Economic evaluation of gambiense human African trypanosomiasis elimination campaigns in five distinct transmission settings in the Democratic Republic of Congo

Background: Gambiense human African trypanosomiasis (gHAT) is marked for elimination of transmission (EOT) by 2030, but the disease persists in several low-income countries. We examine the cost-effectiveness of four gHAT elimination strategies in Democratic Republic of Congo (DRC), which has the highest burden of gHAT. Methods: We compared four strategies against gHAT by coupling a transmission model with a health outcomes model in five settings -- spanning low- to high-risk. Alongside passive surveillance (PS) in fixed health facilities, the strategies included active screening (AS) at average or high coverage levels, both alone or with vector control (VC). A scale-back algorithm was devised to simulate cessation of AS and VC when no cases were reported for three consecutive years. Outcomes were denominated in disability-adjusted life-years (DALYs) and costs until 2040 were denominated in 2018 US$. Results: In high or moderate-risk settings, costs of gHAT strategies are primarily driven by AS and, if used, VC. Due to the cessation of AS and VC most investments (75-80%) will be made by 2030 and VC might be cost-saving while ensuring EOT. In low-risk settings, costs are driven by PS, and minimum-cost strategies consisting of AS and PS lead to EOT by 2030 with high probability. Conclusion: In many settings, the case for EOT by 2030 is a sensible use of resources, and investments in gHAT will decelerate within this decade in moderate- and low-risk regions.

In all strategies, AS (at either 'Mean' or 'Max' coverage) and PS are in place, with supplemental VC deployed 99 in Strategies 3 and 4 (see Figure 1). In Yasa Bonga, VC has taken place since mid-2015, so only the two VC strategies 100 (3 and 4) were considered.

101
The transition between the suppression and post-elimination phases, not included in the previous projections   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 (which was not certified by peer review) The copyright holder for this preprint this version posted August 31, 2020. . https://doi.org/10.1101/2020.08.25.20181982 doi: medRxiv preprint

Costs and cost-effectiveness analysis
We performed our analysis from the perspective of health-payers collectively. Disease-related costs include confirmation, 121 staging via lumbar puncture (when indicated), and drug administration. Other intervention costs include diagnostics, 122 mobile teams, fixed facilities, and target deployment (more details can be found in section S1.4). Costs were parameterised 123 using values from the literature expressed in 2018 US$ (see Section S5). 124 We computed incremental cost-effectiveness ratios (ICERs) to account for parameter uncertainty in the economic 125 evaluation, and we adopted the NMB framework, which expresses the probability that an intervention is optimal at a 126 range of willingness-to-pay thresholds (WTP). Optimal strategies are selected based on the highest mean NMB at a 127 given WTP threshold. For further elaboration on the framework and our implementation, see section S1.5. 128 We examined health impacts and costs in a relatively long-term horizon (2020-2040), discounting at a yearly 129 rate of 3% in accordance with standard conventions [21]. We also performed scenario analyses to examine the impact 130 of our default assumptions around time horizons, discounting, and on the efficacy and cost of VC (see supplemental 131 section S1.8). The feasibility of EOT and cessation of AS are shown in Table 2. While the risk category of each health zone (Table   135 1) influences the year when EOT is expected -places with higher incidence likely meeting EOT later than places 136 with lower incidence -the implementation of VC is predicted to substantially expedite EOT across all moderate-and 137 high-risk settings. adding VC is predicted to bring forward EOT by more than two decades. Zero detections are more informative as an 148 EOT proxy when VC is in situ; if AS is stopped after three years of zero detections, there is up to a 62% probability that 149 RS would be necessary in Kwamouth in the absence of VC, but at most 19% of VC simulations result in RS in Budjala.

150
The health impact and net costs of each strategy between 2020-2040 are shown in Table S23. Yasa Bonga, 151 Boma Bungu, and Budjala are each predicted to have an average of ≤ 5 cases and ≤ 5 gHAT-related deaths over the next 152 20 years. Mosango is predicted to have more cases (≤ 23) and deaths (≤ 13) in the absence of VC. Kwamouth has the 153 most predicted cases and deaths, although the burden may be cut by three-quarters with the deployment of VC. In terms 154 of DALYs, Kwamouth sustains the worst burden even under the best strategy (1807 DALYs) compared to the worst 155 strategy in moderate-risk Mosango (443 DALYs in Mosango). Supplemental outcomes of treatment are found in section 156 S1.2.3; 97% of treated stage-1 patients and 93% of treated stage-2 patients will be cured, an additional 1% and 3%,  Pr. RAS is calculated as a proportion of the iterations when active surveillance must be when active surveillance ceases.
7 . 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 August 31, 2020.  Table 3: Summary of effects and costs 2020-2040. Two differences should be noted between these estimates and those used for decision analysis shown in table 4. First, these estimates are not discounted. Second due to asymmetric distributions, a naive difference in mean costs would not equal the mean differences in costs across simulations -the metric we used in decision analysis. Undetected cases are reflected the deaths, as very few deaths (<1 percent) originate result from treated cases. Estimates shown are means and their 95% predictive intervals (PI).

Cost-effectiveness 173
The cost-effectiveness results are displayed in Table 4 and select features are illustrated in Figure 4. Cost-effectiveness 174 acceptability curves, expressing the same information in a more conventional format, are shown in Figure S9.  Importantly, the analysis favours VC activities even if target density must be doubled at a relatively low cost of $269 and 190 $375 per DALY averted (see Figure S5).

192
The results assuming no discounting on costs or effects ( Figure S10)  . 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 August 31, 2020.  Table 4: Summary of cost-effectiveness, assuming a time horizon of 2020-2040. Cost differences and DALYs averted are relative to the comparator, which is the first strategy listed for each location. DALYs averted and cost differences are discounted at 3 percent per year in accordance with guidelines. In the uncertainty analysis (columns 5-8), the probability that a strategy is optimal is shown (as a proportion of all simulations, accounting for parameter uncertainty). Strategies highlighted in pink are optimal strategies: the strategies for which the mean net monetary benefit (NMB) is highest. ICER: Incremental Cost-effectiveness Ratio, DALY: disability adjusted life-years. For an extended discussion of these terms, see supplement section S1.5.
while healthcare costs and disability-adjusted life-years can be considered narrow criteria to justify investments in

11
. 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 August 31, 2020. . https://doi.org/10.1101/2020.08.25.20181982 doi: medRxiv preprint caution in its use, and 65% of simulated patients were assumed to be treated with NECT or on an in-patient basis due to 229 late-stage detection, low body weight, or the young age of the patient (see sections S5.5.16, S5.5.14, S5.5.15). The strategies, and they found that elimination strategies are likely cost-effective at a WTP of $400-$1500 in high-and 240 moderate-incidence places. One important difference is that we assumed yearly (rather than biennial) AS campaigns 241 at coverage levels that match historical averages (rather than a fixed value of 80% of the population, which would be 242 considered quite high). We also assumed that AS would persist for three years after the last detected case in a health 243 zone, irrespective of the initial incidence of the health zone, whereas Sutherland et al assumed no AS in low-incidence 244 health zones and therefore had to recommend VC activities to reach EOT. In our lower incidence health zones (Boma 245 Bungu and Budjala), the presence of AS meant that EOT could be reached without VC (Table-4).

246
Our analyses reinforce previous findings that VC would be both an expedited method of achieving EOT and 247 cost-effective in one moderate-incidence health zone (Mosango) and one high-incidence health zone (Kwamouth).

248
However, determining the amount of VC necessary to reach a desired tsetse population reduction is a complicated task, is substantially larger than other health zones, the geographic clustering of cases will be important to determine whether 254 all high-risk areas can be addressed in a cost-effective way (see Figure S5).

255
Our findings took into account historical improvements in PS, which was made possible by more recent data 256 and novel model calibration; this element of the strategies has difficult-to-quantify impacts that might explain some of 257 the difference between our results and previous findings. The fact that our analysis was more optimistic about current 258 practice in lower-incidence health zones underscores the potential of a well-equipped health system that can serve 259 self-presenting gHAT cases. Future analyses on the impact of integrated gHAT surveillance is warranted.

260
As this paper goes to press, DRC is contending with the emergence of COVID-19, which has triggered the 261 untimely, but hopefully temporary, suspension of AS activities [26]. Elimination strategies would protect the gains that the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. 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 August 31, 2020.

13
. 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 August 31, 2020. . https://doi.org/10.1101/2020.08.25.20181982 doi: medRxiv preprint identifying target regions for enhanced control of gambiense human African trypanosomiasis in the Democratic 327 Republic of Congo. medRxiv 2020.