Cost of a new method of active screening for human African trypanosomiasis in the Democratic Republic of the Congo

Background Human African trypanosomiases caused by the Trypanosoma brucei gambiense parasite is a lethal disease targeted for eradication. One of the main disease control strategies is active case-finding through outreach campaigns. In 2014, a new method for active screening was developed with mini, motorcycle-based, teams. This study compares the cost of two active case-finding approaches, namely the traditional mobile teams and mini mobile teams, in the two health districts of the Democratic Republic of the Congo. Methods The financial and economic costs of both approaches were estimated from a health care provider perspective. Cost and operational data were collected for 12 months for 1 traditional team and 3 mini teams. The cost per person screened and diagnosed was calculated and univariate sensitivity analysis was conducted to identify the main cost drivers. Results During the study period in total 264,630 people were screened, and 23 HAT cases detected. The cost per person screened was lower for a mini team than for a traditional team in the study setting (US$1.86 versus US$2.08). A comparable result was found in a scenario analysis, assuming both teams would operate in a similar setting, with the cost per person screened by a mini team 15% lower than the cost per person screened by a traditional team (1.86 $ vs 2.14$). The main explanations for this lower cost are that mini teams work with fewer human resources, cheaper means of transportation and do not perform the Capillary Tube Centrifugation test or card agglutination test dilutions. Discussion Active HAT screening with mini mobile teams has a lower cost and could be a cost-effective alternative for active case-finding. Further research is needed to determine if mini mobile teams have similar or better yields than traditional mobile teams in terms of detections and cases successfully treated.


AUTHOR SUMMARY
Human African Trypanosomiasis (HAT) used to be a major public health problem in Sub-Saharan Africa, but the disease is becoming less frequent today as a result of sustained control efforts. Currently, the elimination of sleeping sickness is targeted as a public health problem by 2020 with interruption of transmission by 2030. To achieve these targets, a longterm commitment towards HAT control activities will be necessary with innovative disease control approaches accompanied by economic evaluations to assess their cost and costeffectiveness in the changing context. Today, active case finding conducted through mass outreach campaigns accounts for approximately half of all identified cases in the Democratic Republic of the Congo. However, this strategy has become less efficient, with a dwindling "yield" in terms of the number of identified cases, translating to a higher cost per diagnosed HAT case. Therefore, different approaches to outreach campaigns need to be evaluated with a focus on reaching populations at risk for HAT.
This article presents the costs and outcomes of two approaches to active screening: traditional mobile teams and mini mobile teams.
This study shows that mini mobile teams could be a cost-effective alternative for active screening with a cost-per-person screened of US$1.86 compared to US$2.08. This approach could increase the screening coverage of populations at risk for HAT that are currently not being reached through the traditional approach. Future research is needed to evaluate the difference in HAT cases identified and treated by both approaches. This would allow a costeffectiveness comparison of both strategies based on the cost-per-person diagnosed and treated.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

INTRODUCTION
Human African trypanosomiasis (HAT), or sleeping sickness, is a vector-borne disease believed to be invariably fatal when left untreated. There exist two forms of HAT, one caused by the Trypanosoma brucei (T. b.) rhodesiense and a second caused by the parasite T. b.
gambiense. Infections with T.b. gambiense are responsible for more than 95% of the globally reported HAT cases and are the focus of this study. (1) HAT is considered a public health problem because of the devastating epidemics in the 20 th century, but it is becoming more and more uncommon today thanks to sustained control efforts.
(2) Therefore, the World Health Organization's (WHO) Strategic and Technical Advisory Group for neglected tropical diseases decided to target the elimination of HAT as a public health problem by 2020 and interruption of transmission by 2030. (3) In 2018, 953 new HAT cases were declared globally, well below the targeted maximum of 2,000 cases. (4) The current method to control HAT is a combination of case-finding and treatment, and in some places, vector control as well. Case-finding is conducted either actively, through mass outreach campaigns by large mobile teams (here after called 'traditional teams') or passively in fixed health facilities. Currently, each of these strategies' accounts for approximately half of all identified cases. Active case-finding has proven to be highly effective in poor, remote HAT endemic communities with limited access to health care facilities, but this strategy is labour-intensive, costly, and time-consuming as it generates a high opportunity cost for the populations screened because of the time they have to queue waiting for the service. (5, 6) In a context of near disease elimination, this control strategy also becomes less efficient, with a dwindling "yield" in the number of identified HAT cases, translating to a higher cost per detected case due to the decreasing prevalence and declining participation rates. Additionally, mass screening campaigns are characterised by heavy logistics-operations which limit the possibilities to organise a targeted and responsive screening in high-problem areas and in remote areas that are difficult to access by car. (7) Five years ago, an alternative model for active HAT screening, called screening by "miniteams", was developed, which tries to mitigate the diminishing uptake and efficiency of traditional teams. (8) Qualitative research showed that communities prefer this type of screening because it is more adapted to their daily routine and guarantees more confidentiality, and therefore, they are also more likely to participate. (7,9) Only a few economic evaluations assess the cost and cost-effectiveness of HAT control activities, and they mainly focus on diagnostic algorithms for case detection, treatment . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 26, 2020. . https://doi.org/10.1101/2020.06.25.20139717 doi: medRxiv preprint options, and vector control. (6) If we want to achieve sustainable elimination of transmission, a longer-term commitment towards HAT control activities will be necessary, integrating improved tools and innovative disease control approaches. (10) This study aims to document the cost of two approaches to active HAT screening: traditional mobile teams and mini mobile teams, aiming to facilitate decisions on resource allocation for HAT control in different settings in the context of disease elimination.

Study area
The study was conducted in two health zones in the former Bandundu province: Mosango and Yasa Bonga (Fig. 1). Both traditional and mini mobile teams have operated in these zones since 2016. Diagnostic algorithms . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 26, 2020. . https://doi.org/10.1101/2020.06.25.20139717 doi: medRxiv preprint HAT diagnosis needs to be confirmed before a patient can be treated because of the toxicity and the complexity of the existing HAT treatment regimes. The disease is currently diagnosed through a combination of a serological test followed by more specific parasitological tests. The most frequently used serological test is the Card Agglutination Test for Trypanosomiasis (CATT), which is appropriate for mass population screening and distributed in vials of 50 tests. Once opened, the vials need to be used the same day, and specialised equipment (a rotator) requiring electricity and a cold chain for storage is needed. During the screening campaigns, the lymph nodes of all people with a positive CATT test and/or typical HAT symptoms (HAT suspects) were palpated. Upon detection of typically swollen lymph nodes, a lymph gland puncture (LGP) was performed and the fluid examined for parasites. HAT suspects without typical lymph nodes or with a negative lymph node examination were referred for microscopy tests in the following sequence: Capillary Tube Centrifugation (CTC) followed by the more sensitive Mini Anion Exchange Centrifugation Technique (mAECT). (14, 15) A HAT case was considered confirmed when one of the microscopy tests was positive (LGP, CTC, or mAECT). While traditional teams followed PNLTHA guidelines, mini mobile teams did not perform the CTC since their main energy source is 12-volt batteries charged through solar panels. No suitable 12-volt haematocrit centrifuges necessary for CTC could be found. Furthermore, 2 centrifuges would be too cumbersome on a motorcycle. (14) Disease staging, monitoring, and treatment HAT evolves in two stages: a haemo-lymphatic stage followed by a meningo-encephalitic stage when the parasite penetrates the blood-brain barrier and affects the central nervous CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 26, 2020. . https://doi.org/10.1101/2020.06.25.20139717 doi: medRxiv preprint WHO published new guidelines for the treatment of sleeping sickness following the approval of the oral medicine fexinidazole, which is used for both stages. (17) Traditional teams performed the LP on the spot and usually carry pentamidine for stage-one treatment at the nearest health centre. Stage-two patients were referred to the nearest hospital because NECT treatment requires intravenous infusions and close clinical monitoring.
Contrary to traditional teams, mini teams are not equipped for staging. Therefore, mini teams refer all confirmed HAT cases to hospitals for staging and treatment. In addition, traditional teams performed serial dilutions of CATT on HAT suspects with negative microscopy tests.
People testing positive on CATT 1/8 are considered 'serological cases', to be staged and treated for HAT, like cases detected through LGP, CTC, or mAECT. People testing negative on CATT 1/8 were referred for monitoring by the local health centre. Table 1 provides an overview of the two mobile screening strategies that were examined in this study.

Costing methodology
The study adopted the perspective of the PLNTHA and the public health care system. Data on resource consumption and prices were collected prospectively between May 2017 and April 2018 and complemented with financial records from the HAT control programme.
Costs incurred by households were excluded; research costs for activities that were relevant to the implementation of the interventions were expressed in equivalent local costs.
Costs were categorised as recurrent or capital (defined as equipment with a useful life of more than one year). Both financial and economic costs were estimated. Financial costs represent the actual quantities consumed and prices paid for consumables, including . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 26, 2020. . https://doi.org/10.1101/2020.06.25.20139717 doi: medRxiv preprint transportation costs during the study period as well as any durable equipment that was purchased specifically for these activities. Economic costs were estimated as the value of resources foregone that could have been used in other activities ("opportunity costs"). For capital equipment, the purchase or replacement value was considered and annualised based on the expected useful life and discounted at a discount rate of 3%.
For each approach, the annual cost was divided by the number of people screened for 12 months to calculate the annual cost per person screened. All costs exclude value-added taxes (VAT) since the PNLTHA and its main donors are VAT-exempt in the DRC (DRC VAT rate is 16%). (18) For items shipped to the DRC, the price was increased by 10% to account for the average shipment cost of goods between Europe and the DRC. All costs were recorded in the currency they were incurred in and converted to US$ following the average exchange rate of the study period (EUR to dollar: 1,18; CDF to dollar: 0,00065).

Scenario analysis: adjusting for differences in contexts
We examined the degree to which the different contexts in which the traditional and mobile teams operate would have an impact on the observed cost differences. Because the data are incompatible with a traditional econometric analysis to control for differences in the background epidemiological context, we performed a small simulation study of the costs incurred by each team if they were to operate in populations that are epidemiologically identical.
In our modelled scenario, we assumed that both types of teams operated in the same context have a similar percentage of serological suspects based on the results in the study area but that each team uses their regular diagnostic algorithm (Traditional team: LGP, CTC, mAECT, CATT 1/8; Mini team: LGP, CTC, mAECT). Additionally, one-way sensitivity analysis was performed to consider the specific contribution, after control for background epidemiology, of the specificity and use of serological screening tests, the prevalence of the disease, performance of the mobile teams, impact of changes in the useful life of vehicles and motorcycles (maximum and minimum according to WHO Choice guidelines in Africa), discount rate, other important cost drivers such as the fuel cost and the price of mAECT, and the use of RDT's as serological test. The diagnostic test and epidemiological parameters used for this scenario and the sensitivity analysis can be found in the supplementary information (SI_table1).

RESULTS
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 26, 2020.  Economic costs.  is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 26, 2020. Scenario analysis: costs in identical settings Table 4 shows the cost per person screened for a theoretical scenario in which both types of teams operate in an identical context: 2.14$ for a traditional team and 1.86$ for a mini team.
This cost is slightly higher for the traditional team than observed in Yasa Bonga and Mosango. In the model, the number of CATT positives is estimated based on the average number of CATT positives detected during the study (1.05%). The mini teams observed around 3 times more CATT positive tests than the traditional teams. This results in a higher number of CATT positives in the model for the traditional team and therefore a higher cost for parasitological confirmation and surveillance tests than observed in the study setting.
The overall cost per person screened by a mini team is 0.29$ or 15% lower than by a traditional team. Over 80% of this difference can be explained by the lower costs for human resources and means of transportation (motorcycles and fuel consumption) of a mini team. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 26, 2020. One-way sensitivity analysis of cost drivers During the study period, the diagnostic algorithm of a traditional team included 2 tests (CTC and CATT 1/8) that were not performed by the mini team. Excluding these tests from the traditional team's diagnostic algorithm would lower their cost per person by 0.06$ to 2.08$ per person screened.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 26, 2020. . https://doi.org/10.1101/2020.06.25.20139717 doi: medRxiv preprint The study considered the purchase price of the mAECT, but currently, the gel needed to produce this test is donated. This sensitivity analysis also looked at the impact if this gel would no longer be donated. The cost per person screened would increase and the impact depends on the specificity of the serological tests.
The remaining variables (HAT prevalence, fuel cost, useful life of vehicles and motorcycles) have a much smaller impact on the cost per person screened. Research showed that people at risk for HAT are more likely to participate in screening activities by a mini team as they are contacted in person, do not need to queue, the moment of screening can be adapted to their daily routines and their privacy is respected. (9) Additionally, mini teams could reach areas inaccessible by vehicle, and investment and fuel . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 26, 2020. . https://doi.org/10.1101/2020.06.25.20139717 doi: medRxiv preprint costs of mini teams are much lower than for a traditional team, making them more suitable to be deployed in remote areas, regions inaccessible by car or to boost active screening activities for a short period.
A disadvantage in the current set up of mini teams is that they have difficulties ensuring that all HAT suspects undergo parasitological because such confirmation usually takes place at least 1 or 2 weeks later. Traditional teams, on the other hand, perform screening and confirmation simultaneously, ensuring continuity of care. The delay between screening and confirmation for mini teams could be resolved by making the screeners and the microscopist move around together but then additional equipped microscopists might be needed which would increase the cost per person screened. The problem would also be resolved if Currently, the mini mobile teams are using HAT RDTs, therefore their cost for active screening is most likely between 25% or even 115% higher than reported in this study, depending on the RDT they are using due to the higher purchase cost of the serological tests and the lower specificity.
Overall HAT screening by mini teams could be a cost-efficient alternative for active screening if they have similar or better outcomes in terms of the detection rate and enrolment in treatment. Better accessibility to populations at risk, the sensitivity of the diagnostic algorithm, and the delay between serological and parasitological tests could affect the number of cases identified and the enrolment in treatment and therefore the effectiveness of the teams. This approach should be considered a valid alternative to the traditional way of active HAT screening, but further research is needed to evaluate the difference in HAT cases identified and treated. This would allow calculating the cost per person diagnosed and treated for both strategies and performing a cost-effectiveness comparison of both strategies.

SI_SupplementaryInformation (pdf)
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 26, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 26, 2020. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 26, 2020. . https://doi.org/10.1101/2020.06.25.20139717 doi: medRxiv preprint