The study was conducted in Bauchi State located in the North-East geopolitical zone of Nigeria with a surface area of 49,259 sq. km. Bauchi State is made up of twenty Local Government Areas (LGAs) with 323 Political Wards. Bauchi city serves as both Bauchi Local Government headquarters and the State Capital respectively. According to the 2006 census, Bauchi state has a population of 4,653,066.
Notification of the Outbreak
On 28 February 2019, few hours after the index case was hospitalized at the ATBU Teaching Hospital, the Epidemiology Unit of the Bauchi State Primary Health Care Development Agency (BASPHCDA) was notified by the Bauchi LGA Disease Surveillance and Notification Officer (DSNO). The State Rapid Response Team (RRT) was mobilized and dispatched to the affected communities/Tsangaya schools including the ATBU Teaching Hospital to investigate the outbreak. The RRT utilized several methods for the outbreak investigations including advocacy visits; active case searches at the health facilities and affected communities; verbal autopsy for retrospective cases in the communities; review of case notes/patient registers at health facilities for retrospective cases; laboratory investigations and questionnaires administration. We conducted an unmatched case-control study to identify outbreak-associated risk factors.
For the case-control study, we identified 110 cholera cases from the line list of ongoing and recovered patients. Contact tracing was conducted and a semi open-ended questionnaire was administered to them. For every case patient, two controls were identified.
We defined a suspected case of cholera as “any person or patient aged 5 years or more with acute watery diarrhea with or without vomiting living in Bauchi State from 28February to 31 August 2019”. To maintain specificity, therefore, children under 5 are not included in the case definition of cholera but samples from this age group are collected where such symptoms occur and are separately line-listed.
Confirmed case was defined “as any suspected case in which Vibrio cholerae O1 or O139 has been isolated in stool at the laboratory”.
We defined control as “any person living in Bauchi State aged 5 years or more who is a friend or family member or neighbor of a case without any history of diarrhea from 28 February to 31 August 2019”.
Identification of cases and recruitment of controls
We selected all current cholera cases and those who had recovered from the line list. Using the case definitions, we conducted an active case search in the affected communities and reviewed health records at health facilities. All cases meeting the case definition were recruited. We extracted data on socio-demographic characteristics (sex, age, residence), date of onset of illness, date of presentation at the clinic, presenting signs and symptoms, history of treatment, and outcomes. Next, we generated hypotheses about possible exposure factors that were common to the cases. Controls were recruited from among family members, friends, and neighbors of the cases. Only persons who reside in the communities at least 10 days before the start of the outbreak and with no symptoms similar to the case definition during the stated period were considered eligible for selection as controls. Where more than two suitable controls were available for a case, only two were selected randomly.
A Cholera Treatment Centre at the ATBU Teaching Hospital Bauchi was set up to attend to all cholera cases in the state. This center was supported and managed by partners- Medecins San Frontiers (MSF). Referral and suspected patients in and around Bauchi LGA were brought in for management. The case identification procedures of all the health facilities in the affected wards were also reviewed.
We used interviewer-administered questionnaires to collect the following information from cases and controls; socio-demographic characteristics, clinical information from cases, and possible exposure factors. The determination of risk factors for cholera infections involved the administration of questionnaires to assess the association between general hygiene vis-à-vis cholera infections amongst individuals in the communities where the outbreak occurred. Consent was sought before administering the questionnaire to cases and controls. Data on behavioral habits collected included house characteristics, source of water, fruits and vegetables, management of waste, hygienic practices such as cleaning and disinfection procedures, types of toilet facility, eating habits, number of individuals in a household, and date of onset of the disease. The questionnaire was designed and pretested on 20 individuals in and around Bauchi LGA and the questions were adjusted as necessary. However, data from the pretesting was not included in the final analysis.
We used universal sterile bottles to collect stool samples for laboratory confirmation within 24 to 48 hours of the onset of illness. Each appropriately labeled specimen was immediately transported to the nearest microbiology laboratory within one hour of collection where they were quickly processed for culture, gram staining, and antibiotic susceptibility testing. However, treatment of dehydrated patients was initiated immediately before the laboratory confirmation.
Based on our observations and findings in the community, we assessed the following possible risk factors: Open defecation sites in about 66 communities within the metropolis; Poor environmental hygiene/sanitation as indiscriminate refuse dumpsites were seen all over the metropolis; limited access to safe water from the public water supply network; personal hygiene practices; broken pipes and leakages especially along service lines; wells as the major source of drinking water with most of them poorly protected; siting of latrines close to wells and exposed sewage as well as blocked drainages. We also inspected the homes of the cases for possible exposure factors.
The data were analyzed using univariate analysis. Categorical variables were summarized as frequencies and proportions, while quantitative variables (such as age), were expressed as mean and standard deviation. We calculated case fatality rates and developed a weekly epidemic curve to show the distribution of the cases by date (weeks) of onset of illness and date (weeks) of dead. A bivariate analysis was conducted to identify risk factors associated with a cholera infection. Any variable significant at p < 0.05 was included in the multivariate logistic regression model in a forward and backward stepwise fashion. Variables were included or excluded from the model based on the adjusted Wald test statistics, and only variables with p < 0.05 were retained. For variables that have multiple levels, one variable must be significantly different from the baseline (reference) for the variable to be retained. Possible confounders were controlled for by including these variables in the logistic regression model.
We obtained written informed consent from study participants and caregivers of children before questionnaire administration as most of our respondents were uneducated. We assured the respondents of the confidentiality of information obtained. Permission was also obtained from the management of ATBU Teaching Hospital Bauchi before record review.