Cholera Mortality during Urban Epidemic, Dar es Salaam, Tanzania, August 16, 2015–January 16, 2016

In 2015, a cholera epidemic occurred in Tanzania; most cases and deaths occurred in Dar es Salaam early in the outbreak. We evaluated cholera mortality through passive surveillance, burial permits, and interviews conducted with decedents’ caretakers. Active case finding identified 101 suspected cholera deaths. Routine surveillance had captured only 48 (48%) of all cholera deaths, and burial permit assessments captured the remainder. We interviewed caregivers of 56 decedents to assess cholera management behaviors. Of 51 decedents receiving home care, 5 (10%) used oral rehydration solution after becoming ill. Caregivers reported that 51 (93%) of 55 decedents with known time of death sought care before death; 16 (29%) of 55 delayed seeking care for >6 h. Of the 33 (59%) community decedents, 20 (61%) were said to have been discharged from a health facility before death. Appropriate and early management of cholera cases can reduce the number of cholera deaths.


Dar es Salaam
with reported cholera deaths during August 16, 2015-January 16, 2016 were included in the study.

Case Definitions
A CTC decedent was defined as a patient >2 years of age with suspected or confirmed cholera who died following admission to a hospital or CTC during August 16, 2015-January 16, 2016.
A community decedent was defined as a person with highly suspected or culture-confirmed cholera who died in the community during August 16, 2015-January 16, 2016.

Case Finding
To obtain names and wards of residence of CTC decedents, we reviewed registers and patient records at hospitals and CTCs. We also abstracted available information on the patient's time of admission, treatment, co-morbidities, and laboratory results. To construct a list of community decedents, we reviewed burial permits of persons who were either dead on arrival to referral hospitals and CTCs or were reported to municipal offices. From the permits, we obtained the decedents' age, sex, residence, date of illness onset, and date of death.

Data Collection
We worked with ward EHOs to obtain the street and administrative leaders of decedents. We requested either the EHO or administrative leader guide us to the home of the decedent.

Enumerators interviewed the spouse or caretaker about their knowledge of cholera and water
handling behaviors and about the decedent's demographic characteristics, care-seeking behaviors, treatment received (if any), and barriers to care.
The questionnaire was developed in English, translated into Kiswahili, and back-translated into English. Responses were entered onto Galaxy Tablets (Samsung, Seoul, South Korea) into an XLM form by using Open Data Kit software (https://opendatakit.org/), and transferred daily to an encrypted server. Survey results were aggregated into an Epi Info 7 database (https://www.cdc.gov/epiinfo/index.html) for analysis.

Data Analysis
Data was de-identified and analyzed descriptively by using Epi Info version 7 (Centers for Disease Control and Prevention, Atlanta, Georgia, USA). When possible, deaths occurring in CTCs were compared with those occurring in the community. Because data in this convenience sample were not normally distributed, we used the Mann Whitney U test, a nonparametric Page 3 of 3 method, to compare the 2 groups. These descriptive statistics and nonparametric tests reviewed attributes of persons dying of cholera in CTCs and in the community.

Ethical Consideration
This evaluation was part of the emergency public health response to the cholera epidemic that was ongoing in Tanzania. We limited data collection to a series of cases in Dar es Salaam, and thus, results were not generalizable. To respect confidentiality, we obtained verbal or written informed consent from all respondents. Unique identifiers were protected throughout the course of the investigation and destroyed following data analysis. We performed data analyses only with deidentified data. We obtained ethical clearance through the Muhimbili University of Health and Applied Sciences institutional review board. The US Centers for Disease Control and Prevention institutional review board had determined that this evaluation was not research because the data generated would not be generalizable and would be collected as part of the response to a public health emergency.

Investigation Limitations
This investigation had 2 main limitations. First, because data collection took place >1 month after the date of death, it was possible that accurate recall was a problem for some survey respondents. This risk was mitigated by careful administration of the survey and assistance through prompts when necessary for some questions. In addition, because recall of profound events such as sudden death is often greater than recall of more mundane occurrences, recall bias might not have had a substantial effect. Second, because this investigation was limited to case findings through burial permit data, it is possible that not all community deaths were captured. This might have been especially true earlier in the outbreak when patients who died rapidly from cholera might not have been identified as cholera patients.