Risk Factors Early in the 2010 Cholera Epidemic, Haiti

Le résumé fourni par la source

During the early weeks of the cholera outbreak that began in Haiti in October 2010, we conducted a casecontrol study to identify risk factors.Drinking treated water was strongly protective against illness.Our results highlight the effectiveness of safe water in cholera control.O n October 19, 2010, the Haitian Ministry of Public Health and Population (MSPP) was notifi ed of increased cases of acute watery diarrhea resulting in death among adults in Artibonite Department.Within 2 days, MSPP's Laboratoire National de la Santé Publique had identifi ed toxigenic Vibrio cholerae O1, serotype Ogawa, biotype El Tor in stool specimens (1).The fi rst reports of illness consistent with cholera occurred on October 16, and, by November 19, cholera had reached all 10 Haitian administrative departments (2).Because the fi rst cases were in persons who worked near the Artibonite River, contaminated river water was suspected as the initial source.In a proactive effort to protect the population, MSPP rapidly implemented a cholera prevention campaign that began on October 22, 2010, to discourage the population from drinking river water, distribute water treatment products, and promote water treatment, handwashing, sanitation, and safe food preparation.To inform further prevention activities, we conducted a case-control study during the second and third weeks of the outbreak to identify risk factors for symptomatic cholera. The StudyThis study was conducted in Artibonite Department close to where the fi rst cases were identifi ed.On the basis of detailed hypothesis-generating interviews with patients and known risk factors associated with cholera in other investigations in the Americas, we created a questionnaire to assess multiple exposures, including river and other water-related exposures, sanitation and hygiene practices, foods, and other factors.We enrolled and interviewed participants from October 31 through November 13, 2010, with a 4-day break during November 5-8 because of Hurricane Tomas.To rapidly generate relevant information to guide outbreak response, we set a goal of enrolling 50 case-patients and 100 controls, a sample size that, although limited, was in line with that of previous successful emergency investigations.Eligible case-patients were persons >5 years of age who were hospitalized between October 22 and November 9 for acute watery diarrhea at the Médecins Sans Frontières cholera treatment unit in Petite Rivière, a town in a densely populated rural region near the Artibonite River.Only case-patients with the fi rst case of acute watery diarrhea in their household since October 16 were eligible.Case-patients were interviewed about exposures during the 3 days before illness onset.Within 72 hours of the interview, we visited case-patients at home, where we observed household drinking water sources and storage containers, presence of water treatment products, access to toilet facilities, and the case-patient's handwashing technique.Drinking water was tested for free chlorine as an objective measure of chlorine treatment.Matching by neighborhood (through a systematic door-to-door search from the case-patient's house) and age group (5-15, 16-30, 31-45, and >46 years), we enrolled 2 controls per case-patient at the time of the visit to case-patients' homes from households with no diarrhea since October 16.We interviewed controls about exposures during the same 3 days as the matched case-patient and made the same household observations.The term "improved drinking water source" indicated it met the World Health Organization defi nition, which describes technologies that protect water from outside contamination (3)."Lacking safe water storage" referred to water stored in an open container or bucket without a tap."Proper handwashing technique" was defi ned as observed use of soap and thorough lathering.We performed descriptive statistical analysis and exact conditional logistic regression to compute the most likely estimate or, when small cell sizes required, the median unbiased estimate of matched odds ratios (mORs) with 95% confi dence intervals (CIs).Demographic and household poverty indicators were assessed for effect modifi cation and confounding.Matched ORs adjusting for sex and the 2136

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.