Literature review

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Guidelines were used to guide the review.[6] A systematic search of grey and peer-reviewed literature was conducted using a combination of controlled vocabulary and keywords for cholera and rapid response interventions (S1 Table).

The search was limited to publications between January 2009 and November 2019; English language publications were included in both searches, in addition to French and Spanish publications in the grey literature search. The grey literature search used key word searches as well as review of various organizational websites including organizations involved in implementing and funding cholera responses and technical and coordinating bodies. In addition, key documents already known to the authors, including those identified by the READY Initiative,[7]. and forward citation tracking were used to ensure the review was as inclusive as possible.

The screening process for peer-reviewed literature was managed using Covidence software.[8] Titles and abstracts were independently screened by two reviewers; in cases of disagreement, a third reviewer was used as a tiebreaker. Eligibility of full texts were assessed by a single reviewer, and if excluded, the reason recorded. For inclusion, articles had to include primary data or be a systematic review and discuss cholera rapid response intervention(s). Grey literature documents were screened by one reviewer and classified as primary (evaluation reports, after action reviews, and case study reports) or secondary (website posts or blogs in which an organization provided a project update). Primary documents were advanced to full data extraction, while secondary documents were only used to map outbreaks.

Both the included peer and grey literature were categorized as reporting on CATI or other rapid response interventions and only CATI interventions were included in this review. Data extraction used a Microsoft Excel template with fields organized under the following categories: outbreak information, rapid response activation, team composition and equipment/supplies, activities implemented, cross-sectoral integration, coordination, data collected, performance/results, and challenges/limitations. Following initial data extraction key articles and documents were re-reviewed and additional data extracted around particular themes during results synthesis.

Case studies

Retrospective case studies were used to investigate CATI implementation in DRC (2017–2020), Haiti (2010–2019), Yemen (2016–2020), and Zimbabwe (2018–2019) where the approach was implemented to control cholera outbreaks. Locations were selected in consultation with a strategic advisory group that included members from United Nations Children’s Fund (UNICEF), World Health Organization (WHO), and United States Agency for International Development. Personal contacts, referrals, and snowball sampling were used to identify potential key informants; the DRC, Haiti, and Yemen case studies had eight to nine key informants, whilst Zimbabwe had four.

A semi-structured interview guide was developed that included outbreak background; decision-making; response actors; flow of information; interventions; coordination and integration of WASH, health, and surveillance activities; change in interventions over time; challenges; and lessons learned. The interviews lasted approximately one hour, and follow-up interviews/correspondence were undertaken as needed; participants were also requested to share relevant documentation (DRC = 8, Haiti = 6, Yemen = 0, and Zimbabwe = 1). After completing interviews, results were compiled by topic and country-level reports prepared; drafts were shared with key informants and feedback incorporated as appropriate in final versions. A non-human subject’s research determination was received from the Johns Hopkins Bloomberg School of Public Health Institutional Review Board for the case study component.

Outbreaks

A total of 15 outbreaks of varying durations in 12 countries were identified, including four from DRC,[9] two from Zimbabwe,[10] and one each from Cameroon,[11] Kenya,[12] Nigeria,[13] Sierra Leone,[14,15] Guinea,[15] South Sudan,[16] Yemen,[10,13,17–21] Haiti,[10,13,22–24] Bangladesh,[25,26] Nepal (Table 1).[27]

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Included cholera outbreak locations and reporting dates.

https://doi.org/10.1371/journal.pntd.0010042.t001

Most literature documents focused on a single outbreak with only two addressing multiple outbreaks. Within our literature search period, CATIs were first reported in the 2004 Douala, Cameroon outbreak and use continued through 2020 in the ongoing cholera outbreaks DRC and Yemen case studies.[11,30] The size of outbreaks where CATIs were employed varied widely, ranging from 169 confirmed cases in Kathmandu Valley, Nepal to more than 2.3 million suspected cases in Yemen.[27,31] Additionally, outbreak duration ranged from one month in Juba, South Sudan to over eight years in Haiti;[16] the entire duration and extent of outbreaks was not consistently reported. In Bangladesh, Cameroon, DRC (some of the outbreaks), Kenya, Sierra Leone, and Zimbabwe, cholera occurred in the absence of a humanitarian emergency, whereas in DRC, Nigeria, South Sudan, and Yemen conflict was ongoing and outbreaks in Haiti and Nepal occurred several months after the earthquakes during earthquake recovery. CATIs were implemented in urban settings in Bangladesh, Cameroon, Guinea, and Nepal, and rural settings in Kenya; in the other countries, epidemics occurred in both settings.

Interventions

CATIs typically included a combination of preventive WASH, health, and surveillance activities to interrupt cholera transmission. We observed that interventions varied across the outbreaks and temporally within the same outbreak, particularly in multi-year outbreaks (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. List of WASH, health, and surveillance interventions implemented in CATIs.

https://doi.org/10.1371/journal.pntd.0010042.t002

From the peer-reviewed literature, CATIs either encompassed WASH only (n = 6) or both health and WASH activities (n = 5); there were no CATI teams delivering health interventions alone. All WASH CATIs included water disinfection at the household or community level and/or hygiene education along with distribution of water treatment supplies. Other interventions provided in only some locations included safe water storage (Haiti and Bangladesh); soap provision (Haiti, Bangladesh, South Sudan, Kenya); and oral rehydration salts (Haiti). Health interventions delivered by CATI teams included antibiotic chemoprophylaxis for case households in Haiti (doxycycline for non-pregnant adults, though delivery was inconsistent) and for both case and adjacent households in Cameroon (doxycycline for the general population and amoxicillin for children and pregnant/lactating women). Oral cholera vaccine (OCV) was delivered as part of CATIs in South Sudan, where there was surplus stock from a previous vaccination campaign. Grey literature contained relatively little information about specific interventions. All documents reported on WASH interventions, including household disinfection, hygiene education, water chlorination, water quality testing, sanitation assessment, and waste management. Health interventions were described in four documents, including antibiotic chemoprophylaxis distribution in Haiti, health education in Sierra Leone and Yemen, health counseling and referrals in Yemen, and case investigation. Variation in CATI implementation was also observed in the four case study countries where WASH interventions were reported in all locations and health and surveillance interventions were reported in DRC, Haiti, and Yemen but not Zimbabwe.

Team composition

CATI team composition varied across outbreaks and within the same outbreak in Haiti and Yemen. The teams were comprised of staff from NGOs, governments, and community volunteers. Teams were most commonly comprised of four to seven members with a mixed set of skills, including WASH, health/nursing, hygiene promotion, surveillance, and drivers. Teams with only WASH staff were reported in case study interviews in Yemen, Zimbabwe, and during the early response in Haiti;[23] health CATI teams were reported in Yemen.[30] The skill sets included depended upon the type of interventions delivered. For instance, a nurse accompanied CATI teams in Haiti to administer antibiotics and vaccinators were CATI members in South Sudan to provide OCV. Additionally, the Haiti case study showed that before establishing the mixed teams, comprised of health and WASH staff, the NGO WASH teams would often include a nurse to accompany them.

Implementation

CATI implementation was a multi-faceted process involving numerous steps and multiple organizations typically beginning with an alert of a suspected/confirmed cholera case(s) (Fig 2). In Nepal,[27] Bangladesh,[26] and South Sudan,[16] CATIs were implemented in response to rapid diagnostic test (RDT) or culture confirmed positive cases; both were used at different times in Yemen. [17]However, in Cameroon,[11] DRC,[10] and Zimbabwe,[32] CATIs were implemented for suspected cases that presented for care at a cholera treatment center. RDT was used in Haiti during 2013 and discontinued afterward because of delays to obtain results and reliability concerns. Specific CATI activation alert systems were not reported in Nigeria, Kenya, and Sierra Leone. The common source of information to generate an alert was the shared line list from the health treatment facility.[13,20–23] Additionally, suspected cases reported by community health workers, media reports, and other informal surveillance mechanisms also triggered alerts in DRC,[9] Haiti, and Yemen.[13] The time between the identification of a suspected case and the trigger of an alert was critical for rapid implementation of CATIs and community level alert systems were used to prioritize response.[23] Delays in sharing case information with CATI teams were frequently reported.[13,20–23] For instance, in Yemen, cholera case line lists had to be approved by central authorities before being shared[13,17] and in Haiti, case information was shared only on a weekly basis during the initial response which delayed the CATI activation. In Nepal, delays occurred as CATIs were implemented only after receiving a positive stool culture, which took an average of 3.9 days following hospital admission.[27] In Bangladesh, DRC, and Haiti, after establishing daily line lists access, and in South Sudan and Zimbabwe, the CATI teams were able to respond more quickly than in the aforementioned locations[16,22,25,26] as they had rapid access to the case information. In instances when the number of suspect/confirmed cases exceeded the capacity of the CATI teams, a set of prioritization criteria were used, including number of cases from one location (Haiti, Zimbabwe), areas without recent cases (Haiti, Zimbabwe), if a case was RDT/culture positive (Yemen), and case severity status/death (Haiti). [13,22,23]

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Common steps in the CATI implementation process.

https://doi.org/10.1371/journal.pntd.0010042.g002

When CATI teams had difficultly locating case households, assistance was often sought from community members. The selection process and the number of surrounding households to be included in the CATI varied greatly across outbreaks, and depended on the density and arrangement of houses, implementation strategy, capacity, and resource availability. For instance, in Bangladesh and South Sudan, only household contacts received the interventions whereas the early WASH CATIs in Yemen included about 100 households and in the later Yemen response and Haiti a 50–100 meter radius for household selection was reported;[13] in the Zimbabwe case study, the entire apartment level was included when responding in urban settings.[30] As mentioned earlier, a combination of preventive WASH, health, and surveillance activities are used to interrupt cholera transmission, with the same or different interventions provided to case and neighboring households. For example, in Haiti antibiotic chemoprophylaxis was only administered to household contacts;[22] and latrine disinfection was conducted only at case households in Haiti, Zimbabwe, and in Cameroon.[11] In addition to case- and neighbor-targeted household-specific interventions, CATIs in DRC,[9] Sierra Leone and Guinea,[15] Yemen,[18] and Zimbabwe (case study) included community-level activities such as water point repair, WASH risk factor assessment, testing of communal water sources, construction of communal sanitation facilities, and awareness raising.

Implementation challenges

Several common difficulties were identified including: 1) implementation in logistically challenging contexts[13,17,23,27] and poor access to targeted populations;[9,12,16] 2) cholera cases exceeding the limited capacity to implement all CATIs;[10,13,22,27] 3) coordination among actors and integration of WASH, health, and surveillance activities;[10,13,15,19] 4) timely access to case information to rapidly implement CATI;[10,23] and 5) limited funding to adequately implement the interventions.[9] The grey literature review and all four case studies reported coordination between WASH and health actors as hard to achieve, at times resulting in duplication of activities by the WASH and health CATI teams. Aspects such as access to case data as well as coordination among partners gradually improved over time, which enhanced response timeliness in Haiti[23] as well as in Yemen.[30]

Evaluation measures

Multiple measures were used to evaluate different aspects of CATI performance including: 1) timeliness (e.g. time between case presentation/confirmation and CATI start or time between outbreak declaration/index case confirmation and start of the first CATI); 2) coverage; 3) cost of the intervention; 4) household’s knowledge and practice of measures aiming to interrupt cholera transmission; 5) outputs (items distributed during intervention); and 6) impact (incidence of cases). While the Bangladesh and Haiti studies reported about the complete CATI approach, none of the documents reviewed evaluated the effect of individual interventions (e.g. only chlorine tablet distribution or only awareness session) in reducing the number of new cases or interrupting the cholera transmission pathways.

Timeliness of intervention delivery was reported in four peer-reviewed[16,22,23,27] and two grey literature documents.[10,13] The elapsed time between case admission and the start of CATIs varied across outbreaks, and usually CATI teams based at cholera treatment facilities were able to initiate interventions faster than separate organizations not based in the facilities. In Cameroon,[11] DRC,[33] Haiti,[24] and Bangladesh[26] the interventions started at case admission or discharge where antibiotic prophylaxis and hygiene kits were distributed to the case’s household. In South Sudan, Nepal, and Zimbabwe, interventions at the case household began between one and four days after case identification. In Nepal, delivery started on average 1.7 days after culture result and 3.9 days after admission,[27] and in Zimbabwe, the intervention started within 48 hours of the presentation of a suspected case at the treatment center,[9] and in South Sudan, CATIs occurred 1–6 days after the suspected case was presented at the health facility with an average delay of 3.4 days.[16]

Intervention coverage was reported in 10 peer-reviewed articles and four grey literature reports, though different indicators were used. For example, 51% cases in South Sudan[16] and 92% in DRC received interventions;[9] whereas 17% and 30% of households and communities, respectively, received interventions in Nepal.[27] In Haiti, coverage was reported in terms of responded cholera alerts (49%),[23] identified outbreaks responded (53% in Centre department)[22] and cases admitted to cholera treatment centers in Carrefour Haiti (65%).[24] The two Bangladesh research studies reported 100% coverage of the admitted cases, which is not typical in most outbreak responses. One peer-reviewed article and two grey literature documents reported CATI cost information using different metrics. In Sierra Leone the per capita cost was US$2.32 (and was compared with a previous outbreak response in Zimbabwe where the cost was US$2.85)[14] and in Bangladesh the cost was US$45.5 per household.[26] UNICEF’s global review of cholera rapid response teams reported monthly per team costs of $10,234 in Haiti and $2,400 (urban) to $3,000 (rural) in Yemen.[13]

A range of different effectiveness measures were reported in peer-reviewed articles. Behavior change measures, such as knowledge of cholera prevention, treatment, and hygiene practices;[12,25] and use of appropriate hygiene practices and/or items distributed in hygiene kits[26,27] were commonly reported. Only the Bangladesh study and the Haiti Central department study reported evaluated differences in cholera incidence between the households that received a timely CATI and those that did not.[22,26] For instance, in Kenya, a significant difference in respondent-reported water treatment as compared to controls (56% vs 37%, p<0.001) was reported, however, cholera knowledge (7/7 indicators) and behaviors (5/6 indicators) were similar between the two groups.[12] In Bangladesh, as compared to a control group, the intervention group had significantly higher odds of handwashing (odds ratio = 14.7, CI: 8.3–25.9); was more likely to have soap present in cooking areas (98% vs. 15%, p<0.001) and latrine areas (98% vs 13%, p< 0.0001); stored drinking water was more likely to have adequate concentrations of free chlorine (94% vs 1%, p<0.0001) and less likely to be contaminated with V. Cholerae (0% vs 6%, p = 0.06); and there was a 47% lower incidence of V. cholerae infection (symptomatic and asymptomatic) in intervention group case contacts during the intervention period.[25,26] In Nepal, 30.2% of surveyed households reported hearing awareness messaging via community campaigns and 16.5% reported home visit with various types of health education messages and supplies. Individuals that received home visits were more likely to have heard of cholera (adjusted odds ratio = 2.38, p<0.05).[27] In Haiti, estimates of the effectiveness of national CATIs indicated that timely response (≤1 day after outbreak identification compared to a delayed response of >7 days) reduced new cases by 76% (CI: 59–86%) and outbreak duration by 61% (CI: 41–75%) whereas an intense response (≥1 complete CATI per week compared to a weaker response of ≤0.25 complete CATIs weekly) reduced new cases by 59% (CI: 11–81%) and outbreak duration by 73% (CI: 49–86%).[22]