint could be used to offer additional routine haematology and chemical pathology mobile testing.
A wide range of tuberculosis positivity rates were reported for the three mobile units in this study. This highlights the importance of identifying high-burden settings with high tuberculosis prevalence for effective deployment of mobile testing. The reported cost to find a single tuberculosis-positive case would vary substantially based on the setting where testing is offered. Offering mobile testing in high-burden areas with a large population would substantially reduce the overall diagnostic cost and simultaneously offer immediate access to treatment. The higher cost of mobile testing should be weighed against the impact of earlier diagnosis, improved coverage, same-day treatment and care, as well as reduced loss to follow-up.
For mobile tuberculosis testing, scenarios should be identified that match the increased costs of mobile testing with improved patient outcomes such as rapid tuberculosis case identification and same-day antiretroviral therapy initiation. A clinical outcome study should be embedded within any future mobile testing to assess the impact on patient outcomes. Similarly, detailed cost-effectiveness studies are needed to provide evidence of how mobile tuberculosis testing can save lives and fully realise the potential of targeting high-risk groups.
This study used actual costs from the 2018 calendar year that would be more accurate than a desktop exercise. However, some staffing estimates are based on the typical number of days of mobile testing and this could have underestimated the costs. More so, the costs reported are based on the clinical referral of patients for testing. In a different clinical scenario with higher patient volumes, the costs could be very different. There are several assumptions made for this costing analysis that could have affected the reported cost per result. The number of Xpert platforms in each mobile unit, the level and type of staff employed (technologist versus technician), full-time equivalent assumptions, and the exclusion of some costs, such as overheads, would affect the reported cost per result.
This study reported that mobile tuberculosis testing is more expensive than traditional laboratory testing. However, mobile testing holds the potential to offer rapid tuberculosis case detection and improve coverage and diagnostics in communities with a high burden of disease. Furthermore, mobiles could be dovetailed to be used to deliver same-day antiretroviral therapy initiation. Further cost-effectiveness studies are needed using the patient outcome data reported.
The authors thank the staff that operated the mobile units. We also wish to thank the Global Fund for making this project possible and the Aurum Institute (clinical partner).
The authors declare that they have no financial or personal relationships that may have inappropriately influenced them in writing this article.
N.C. and L.M.C. designed the study, developed the methodology and conducted the research. N.C. conducted the costing analysis. A.L.M. provided the data required for the costing analysis. D.K.G. and W.S.S. provided editorial comments and technical input. D.K.G. supervised the study by providing leadership and oversight as the project leader. All authors reviewed the results and contributed to the manuscript development.
No funding was obtained for this study. The Global Fund to Fight AIDS, Tuberculosis and Malaria covered the cost of mobile testing in the peri-mining communities (ZAF-C-NDOH [National Department of Health]).
The authors do not have permission to share the data used for this study.
The authors declare that the views expressed in the submitted article are our own and not the official position of any institution or funder.