Impact of a multi-pronged cholera intervention in an endemic setting

Cholera is a bacterial water-borne diarrheal disease transmitted via the fecal-oral route that causes high morbidity in sub-Saharan Africa and Asia. It is preventable with vaccination, and Water, Sanitation, and Hygiene (WASH) improvements. However, the impact of vaccination in endemic settings remains unclear. Cholera is endemic in the city of Kalemie, on the shore of Lake Tanganyika, in the Democratic Republic of Congo, where both seasonal mobility and the lake, a potential environmental reservoir, may promote transmission. Kalemie received a vaccination campaign and WASH improvements in 2013–2016. We assessed the impact of this intervention to inform future control strategies in endemic settings. We fit compartmental models considering seasonal mobility and environmentally-based transmission. We estimated the number of cases the intervention avoided, and the relative contributions of the elements promoting local cholera transmission. We estimated the intervention avoided 5,259 cases (95% credible interval: 1,576.6–11,337.8) over 118 weeks. Transmission did not rely on seasonal mobility and was primarily environmentally-driven. Removing environmental exposure or contamination could control local transmission. Repeated environmental exposure could maintain high population immunity and decrease the impact of vaccination in similar endemic areas. Addressing environmental exposure and contamination should be the primary target of interventions in such settings.


Introduction
Cholera is a bacterial, water-borne diarrheal disease.Typical cholera symptoms include vomiting and diarrhea with rice-water stools, potentially leading to severe dehydration.The mortality rate can reach 70% among severe cases if appropriate treatment, mainly rehydration, is not provided (1).However, individual symptoms can vary widely depending on age and prior exposures, ranging from asymptomatic infections, to mild infections with symptoms that are hardly distinguishable from other diarrheal diseases, to the characteristic severe watery diarrhea (2).As many as 80% of infections can be asymptomatic in endemic areas (1), resulting in underestimates of cholera burden.Cholera is transmitted through the fecal-oral route.The causal agent, Vibrio cholerae (V.cholerae), survives in aquatic environments and is present in the excreta (stools and vomit) of infected individuals.Infection is acquired by ingesting a sufficient bacterial load from the environment (indirect transmission) or contact with infectious excreta (direct transmission).V. cholerae abundance in aquatic reservoirs varies through interactions with biotic and abiotic factors.Aquatic flora, phytoplankton, and fauna, including copepods, have been described to modulate V. cholerae aquatic populations.In parallel, environmental parameters including water temperature and salinity also influence the V. cholerae life cycle in its aquatic reservoir (3)(4)(5)(6)(7)(8).Viable V. cholerae can persist in the environment in suboptimal conditions for over 15 months in a non-culturable state (7), from which it can revert to a culturable state in favorable conditions.There are over 200 serogroups of V. cholerae, but only the serogroups O1 and O139 can cause the disease cholera (9).They are often introduced in natural or manmade water reservoirs through inappropriate waste management (6,10,11).Consuming contaminated water can then trigger outbreaks, potentially leading to additional contamination of water reservoirs.In nonendemic areas, the environmental contribution to cholera transmission is often low, but in endemic areas the relative contribution of direct and indirect transmission routes is unknown.This could make the benefits we would expect from interventions, as traditionally implemented in outbreak response, less clear in such settings because they do not necessarily target the dominant transmission route.The 7th cholera pandemic has been ongoing since 1961.In addition to its traditional endemic area in the Bengal region, cholera established endemic areas in sub-Saharan Africa (SSA) following its introduction in 1970 (12).SSA now experiences the greatest annual global morbidity and mortality of cholera (13), excluding major epidemic events that occurred in Haiti and Yemen.The Global Task Force on Cholera Control (GTFCC) has set a road map to eliminate cholera in 20 endemic countries by 2030 (14), defining SSA as an important target.Generally, diseases or pathogens are considered 'endemic' when they display persistent local transmission for an extended period of time in a geographic area.For cholera specifically, the World Health Organization (WHO) defines a nation as endemic when local transmission caused cases in the previous three years.This definition encompasses a wide variety of transmission patterns.Cholera transmission can be prevented by improving water and sanitation infrastructures and with vaccination.Water, sanitation, and hygiene (WASH) improvements have historically been the primary tool to prevent cholera.WASH improvements are resource intensive and implementation requires time to prevent cholera (15).They are extremely effective, as waste management and water infrastructures have largely prevented cholera transmission in high income countries (16).Cholera cannot be eliminated from an area without WASH improvements (17), however resource scarcities limit such improvements in the countries carrying most of the global cholera burden.Sub-Saharan African nations have some of the poorest access to clean water and improved toilets in the world as well as the greatest disparities in infrastructures between rural and urban areas (18).The scale of the WASH improvements necessary to address cholera in the SSA region is vast.In comparison, implementing a vaccination campaign is fast and can reduce cholera transmission quickly.The empirical results of the reactive use of oral cholera vaccines (OCV) in 2012 in Guinea and theoretical results from modeling studies demonstrated the potential utility of vaccination as a tool for preventing cholera quickly (19,20).A quick vaccine rollout leads to a fast increase in population immunity that can mitigate cholera transmission, but this is a short term benefit because the acquired protection declines after about 2-3 years (21).The increasing stockpile of OCV allowed for more frequent use of vaccines in outbreak response and its novel use in endemic areas (22,23).
. Both OCV and WASH improvements are important components of the multisectoral interventions required to achieve cholera control or elimination in areas with high cholera burden (14,23).However, the benefit of OCV is straightforward only in epidemic settings (22,(24)(25)(26).The short-term impact of OCV correlates with the resulting increase of the population immunity, but this benefit might be narrow in an endemic setting where exposure to cholera is frequent.As such, quantifying the impact of interventions using OCV in endemic settings will provide valuable information to inform control strategies and achieve the ambitious goals set by the GTFCC.Within SSA, Cholera is endemic in the African Great Lakes region.The Democratic Republic of Congo (DRC) has consistently carried one of the highest cholera burdens in this region (13).Cholera is endemic in the city of Kalemie in Tanganyika Province in DRC; the city lies on the shore of Lake Tanganyika (Figure 1A and Figure 1B).The area displays annual peaks of cholera cases, typically during rainy seasons (Figure 1C), and reports suspected cholera cases all year.Lake Tanganyika is a potential environmental reservoir that likely promotes cholera persistence.In parallel, the local population is highly mobile with 24.7% of the residents of the Tanganyika Province reporting travelling at least once in the previous 12 months for a duration of at least 1 month (27).There is also strong fishing activity associated with the lake, with fishermen moving seasonally and experiencing exposure to the lake and low sanitation conditions (28).These population factors could promote cholera persistence through metapopulation dynamics (28).The city of Kalemie received an intervention in 2013-2016 that included both an OCV campaign and WASH improvements.The health system in DRC is organized in several levels in increasingly smaller geographical units: Provinces, health zones, and health areas.Public health interventions are often organized and implemented at health zone level.The vaccination campaign in Kalemie intended to target about 120,000 people with two doses of OCV, Shanchol TM , and focused on the four health areas where the attack rates had historically been the highest as of November 2013.The vaccination campaign was interrupted after 3 days due to security issues and resumed in July 2014 after redefining the target .population.The expiration of vaccine doses ultimately led to reducing the target to about 52,000 people in two health areas (29).The WASH intervention focused on improving access to clean water.Doctors Without Borders (Médecins Sans Frontières, MSF) extended access to tap water in the northern part of the city and laid pipes, built water reservoirs, distributed water filters, and set up public drinking fountains in collaboration with Solidarites International.In addition, sand filters were installed on paths where people draw water from the lake, and chlorination activities were performed during outbreaks.The WASH intervention incurred delays in the aftermath of the security issue, but was completed.We fit a compartmental model including both interhuman and environmental cholera transmission to assess the short-term impact of this multi-pronged intervention in Kalemie.We considered the potential influence of seasonal migration and environmental drivers and their relative contributions to local transmission.

Results
The compartmental model with interhuman and environmental transmission reproduced the reported weekly cholera cases reasonably well, although it struggled to capture the magnitude of the peaks in cases (Figure 2A).The model indicated high local immunity, fluctuating between 93.2% and 99.5% (Figure 2B).This high immunity is the likely consequence of annual outbreaks and persistent environmental exposure, which we explain further below.The targeted vaccinations occurred at times when population immunity was at 98.2% (95% credible interval (CrI): 97.6-98.7) in November 2013, 93.4% (95% CrI: 91.3-95.4) in July 2014, and 93.4% (95% CrI: 91.3-95.3)during the catch-up in August 2014.All the scenarios omitting vaccination (WASH only, and no WASH and no vaccination) visibly lacked a reduction in the susceptible proportion of the population in July 2014 (Figure 3A).Over this period of 118 weeks, we estimated an average number of cases avoided of: 2,623 (mean: 2,626.6,95% CrI: 1,910.2-3,391.6)cases avoided by vaccination alone (scenario with WASH only), 881 (mean: 881.3, 95% CrI: 94.9-1,764.1)cases avoided by WASH alone (scenario with vaccination only), and 3,484 (mean: 3,484.4,95% CrI: 2,411.7-4,832.8)cases avoided by implementing both vaccination and WASH (scenario with no vaccination and no WASH improvements) (Figure 3B).Alternative vaccination scenarios revealed that small-scale interventions, or interventions with small target population sizes, had a very limited impact in this population with high immunity (Figure 3C).However, the timing of a pulse of vaccination substantially influenced its potential impact.Specifically, timing the vaccination to occur during the lowest point of population immunity and before an outbreak began increased its impact.The best performing vaccination scenario avoided 6,858 cases (mean: 6,858.5, 95%CrI: 4,809.7-8,600.8)over 118 weeks for 200,000 vaccinated people, which is more than twice as many cases that were avoided by vaccination alone in the intervention that occurred.Despite this, the high level of local immunity resulted in vaccinating a large proportion of immune individuals, reducing the impact of the vaccination.Regarding the relative contributions of mobility and environmental reservoirs to cholera transmission, the model suggested that seasonal migration had a minimal influence on the observed cholera dynamics relative to the environmental reservoir (Figure 4A); a scenario with no seasonal migration did not substantially alter the number of cases.However, removing environmental exposure avoided 93,896 cases (mean: 93,896.5, 95% CrI: 72,842.6-122,297.5) and preventing environmental contamination from infectious human excreta avoided 91,519 cases (mean: 91,519.0,95% CrI: 68,381.1-124,373.1).Each of these interventions virtually interrupted local cholera transmission (Figure 4B).Environmental contamination with infectious excreta appeared necessary to maintain a bacterial load sufficient to support environmentally-driven transmission because the fluctuation of V. cholerae population averaged towards net decay (Figure 4D, right).The high immunity inferred by the model was maintained through annual flare-ups in Kalemie and constant environmental exposure.The environmental component of the force of infection (  =     +  (1 +   (  )) ) was consistently greater than the interhuman transmission component ( ℎ =  ℎ   ) (Figure 4C).The latter,  ℎ , remained low because epidemic flare-ups did not lead to a high prevalence of infection as they would in a mostly susceptible population.Conversely,   strongly increased with pulses of net bacterial growth due to environmental drivers, despite an overall trend favoring net decay (Figure 4D left and right).

Discussion
The impact of the intervention performed in Kalemie was modest when measured by cases avoided, which we estimate around 3,484 (mean: 3,484.4,95% CrI: 2,411.7-4,832.8)for both intervention arms combined.Following the interruption of planned vaccination activities, the reduction of the target population size after resuming activities, the narrow scale of the WASH improvements, and the high level of immunity in the population likely all contributed to mitigating the impact of the intervention.Benefitting from vaccination in endemic cholera settings, as defined by WHO, requires an understanding of dominant local transmission routes.Our model suggests that in settings where an environmental reservoir provides consistent exposure and maintains high immunity in the population, the impact of vaccination is minimal.The practical limitation of not being able to identify and target the population of susceptible individuals leads to vaccinating a majority of immune individuals.Achieving very high vaccine coverage would immunize a greater number of susceptible individuals, but at the cost of giving many additional doses to immune individuals.This cost could be reduced slightly by targeting the age group most represented among susceptibles or by guiding vaccination with serosurveys.The age profile of the suspected cholera cases residing in Kalemie (median age of 15 years, and interquartile range (IQR) . of 3-34 years) during this period would support restricting the maximum age of the target population to increase the impact of the vaccination campaign.However, applying an age cut-off would require high resolution epidemiological data.Similarly, using serosurveys to guide vaccination efforts to target susceptibles would incur a substantial additional financial cost in addition to the difficulty of applying a binary interpretation to serosurvey results.Our findings are consistent with previous analyses suggesting that allocation of OCV doses should prioritize reactive vaccination in epidemic settings over proactive vaccination in endemic settings (31).However, because endemicity is more nuanced than the current WHO definition suggests, OCV could still play an important role in some endemic settings.Ongoing studies are assessing the impact of recent vaccination campaigns performed in other endemic settings in DRC: in Uvira (32), which is also on the shore of Lake Tanganyika, and in Goma (33), which is on the shore of Lake Kivu, a Great Lake that sits just north of Lake Tanganyika .These studies will provide valuable insights on the impact of OCV with alternative strategies.Our model suggests that a well-timed large-scale vaccination could improve the impact of vaccination in Kalemie city, potentially avoiding an average of 6,858.5 cases (95%CrI: 4,809.7-8,600.8)for 200,000 vaccinated individuals.However, this requires implementing a very large vaccination campaign with precise timing.Such vaccination campaigns would be difficult to implement due to logistical challenges and high costs, and they would still achieve only short term and small-scale benefits.On the other hand, our findings suggest that WASH improvements on a scale large enough to prevent environmental exposure and contamination for the whole population could have a dramatic impact.Although we estimated that the WASH improvements in Kalemie prevented a modest number of cases, this is likely partially due to the short period of time considered to assess the impact of an intervention.The main components of this WASH intervention consisted of extending the pipe network and building a water reservoir, which were completed incrementally during that 118-week period.While extending the access to the pipe network is an important step, it does not guarantee reliable and consistent access to chlorinated tap water (34).The immense magnitude of the improvements required to ensure both access to safe water and efficient waste management, not only in Kalemie but throughout the choleraaffected nation of DRC, appears necessary to control cholera.Implementing WASH improvements should be considered a priority not only to control cholera, but also to prevent the transmission of other water-borne and fecal-oral pathogens that contribute to the disease burden in DRC (35).This approach will also help achieve the 6th goal of the Sustainable Development Goals (36), which is to ensure availability and sustainable management of water and sanitation for all, in a country where WASH improvements are critically needed (18).Kalemie is not unique regarding strong environmentally driven cholera transmission.Substantial environmental contribution for cholera cases has been reported in Haiti and Zimbabwe (20).Environmental drivers are also important drivers in other endemic settings like Bangladesh and India, where they act differently: flooding in the early and late phase of the monsoon are strongly associated with higher cholera incidence (37,38), while the peak of the monsoon is associated with a cholera lull due the "dilution" of V. cholerae in its reservoir (39).Although mobility does not play a role in local persistence within the city of Kalemie, movement allows Kalemie to be a source of cholera that can seed outbreaks in surrounding areas where exposure is less frequent.The older age of the suspected cholera cases residing outside of Kalemie (median age of 24.5 years, and IQR: 5.75-39.25 years) is consistent with lower exposure rates and source-sink dynamics.Our estimates indicating a major role of environmentally driven transmission in Kalemie's local cholera dynamics appear plausible.Confirming our estimates of population immunity and the source of bacterial infection would require serological data and substantial microbiological monitoring of the lake water in the area.Evidence of environmental presence of toxigenic V. cholerae is scarce but it has been found in Lake Tanganyika on the Tanzanian shore in fish and water samples with some evidence of increased positive samples during rainy seasons (40)(41)(42).Evidence of fecal contamination in coastal water would support the importance of the strong amplification of environmental presence through environmental contamination during outbreaks (43).The natural variation in abundance of the V. cholerae population in the lake leans in favor of net decay.Previous modeling studies assumed bacterial growth rates to consistently be in favor of net decay, whether it varied over time or it was constant negative growth (44,45).Our model allowed environmental bacterial abundance to vary over time based on environmental inputs, allowing temporary switches to net bacterial growth.It played an important role in creating pulses of a high, environmentally-driven force of infection.The overall net bacterial decay in the lake also highlights that regularly replenishing local bacterial population through environmental contamination is a critical component of local persistence.This emphasizes the potential compounded benefits of comprehensive improvements to sanitary infrastructures and access to clean water.Although our model is not designed to assess the long-term dynamics of cholera in Kalemie city, the importance of water contamination is consistent with observations in other modern outbreaks.The cholera outbreak that occurred in Haiti in 2010 following its introduction by United Nations peacekeepers was primarily transmitted along the Artibonite river in its early phase (46).Fecal contamination of water has been commonly found in Haiti (47), where access to safe drinking water and sanitation is among the lowest in the world (18).This provides conditions conducive to environmentally driven transmission for cholera.In contrast, the neighboring country on Hispaniola, the Dominican Republic, has good access to safe drinking water and strong sanitation infrastructures (18) and experienced little cholera transmission during the massive outbreak in Haiti.The impact of improving the quality of consumed water (reducing environmental exposure) in our simulation is impressive but the impact of removing environmental contamination appears almost as effective.However, this would be difficult to compare in practice since a comprehensive WASH intervention would aim to both improve the quality of consumed water and minimize the contamination of the local water reservoir.We did not consider cholera-induced mortality because of the low number of cholera-induced deaths in this population and the local experience in managing cholera.However, there is evidence that a substantial portion of cholera mortality has previously occurred in the community (48) so we cannot rule out that some cholera-induced mortality is not captured in the reported data.The absence of dedicated mortality surveys in Kalemie prevents us from estimating the number of deaths avoided by the multipronged intervention.Our model did not consider the potential immunological impact of vaccination on already immune individuals, which could extend the period during which they are protected against cholera infection.This could have led us to slightly underestimate the duration of immunity but it is unlikely to have substantial impact on our estimates considering the short study period (118 weeks, which is a little more than 2 years) compared to our estimated average immunity period (1/= 6.2 years, 95% CrI: 4.6-8.0years).We also assumed that immunity waning for susceptible individuals who were successfully vaccinated was similar to the immunity waning after a natural infection, although it is likely shorter (49).Again, this would have little impact on our estimates considering the small proportion of susceptible individuals in the population when doses were distributed in our model as well as the previously mentioned reasons (50).We included the potential impact of the WASH intervention in a simplistic way, assuming a linear variation of the environmental transmission rate due to the WASH intervention.In the absence of more detailed information, this method required the fewest additional assumptions.To estimate environmental drivers, we used measurements of chlorophyll-a levels and surface water temperature in the lake in addition to the influence of rain.The association between V. cholerae and other elements of its aquatic reservoir is vaguely understood, but the details remain unclear (7,51).We cannot assess how accurately we captured the main fluctuations of the environmental bacterial population in the absence of thorough environmental sampling in the area.However, the environmental drivers we considered have all been shown to modulate V. cholerae environmental abundance or exposure to the environmental reservoir.Phytoplankton growth, indirectly measured through chlorophyll-a, has been associated with cholera outbreaks in several studies, and cyanobacteria more specifically are a credible reservoir for V. cholerae (5,7).Water temperature influences phytoplankton growth (7), and the consequence of rainfall on environmental exposure and environmental contamination to/from V. cholerae is credible in this setting along a lake with low access to water and sanitation infrastructures (52).Finally, the complex combination of parameters to estimate critical metrics including the number of infected individuals, the net bacterial growth, and its consequence on the environmental component of the force of infection leads to some loss of identifiability.However, we reached convergence on the critical metrics they helped estimate, and our model captured true values of parameters in 95% CrI using simulated data (see Supplementary Information (SI)).We remain confident in our main conclusions.We explicitly included a direct proxy of human presence: anthropogenic nighttime radiance.Nighttime radiance is a reliable indicator of human presence and has previously been used to infer population mobility in both high and low-income countries (53)(54)(55)(56).We are confident in our finding regarding the low contribution of seasonal mobility.We also explicitly included relevant environmental drivers making the environmental component of our model robust, though it might be limited by the spatiotemporal resolution and availability of remote sensing data, particularly for chlorophyll-a.There is a scarcity of impact assessments of cholera interventions in endemic settings beyond estimates of vaccine effectiveness and vaccine coverage.Studies like this one are crucial to guide cholera elimination.OCV and WASH improvements are core components of the toolbox to control or eliminate cholera.However, the value of OCV in reactive vaccination in epidemic settings has not been clear in areas with various patterns of endemicities.The assumption that most of the target population is susceptible becomes less accurate as transmission is increasingly environmentally driven.Reducing cholera transmission in endemic areas will require a location-specific understanding of the transmission routes to tailor a strategy; a "one size fits all" approach is unlikely to achieve satisfying results.Coordinated geographically-specific strategies might also be necessary to achieve regional cholera control.

Materials and methods
We fit a Susceptible-Infected-Recovered-Susceptible model that included a compartment, B, for the V. cholerae population in the environmental reservoir, Lake Tanganyika (57).Our SIRB model explicitly considered the influence of environmental factors on bacterial growth in the lake and the potential consequences of seasonal population mobility.We fit the SIRB model to the reported suspected cholera cases presenting at the only cholera treatment center (CTC) in the city of Kalemie from November 2013 to February 2016.Detailed surveillance data were gathered in an electronic register with support from MSF.This level of resolution is available only for this period of time.Only residents of Kalemie city were included in this analysis.We estimated the number of vaccinated individuals from vaccine coverage data collected in a survey performed by MSF (29) and the associated population size estimates (see SI). Seasonal population mobility (58) can influence local cholera dynamics through regular reintroductions from areas with ongoing transmission.We included this connectivity by quantifying the seasonal variation of contemporaneous anthropogenic nighttime radiance extracted from Visible Infrared Imaging Radiometer Suite (VIIRS) data (59) (see SI).We considered the influence of water temperature and phytoplankton, such as cyanobacteria, as environmental drivers on aquatic bacterial growth (7).Water temperature and phytoplankton were included in the model through lake surface temperature (SST t ) and chlorophyll-a (chlor t ).These values were extracted from Moderate Resolution Imaging Spectroradiometer data (60) (see SI). Precipitation (rain t ) was also considered as a potential factor that could increase exposure to environmental reservoir and the environmental contamination with infectious human excreta (52) (  f (  ) ) by contaminating drinking water sources and flooding defecation sites (  f (  ) ).We extracted precipitation estimates from meteorological forcings data (61).The structure of the model is as follows: . We used a negative binomial process to link the predicted number of weekly incident cases (  ) and the weekly reported number of suspected cases (  ):   ~(  ,   ), with r, a constant reporting rate, and   , an overdispersion parameter that scaled with the predicted number of new cases.The negative binomial distribution can handle overdispersion and its scaling overdispersion parameter allows variance estimates to better scale in situations with fast and large variations of the incidence, like in epidemics.Susceptible individuals become infected through exposure to the environmental reservoir,   or through interhuman transmission,  ℎ .The WASH intervention decreased the environmental exposure rate   by a quantity   by the end of the study period.  was assumed to decrease linearly from   to   -  from the time the first WASH improvement was completed (ISO week 40 in 2014).The model did not allow the environmental contamination to vary because the intervention did not target human waste management.The infection probability from an exposure to the environment followed a doseeffect relationship, with  being the half saturation constant.Infectious individuals transitioned to the recovered compartment at rate .Susceptible individuals could gain immunity through vaccination, , 1 week after receiving the vaccine (19).This was included in the model through a step function of the number of people who received 1 or 2 doses ( 1 and  2 ) of ShancholTM.The model presented here assumed a vaccine effectiveness of 80% ( 1 ) and 85% ( 2 ) respectively for one and two dose regimens.We considered a range of alternative values for those two parameters, including estimates from studies done in the aftermath of reactive vaccination campaigns performed in Zambia and Guinea (19,62) (see SI). Considering the wide age range of the target population (everyone older than 1 year), we assumed that the proportions of susceptible, infected, and recovered among the vaccinated individuals were the same as in the general population at the time the doses were distributed.Immunity wanes at rate , returning immune individuals to the susceptible compartment.We assumed the vaccine had no effect on individuals who were immune, whether immunity was acquired through prior infection or vaccination.We also assumed that vaccination had no impact on those who were infected at the time of the vaccination.We added a penalty term (ϴ) to account for the spatially targeted nature of the vaccination campaign, which focused on areas with historically high attack rates where residents experience more cholera exposures, resulting in a low proportion of susceptibles.We considered a range of possible values for ϴ over several models (between 0.7 and 1) (see SI).This model did not include the births, deaths, and the age structure of the host population because of the short time period of 118 weeks that was considered.We also assumed no cholera-induced mortality due to case management; only 5 deaths were reported among the 1634 resident suspected cholera cases, or 0.3%, during the study period (see SI).
Population size was allowed to vary through seasonal migration (  (  ) and   (  )).We assumed that the net migration flow varied linearly with the first derivative of the nighttime radiance data in the area (  ) (53).We first fit a generalized additive model with a cyclical spline to the radiance data and then extracted its first derivative (see SI).We considered the migration of susceptible and infected individuals but not immune individuals, who do not actively contribute to transmission.We considered a range of values for the ratio of susceptible and infectious individuals among the mobile population (  / ) over several models (between 10 and 100) (see SI).We explored alternative model structures including allowing only susceptible individuals to be mobile (  (  ) = 0) as well as excluding all seasonal mobility as a sensitivity analysis (  (  ) = 0    (  ) = 0) (see SI).The bacterial population in the environment increased with contamination of the lake from the excreta of infected individuals (), and a time dependent bacterial growth rate (  ) that varied with environmental drivers.Conversely, it decreased through bacterial decay ().To minimize identifiability issues between  and the components of   we clipped the parameter space of the decay rate using previously reported values (44,63) (see SI). the proportion of susceptible, infected, and recovered among the vaccine-targeted individuals was the same as in the general population at the time the doses were distributed.For these scenarios we considered a non-spatially targeted vaccination campaign (see SI).We investigated the relative contributions of population mobility and environmental exposure in a context without intervention (=0 and   =0, meaning no vaccination and no WASH improvements).
We simulated alternative scenarios with no seasonal mobility (  (  ) =0 and   (  ) =0), or no environmental exposure (  =0), or no environmental contamination (=0) and calculated the cumulative difference between these and a scenario with seasonal mobility, environmental exposure and contamination (  (  ) ,   (  ) ,   , and  unchanged) for each of 10,000 sets of parameters sampled from the posterior distribution.

Figure 1 :
Figure 1: Overview of the location and the seasonality of cholera cases in the study area.(A) Map of the DRC with population density (with log10(population+1) transformation), the boundaries of the health zones in white, major roads in light green, a red box around the health zones of Kalemie and Nyemba (seen in B), a red circle on the city of Kalemie, and Lake Tanganyika in blue.(B) Detailed map of the city of Kalemie (red circle), which sits across the two health zones of Kalemie and Nyemba (dark red boundaries), and population density (low in grey and high in red).(C) Boxplot of the total number of reported suspected cholera cases by week (based on the International Organization for Standardization (ISO) system) in the health zones of Kalemie and Nyemba from 2002 to 2014 (30), with week number and the typical rainy season weeks shaded in grey.

Figure 2 :
Figure 2: Incident cases and model fit and variation of the percentage of infected, recovered, and susceptible over time based on the model.(A) Weekly reported suspected cholera cases residing in the city of Kalemie (empty circles) from November 2013 to February 2016 and mean prediction of the reported weekly cholera cases by the model (dark line) and its 95% credible interval (grey envelope) (B) Mean prediction of the percent of the population infected (prevalence), recovered, and susceptible (dark lines) and their 95% credible interval (grey envelopes) by the model from November 2013 to February 2016.Typical rainy seasons are shaded in grey and the timing of the distribution of vaccine doses in vertical dashed grey lines.

Figure 3 :
Figure 3: Estimated impact of the components of the intervention and impact of alternative vaccination strategies.(A)Model predictions of the mean percent of the population that was infected, recovered, and susceptible considering: the intervention as it happened of WASH and vaccination (dark blue), WASH only (blue), vaccination only (green),

Figure 4 :
Figure 4: Contributions of seasonal mobility and the environmental reservoir in cholera transmission.(A) Model predictions of the mean percent of infected, recovered, and susceptible individuals in the population with seasonal migration and environmental exposure and contamination (light blue), environmental exposure and contamination (no seasonal mobility) (pink), seasonal mobility and environmental contamination (no environmental exposure) (beige), and seasonal mobility and environmental exposure (no environmental contamination) (grey) from November 2013 to February 2016.Typical rainy seasons are shaded in grey.(B) Violin plots of numbers of cholera cases avoided by the