Micro-costing analysis of integrating SARS-CoV-2 antigen rapid detection tests into routine health facility services: a comparison of test-all versus screen-and-test models in Cameroon and Kenya

Abstract

Introduction

Understanding costs associated with identification of emerging infections is critical to inform the policy. Nested within a randomised controlled trial that found that a test-all (TA) model for rapid SARS-CoV-2 testing identified more SARS-CoV-2 cases than the standard screen-and-test (ST) model. Our study assessed the cost-effectiveness of integrating SARS-CoV-2 services into maternal, neonatal and child health (MNCH), HIV and tuberculosis (TB) clinics using the two models in Cameroon and Kenya.

Methods

The total costs of implementing the TA and ST models in Cameroon and Kenya were estimated from a health systems perspective using a micro-costing method. The cost per client tested (CPCT) and tested positive (CPCTP) for SARS-CoV-2 were estimated by dividing the total cost of each model by the number of clients tested and tested positive, respectively. A decision tree and cost-effectiveness acceptability curve were used to compare the cost-effectiveness of the two models.

Results

In Cameroon, the total cost of the TA model was US$141,942, while the ST model was US$48,020. In the TA model, the CPCT was US$7.66 and the CPCTP was US$508.75, whereas in the ST model, they were US$25.02 and US$727.58, respectively. In the TA model, the biggest cost was SARS-CoV-2 antigen rapid detection tests (Ag-RDTs) at 61% (US$86,853), whereas in the ST model, it was personnel at 39% (US$18,592). In Kenya, the total cost was US$39,264 in the TA model and US$27,500 in the ST model. The TA CPCT was US$13.04 and the CPCTP was US$1,189.81, whereas in the ST model the costs were $125.00 and $1,250.01 respectively. In both models in Kenya, the biggest expenditure was personnel, at 45% ($17,696) of cost in TA and 56% ($15,267) in ST. In both countries, the TA model was more cost-effective.

Conclusions

Implementation of the TA model is a more cost-effective approach to increase early identification of individuals with SARS-CoV-2 infection.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Very little data exist on the costs and cost-effectiveness of the different SARS-CoV-2 testing strategies in Sub-Saharan Africa.

WHAT THIS STUDY ADDS

• We estimated the costs and cost-effectiveness of TA and ST models for integrating SARS-CoV-2 testing in MNCH, TB and HIV clinics in selected regions of Cameroon and Kenya.

• We found the TA model to be more cost-effective in both countries.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• The results shared can be used for budgeting, identification of priority areas for SARS-CoV-2 testing and expansion of the model in the continent.

• Our results suggest that using a TA model, rather than a ST model, may be more cost-effective for rapid testing during future pandemics, particularly when significant transmission occurs from asymptomatic individuals.

Introduction

As of February 2023, Cameroon had the highest number of confirmed SARS-CoV-2 cases (124,679) and deaths (1,968) in the Central Africa region.¹ Kenya had the second highest number of confirmed SARS-CoV-2 cases (342,953) and deaths (5,688) in East Africa.¹ However, these numbers of cases and deaths are likely underestimated given the limited access to SARS-CoV-2 testing. Understanding costs associated with identification of emerging infections such as SARS-CoV-2 is critical to inform testing availability and national policy.

The pandemic significantly affected the health systems of African countries and the management of major pre-existing infectious diseases such as HIV and tuberculosis (TB). A significant proportion of SARS-CoV-2 cases have been found to be asymptomatic.² A decision analytical model based on eight studies estimated that 59% of SARS-CoV-2 transmissions occur from asymptomatic cases.³ Furthermore, in a systematic review that included data from 16 studies, the prevalence of asymptomatic individuals was 48.2% from a total of 2788 confirmed SARS-CoV-2 cases.⁴

Identification of asymptomatic individuals is crucial for controlling the spread of SARS-CoV-2 in the African continent since SARS-CoV-2 can be easily transmitted by such individuals. The WHO approved the use of SARS-CoV-2 antigen rapid detection tests (Ag-RDTs) for case detection in symptomatic and asymptomatic individuals, contact testing, outbreak investigations and to monitor the trend of SARS-CoV-2 incidence at the community level.⁵ SARS-CoV-2 Ag-RDTs are much cheaper than PCR and can be provided at the point-of-care and in communities. Thus, they have the potential to identify infected individuals early, enabling prompt referral for care and initiation of treatment and quicker isolation of positive cases and their contacts.

In Cameroon and Kenya, SARS-CoV-2 testing was decentralised to primary healthcare facilities and district hospitals. Testing is usually performed with Ag-RDTs; however; testing is not integrated within key health services. Clinic attendees are screened for SARS-CoV-2 infection by inquiring about present symptoms of COVID-19 and known exposure to individuals infected with SARS-CoV-2. In Cameroon, screening occurs at various points of entry in the health facility, while in Kenya, it is mainly performed in outpatient departments. In both countries, suspected cases who have screened positive are referred to specific SARS-CoV-2 testing points within that health facility or at other health facilities. These national programmes use a screen-and-test (ST) model; however, this model does not identify those with asymptomatic infection, who also contribute to the spread of SARS-CoV-2 infection.³

As part of the Unitaid-funded Catalysing COVID-19 Action (CCA) Project, SARS-CoV-2 Ag-RDTs were integrated within maternal, neonatal and child health (MNCH), HIV and TB clinics as part of routine services in selected facilities according to national guidelines. The INTEGRATE study, a cluster randomised trial, was conducted in a subset of CCA facilities comparing the standard ST model of testing integration to a “test-all” (TA) model, where clinic attendees were offered SARS-CoV-2 testing regardless of symptoms. The TA model identified more SARS-CoV-2 cases in these high-risk populations than the ST model. There is no information available about the costs or cost-effectiveness of implementing a TA model in either country.

In this study, we determined the costs associated with implementation of a TA model of testing integration compared with a ST model in MNCH, HIV and TB clinics in Cameroon and Kenya. Furthermore, we estimated the CPCT and CPCTP for SARS-CoV-2 using Ag-RDTs and compared the cost-effectiveness of the two models in both countries.

Methods

Study design

From May 2022 to January 2023, the Elizabeth Glaser Paediatric AIDS Foundation (EGPAF) implemented the INTEGRATE Study in collaboration with the Ministries of Health (MOHs) in Kenya and Cameroon. The INTEGRATE Study was a randomised control trial conducted in 10 health facilities each in Cameroon (Central, Littoral and Western regions) and Kenya (Kiambu County). Facilities were randomised to either the TA model (testing offered regardless of screening outcome) or the ST model (testing offered if individual deemed eligible after screening).

These facilities integrated SARS-CoV-2 Ag-RDT into routine MNCH, HIV and TB services. Clinic attendees aged >2 years at HIV, TB and MNCH clinics were offered SARS-CoV-2 testing using Ag-RDTs (Panbio COVID-19 Ag test from Abbott Rapid Diagnostics, product code 41FK10; and Standard Q COVID-19 Ag Test from SD Biosensor, product code 99COV30D-EN01). We assumed that the tests had the same specificity and sensitivity.

Costing

Costs were estimated from a health systems perspective using a micro-costing method, combining top-down and bottom-up approaches to obtain resource use and costs per line item. All project costs were converted to 2022 U.S. dollars (US$) using the average exchange rate for 2022 in Cameroon and Kenya from Bank of Central African States and Central Bank of Kenya, respectively, similarly to Granich and colleagues.⁶

For each country, financial costs were divided into three groups: human resources, recurrent costs (travel, supplies and meetings) and capital (equipment and furniture, and training). Financial costs were extracted from EGPAF, health facility or MOH financial systems, as applicable. Data were collected by research assistants using an Excel cost allocation tool.

Time–motion analysis was employed to record the time that site-level personnel took to perform the activities related to SARS-CoV-2 testing, such as screening, pre-test and post-test counselling, test preparation, sample collection, testing and result documentation. The average time each cadre spent per patient was multiplied by the total number of patients to obtain the total time, which was then multiplied by the salary per hour to estimate the cost for each of these cadres.

Additional staff were hired to support the implementation of the SARS-CoV-2 testing in both countries. In Kenya, community health volunteers were hired to perform SARS-CoV-2 screening, and laboratory technicians were hired to test for SARS-CoV-2 in both ST and TA models. In Cameroon, the screening and testing for SARS-CoV-2 were performed by existing staff, and they were integrated into routine service delivery in all health facilities. However, in the TA facilities, additional testing agents were hired to meet the increased demand. For above site-level cadres, the level of effort (LOE) they dedicated to the activities related to the project (ie, supportive mentorship and supervision, monitoring and evaluation) was obtained from monthly LOE reports. We excluded any time spent on activities that were unique to donor reporting/engagement and research since these are not part of implementation of SARS-CoV-2 screening and testing. The total hours spent on these activities was converted into a percentage of the total number of hours, and this percentage was then used to calculate the cost of the LOE allocated to the project from the total salary. Salary information was obtained from the EGPAF finance department, health facility or MOH. In Cameroon, above-site personnel included monitoring and evaluation (M&E) officers, who oversaw data collection and conducted site and distance monitoring as well as data quality assessment; community engagement associate officers who were involved in demand creation and facility and community sensitisation and data clerks who collected individual forms from facilities and performed data verification and/or validation. One project officer and three testing, care and treatment technical officers, one in each region, oversaw implementation of the project at facility entry points and provided training, coaching and mentorship, and data quality assessment and verification. In Kenya, technical officers supported demand creation for SARS-CoV-2 services at facility and community levels and coordinated SARS-CoV-2 testing activities; M&E officers provided training to healthcare workers on the electronic medical record (EMR) system, mentorship on data collection using the EMR, and troubleshooting and resolving issues within the database.

Equipment (ie, tablets and laptops) and furniture (ie, chairs, folding screens and tables) were treated as capital costs annualised over a 5-year and 10-year useful life, respectively, applying a discount rate of 3% as per WHO guidelines,⁷ as previously done.^{8 9} Costs were annualised by dividing the total cost of the equipment by the annuity, as described previously by Walker and Kumaranayake.¹⁰ Trainings were also treated as capital costs and annualised over 2 years, as previously done by Vyas et al.¹¹ We applied a discount rate of 3% and calculated the annualised costs as described above. In both Kenya and Cameroon, there was on-site training on SARS-CoV-2 testing and case management. Costs included training materials (ie, stationery and printing), lunch and snack, facilitator fees and transport.

Recurrent costs were supplies, travel and meetings. Supplies costs included SARS-CoV-2 Ag-RDTs, testing consumables, airtime, non-medical supplies, stationery and printing. As part of implementation of the project, EGPAF staff travelled to attend meetings, site support supervision and mentoring activities, and stakeholder engagement meetings. Costs included fuel, taxi fare, per diem and accommodation. EGPAF covered costs for a range of meetings for sensitisation, stakeholder engagement and activity coordination related to project implementation. Costs included meeting materials (ie, stationery and printing), lunch and/or snack, facilitator fees and transport.

Cost analysis

Cost per client

The CPCT and CPCTP for SARS-CoV-2 were calculated by dividing the total annual costs of the TA or ST models by the number of clients tested and tested positive for SARS-CoV-2 in each model, respectively.

Threshold analysis

We calculated the minimum number of clients positive for SARS- CoV-2 that need to be found in ST to equal TA’s cost per client tested (CPCT) positive for SARS- CoV-2 by dividing the total cost of the ST by the CPCTP for SARS- CoV- 2 in the TA model. To calculate the maximum total cost of ST to equal the TA’s CPCTP for SARS-CoV-2, we multiplied the TA’s CPCTP for SARS-CoV-2 and the number of clients tested positive for SARS-CoV-2 in ST. Finally, the minimum number of clients that need to be tested in ST to equal TA’s CPCTP for SARS-CoV-2 (if the positivity yield remains the same) was calculated by dividing the minimum number of SARS-CoV-2-positive clients who need to be found in ST to equal TA’s cost per SARS-CoV-2-positive client by the percentage of contacts tested SARS-CoV-2 positive in ST.

Sensitivity analysis

A one-way sensitivity analysis was performed to assess the impact in variation of each cost input on the CPCT and CPCTP for SARS-CoV-2 infection. The one-way sensitivity analysis consisted of varying each cost input by applying a ±10% variation range while the others remained the same, as previously done by us and others.^{11 12} The four parameters with the biggest impact on the CPCT positive in the one-way sensitivity analysis were then used for multivariate sensitivity analysis. The lowest and the highest market price of Ag-RDT were used for the best and worst case scenarios, respectively, and we used the percentage of variation of the Ag-RDT for the other supplies. For personnel costs, for the best and worst case scenario, we used a 25% decrease or increase, respectively. Whereas, for the number of clients tested positive for SAR-CoV-2, for the best case scenario, we used a positivity yield of 31.4%¹³ and 0.8%¹⁴ for the worst case scenario.

TreeAge model

A simple decision tree was designed in TreeAge Pro Healthcare 2023, based on a model previously designed by Hussain and colleagues,¹⁵ to compare the cost-effectiveness of the TA and ST models. For the cost metric, we used the CPCT, and for the effectiveness metric, we used the number of individuals who tested positive for SARS-CoV-2, similar to others.^{15 16}

We also conducted probabilistic sensitivity analyses using Monte Carlo simulations with 10 000 iterations to explore the effects of combined uncertainties in key parameters. Gamma distributions were used for CPCTP, whereas beta distributions were used for the probability of clients testing positive for SARS-CoV-2, similar to others.^{15 16} We used a maximum willingness-to-pay threshold of three times the gross domestic product (GDP) per capita (maximum price a payer is willing to pay per individual according to WHO guidelines),^{7 16} and thus the willingness-to-pay of US$5,000 was set for Cameroon and US$6,245 for Kenya.¹⁷ We used the cost-effectiveness acceptability curve to compare the probability of the cost-effectiveness of the two models, as done previously.¹⁶

Patient and public involvement

Patients or the public were not involved in the design, conduct, reporting or dissemination plans of our research.

Results

Total cost and cost per client

In Cameroon, a total of 18 526 and 1919 clients were tested through the TA and ST models, respectively. Of these, 279 (1.5%) and 66 (3.4%) tested positive in the TA and ST models, respectively. There was little variation in SARS-CoV-2 positivity outcome between health facilities (intraclass correlation coefficient, (ICC) = 0.0021). The total costs of implementation were US$141,942 for the TA model and US$48,020 for the ST model (table 1). In the TA model, the CPCT was US$7.66 and the CPCTP for SARS-CoV-2 was US$508.75, whereas in the ST model, the CPCT was US$25.02 and the CPCTP was US$728. The biggest expenditure in the TA model was SARS-CoV-2 Ag-RDTs (61% of total cost) and personnel (39% of the total costs) in the ST model. It is worth noting that for the TA model, the cost of SARS-CoV-2 Ag-RDTs corresponded to 61% of the total cost, whereas for the ST model the tests corresponded only to 19%. The CPCT and CPCTP for SARS-CoV-2 was lower in the TA model (table 1 and online supplemental figure S1).

Table 1. Cost of providing testing serving through the TA and ST models.

Category	Sub-category	Cameroon		Kenya
		Test all (US$)	Screen and test (US$)	Test all (US$)	Screen and test (US$)
Personnel	Site level	22 404	6794	7683	5254
	Site support	12 742	11 798	10 013	10 013
	Total	35 147	18 592	17 696	15 267
Supplies	Stationery & medical consumables	7768	8459	8053	7119
	SARS-CoV-2 Ag-RDT	86 853	8997	9030	660
	Total	94 620	17 456	17 083	7779
Capital	Annualised equipment and furniture	419	303	838	801
	Annualised training	6062	6526	382	428
	Total	6481	6828	1220	1229
Meeting		2670	2088	1959	1918
Travel		3024	3056	1307	1307
Total		141 942	48 020	39 264	27 500
Number of clients tested		18 526	1919	3010	220
Number of clients tested positive for SARS-CoV-2		279	66	33	22
Cost/client tested		7.66	25.02	13.04	125.00
Cost/client tested positive		508.75	727.58	1189.82	1250.00

Ag-RDT, antigen rapid detection tests.

In Kenya, a total of 3010 and 220 clients were tested for SARS-CoV-2 through TA and ST, respectively. Of these, 33 (1.1%) and 22 (10.0%) tested positive in TA and ST models, respectively. There was little variation in SARS-CoV-2 positivity outcome between health facilities (ICC=0.0013). The total cost of providing TA services was US$39,264 and for ST was US$27,500. In the TA model, the CPCT for SARS-CoV-2 was US$13.04 and the CPCTP was US$1,189.81, whereas in the ST model, the CPCT for SARS-CoV-2 was US$125.00 and the CPCTP was US$1,250.01. In both arms, the biggest expenditure was personnel, which corresponded to 45% in the TA and 55.5% in the ST model. SARS-CoV-2 Ag-RDTs were 23% of the total cost of TA and only 2.4% of the ST. The training costs were much higher in Cameroon compared with Kenya due to strategies used. In Cameroon, each health facility had onsite COVID-19 case management training, whereas in Kenya, one large case management training hosted personnel from all health facilities involved.

Like in Cameroon, the CPCT and CPCTP for SARS-CoV-2 was lower in the TA in Kenya (table 1 and online supplemental figure S1).

Threshold analysis revealed that the number of clients tested positive for SARS-CoV-2 would have to increase from 66 to 94 (positivity yield increase from 3.4% to 4.9%) in order for ST to have the same CPCT positive for SARS-CoV-2 as TA, whereas in Kenya, it would have to increase from 22 to 23 (positivity yield increase from 10% to 10.5%). The maximum total cost of ST to TA’s CPCTP (assuming the same positivity yield) in Cameroon was US$33,578 (30% reduction), and in Kenya, it was US$26,176 (a 4.8% reduction). The minimum number of clients tested for SARS-CoV-2 in ST to equal TA’s CPCTP was 2744 (a 43% increase) in Cameroon and 231 (a 5% increase) in Kenya (online supplemental table S1).

Sensitivity analysis

In both models and countries, when cost inputs were varied by plus or minus 10%, the number of clients tested and tested positive for SARS-CoV-2 had the biggest impact (and inverse correlation) on the CPCT (figures1a,b 2a,b) and CPCTP (onlinesupplemental figures S2a,b S3a,b), respectively. In the TA model in Cameroon, the second category that caused the biggest impact was the purchase of SARS-CoV-2 Ag-RDTs, whereas in the TA model in Kenya and ST model in both countries, the cost of personnel was the second category that caused the biggest impact. At the current price point for SARS-CoV-2 Ag-RDTs, the TA model is already more cost-effective than the ST model, and given that it is expected that the price of SARS-CoV-2 Ag-RDTs will reduce significantly due to increased capacity and availability, this trend will probably continue.

Figure 1. One-way sensitivity analysis: average cost per client tested in the TA model in (a) Cameroon and (b) Kenya. Note: Two values for each input were used (±10%), the lowest in the range (green) and highest in the range (red), while the rest of the parameters remained the same. TA, test-all.

Figure 2. One-way sensitivity analysis: average cost per client tested in the ST model in a) Cameroon and b) Kenya. Note: Two values for each input were used (±10%), the lowest in the range (green) and highest in the range (red), while the rest of the parameters remained the same. ST, screen-and-test.

In the multivariate sensitivity analysis, TA had the lowest CPCTP in both the countries, in the best and worst scenarios (table 2 and online supplemental table S2).

Table 2. Multivariate sensitivity analysis of CPCTP in Cameroon and Kenya.

Scenario	Cameroon			Kenya
	Test all (CPCTP)	Screen & test (CPCTP)	Lowest CPCTP	Test all (CPCTP)	Screen & test (CPCTP)	Lowest CPCTP
Base Case	509	728	Test All	1190	1250	Test All
Best Case	11	51	Test All	27	280	Test All
Worst Case	1277	3894	Test All	2666	23 098	Test All

CPCTP, cost per client tested and tested positive.

Cost-effectiveness

The CPCT in the TA model in Cameroon was 3.3 times lower than in the ST model, whereas in Kenya, this cost was 9.6 times lower in the TA model (online supplemental table S3). In Cameroon, the probability of client testing positive for SARS-CoV-2 in the ST model was 2.3 times higher than in the TA model, whereas, in Kenya, this probability was 9.1 times higher in the ST model compared with TA (online supplemental table S3).

We applied a decision tree model to compare the cost-effectiveness of the TA and ST models in Cameroon (online supplemental figure S4a) and Kenya (online supplemental figure S4b). In both countries, the TA model was more cost-effective than the ST model.

Cost-effectiveness acceptability curves analysis showed that in Cameroon up to a willingness-to-pay per client tested was US$915, the TA model has the highest probability of being more cost-effective (figure 3a), and in Kenya, the TA model has the highest probability of being cost-effective up to US$1,249 (figure 3b).

Figure 3. Cost-effectiveness acceptability curves for ST and TA models for a range of willingness-to-pay per client tested in (a) Cameroon and (b) Kenya. ST, screen-and-test; TA, test-all.

Discussion

In both countries, we found that integrating SARS-CoV-2 Ag RDT into key clinical service delivery clinics using a TA model was more cost effective than the standard ST model, up to a relatively high willingness-to-pay per client tested was US$915 in Cameroon and US$1,249 in Kenya. In Cameroon, given that the TA model identified more clients testing positive for SARS-CoV-2 and is more cost-effective, its expansion would provide better value for money and help prevent the spread of the disease. According to guidelines in the country, individuals who test positive for SARS-CoV-2 are isolated and treated and, given that a single infected individual not isolated or treated can cause on average 3.74 secondary cases of SARS-CoV-2 in Cameroon,¹⁸ the identification of more cases could limit the spread of the disease. However, in Kenya, threshold analysis showed that only a 0.5% increase in the positivity yield of SARS-CoV-2 testing or 4.8% decrease in the total cost in ST model would lower the CPCTP to equal that of TAs, suggesting that expansion of TA in the country may not be the best strategy, although larger studies are needed to verify our hypothesis.

There are no studies in sub-Saharan Africa comparing ST and TA. A study in Mozambique, which assessed the cost of screening and testing symptomatic suspected patients for SARS-CoV-2 (strategy similar to ST), estimated that the average CPCT was US$13.1 (in 2022 US$).¹⁹ In Cameroon, the CPCT in TA and ST were US$7.66 and US$25.02, respectively, whereas in Kenya, the CPCT in TA and ST were US$13.04 and US$125.00, respectively.

The TA model in Cameroon has the lowest CPCT, and this is mainly due to differences in the number of clients tested, which has an inverse correlation with CPCT. Thus, the higher number of clients tested was the main contributing factor. Furthermore, there were differences in the methodology, for instance, the study in Mozambique did not include training cost.

Within our study, there were major differences between models and countries. The lower CPCT in TA in both countries is in line with the inverse correlation between the number of clients tested (and tested positive) and the CPCT (and CPCTP). Furthermore, the difference in personnel costs, which also have a major impact on both cost per client estimates, helps explain the differences. In both countries, site-level personnel involved in the ST model took longer to perform activities related to SARS-CoV-2 Ag-RDTs. It is worth noting that, in this study, the LOE spent by above-site personnel is a conservative estimate based on a donor-funded project, and some of that time would be reduced in an MOH implementation context.

The TA model is more cost-effective but requires a much higher investment; in Cameroon, the total cost of TA was US$93,922 more than ST, which may not be affordable to low-middle income countries. Furthermore, for the TA model to be more cost-effective, it requires a much higher testing volume than ST which may not be feasible when there is a shortage of SARS-CoV-2 Ag-RDT and other supplies. Finally, the TA model requires more personnel to deal with the higher number of clients to be tested, and given that most sub-Saharan Africa countries have a shortage of healthcare workers,²⁰ this model may not be viable to implement nationwide. So while the TA model is cost effective, governments/policies may need to make decisions about feasibility depending on the epidemiology of the disease and availability of resources.

WHO recommends integrating COVID-19 testing services with testing for other respiratory illnesses, such as influenza and respiratory syncytial virus.²¹ The estimated seroprevalence among unvaccinated persons was 9.5% in Cameroon²² and 12.7% in Kenya²³ in terms of vaccination coverage by March 2023, and only 12.93% and 26.63% of the population received at least 1 dose of COVID-19 vaccine in Cameroon and Kenya, respectively.²⁴ During COVID-19 off-peak seasons, when the prevalence is lower and there is low transmission, the positive predictive value of Ag-RDTs will be low, thus targeting people at increased risk for hospitalisation or severe COVID-19 would be a better strategy.²¹ Given that the prevalence of COVID-19 in the continent is expected to remain low in the future,²⁵ targeted testing in hospitals could be the best strategy in most contexts.

According to a recent study in South Africa, the median hospitalisation cost per patient with SARS-CoV-2 infection/COVID-19 is US$3,154 whereas the cost management of mild/asymptomatic cases at home, according to a study in Ghana, is US$317.^{26 27} Thus, early identification of individuals infected with SARS-CoV-2 would enable prompt appropriate care management and may prevent expensive hospitalisation and treatment costs.28,30

Conclusion

Overall, based on the results of this study, the TA model is the most appropriate testing strategy to identify individuals infected with SARS-CoV-2 in both the countries. As shown in the sensitivity analysis, the yield of number of clients tested positive for SARS-CoV-2 has an inverse correlation with cost per client, which implies that increasing testing yield through timing of testing with SARS-CoV-2 wave and enhancing organisational capacity would have a major positive impact on CPCT and CPCTP.

Expansion of the TA model, as done in this project, would help identify individuals infected with SARS-CoV-2 early, which has been shown to be critical for the effectiveness of the treatment and/or quarantining.³¹ However, in the context of low prevalence, there is an increased risk of false positive Ag-RDTs; thus, a more specific nucleic acid amplification testing can be used for confirmation.³² TA strategy should be implemented in areas of high prevalence and, the given low adherence to quarantine measures in sub-Saharan Africa³³ should be accompanied by community education to explain the benefits of isolation, particularly of asymptomatic individuals. Given the significant amount needed to be invested for TA strategy, countries may instead consider investing in mass vaccination, which has been shown in certain contexts and age groups to be cost-effective and to reduce the spread of the disease.³⁴

In the current study, we identified the main cost drivers of mass testing and the inputs that have the biggest impact on the CPCT. When budgeting for similar interventions, the estimation of the target population and the LOE of staff must be as accurate as possible since the number of clients tested and personnel costs have the biggest impact on the CPCT.

Limitations

Our study had some limitations. The implementation of INTEGRATE was between May 2022 and January 2023, after the peak of COVID-19, with the monthly average number of active cases in both countries below 1,000.¹ If testing took place in 2021, peak of COVID-19 transmission, the CPCTP would have been much lower in both models and countries. Furthermore, the facilities had very limited geographic spread, meaning that they are not representative of both countries. There may be differences in terms of specificity and sensitivity of Ag-RDTs used in this study that might have impacted the results.

Data availability statement

Data sharing not applicable as no datasets generated and/or analysed for this study.