Asymptomatic infection and family contact patterns in households of Ebola Virus Disease survivors, Sierra Leone 2015 – User Guide

Persistent Identifier

10.17037/DATA.41

Creators

• LSHTM: Judith R. Glynn;* Hilary Bower, Catherine Houlihan (laboratory data), Shefali Oza (Kerry town ETC data)
• Save the Children, Sierra Leone: Sembia Johnson, Cecilia Turay, Daniel Sesay, Saidu Heisenburg Mansaray, Osman Kamara, Joshua Kamara, Mohamed S. Bangura. UK: Francesco Checchi
• University of Rome: Carla Montesano (laboratory data)
• Public Health England: Steve Dicks, Dhan Samuel, Richard Tedder

*Corresponding investigator: judith.glynn@lshtm.ac.uk

Description

The data set contains information on 937 people (living and dead) who were resident in the households of people who survived Ebola Virus Disease (EVD) in the Kerry Town Ebola Treatment Centre in Western Area Province, Sierra Leone, during the 2014-2016 epidemic. It includes individual and household characteristics, information on exposure levels, symptoms experienced by individuals during the period when the household was affected by EVD, outcomes, possible routes of transmission, and ELISA results from antibody testing for Ebola IgG in oral fluid. Survivor households were chosen because they were more easily contactable through the Save the Children survivors outreach programme: 123 of 152 survivors and their households were interviewed.

Details are available in the following published papers:

• Bower H, Smout E, Bangura MS, Kamara O, Turay C, Johnson S, et al. Deaths, late deaths, and role of infecting dose in Ebola virus disease in Sierra Leone: retrospective cohort study. BMJ. May 17 2016;353
• Bower H, Johnson S, Bangura MS, Kamara AJ, Kamara O, Mansaray SH, et al. Exposure-Specific and Age-Specific Attack Rates for Ebola Virus Disease in Ebola-Affected Households, Sierra Leone. Emerg Infect Dis. 2016 Aug 15;22(8)
• Bower H, Johnson S, Bangura MS, Kamara AJ, Kamara O, Mansaray SH, et al. Effects of Mother's Illness and Breastfeeding on Risk of Ebola Virus Disease in a Cohort of Very Young Children. PLoS Negl Trop Dis. 2016;10(4):e0004622

Data collection methods

The study used qualitative interview methods to obtain quantitative data. All survivors discharged from Kerry Town Ebola Treatment Centre (ETC), Sierra Leone between November 2014 and March 2015 and living in the Western Area (or their parents/guardians) were contacted by the study team during July-August 2015 and invited to bring all household members who were present at the time Ebola was affecting their household to an interview. Households were defined as people eating from the same ‘pot’ at the time Ebola was in the household, regardless of the time they had spent in the household, and including people who joined the household during the Ebola period.

Each household was interviewed separately. Individual written informed consent was requested from all members, with parents/guardians providing consent for those under 18 years. At the interview we first drew up an inventory of all household members, noting their age, sex, and whether or not they had had or had died of Ebola virus disease. A member of the study team then demonstrated the oral swab and self-administered oral swabs were collected from all consenting participants (with parents helping young children). These were stored at -20◦C for later testing for Ebola IgG in the Immunology and Applied Biotechnology Research Laboratory of University of Makeni, Sierra Leone, in January 2016. As a group, household members were then asked to describe what had happened in the household in their own words. For each person reported by the family as having had Ebola we asked about symptoms while at home, who had helped them while they were ill, shared a bed or had other contact with them, and who had contact with their body if they died. Adults spoke for young children and corroborated information from older children. In this way, using probing questions, we assigned a maximum exposure for each household member using predefined levels, which we had developed based on the literature and discussion with staff from the Ebola Treatment Centre. (see Additional Information section below for detail of exposure measurement).

We also asked about exposures outside the home and classified them on the same scale. For those not reported to have had Ebola we asked about symptoms at the time others in the household had Ebola. EVD cases were defined as laboratory-confirmed survivors and deaths from Kerry Town or other ETCs, and all those whom the family reported as having died of Ebola. Deaths for which the family was unsure of the cause, and, for some analyses, symptomatic individuals who were not tested or diagnosed with Ebola at the time, were classified as probable EVD if they fitted the Sierra Leone case definition of probable cases. The most likely primary or co-primary cases were identified for each household, and order of illness among other cases recorded if mentioned. To avoid overburdening participants, we did not collect time sequences or dates. We also did not directly ask family members who they thought transmitted to whom to avoid stimulating blame or ill-feeling among already traumatised families.

All interviews were done in the participants’ language by multi-lingual study teams, and key outcomes were recorded in English on household inventory and individual forms. During the interview, one team member was the focal point, maintaining eye contact with the participants and guiding the discussion. A minimum of two other study team members listened and made notes, asking questions only if an issue was overlooked. This avoided multiple conversations taking place at the same time as well as ‘tick-box’ questioning. Study members were trained extensively so as not to need to use a paper interview guide to avoid loss of connection with the family group. Immediately after the family departed, the team sat together to discuss the story to make sure all elements were understood and data, including each person’s maximum level of exposure, were finalised.

Oral fluid samples were tested for Ebola virus glycoprotein IgG using an IgG Capture assay based on the EBOV Mayinga GP antigen (rGPδTM, IBT Bioservices Inc. USA cat.0501-016). Two positive controls and four negative controls were included in each plate. The cut-off for a reactive result was defined per plate as the mean optical density (OD) of the negative controls plus a fixed OD measure (0.1). Data are given as “normalised ODs”, i.e. the ratio of the test OD to the cut-off, so results >1 are reactive. Tests were repeated for positive samples and some negative samples, including those with unexpected results and samples nearer the cut-off.

Data analysis and preparation

A database template was created in EPI-data based on the study forms and data double-entered by a trained data entry clerk and the study coordinator. After validation, quality checks and double entry corrections, data were exported to STATA and tabulated to further check inconsistencies and identify missing data. These were cross-checked with the paper records. The original variables were then used to construct others where necessary for different analyses.

Geographic regions

Western Area Province, Sierra Leone

Key dates

Data capture took place from 20 July to 27 August 2015.

Quality Controls

The interview team was trained in qualitative interviewing techniques and data-recording over a 10-day period with extensive use of role play. Team supervisors were present at each interview. Questionnaires were collected and checked each evening and queries followed-up with the interviewers the following day. Data were double-entered in EPI-Data using software data checks. Initial comparison and cleaning were done in EPI-data. Additional checks were performed after transfer to STATA 14.

Population

Human population

Privacy

Each household was interviewed privately in locations away from their home, usually in school classrooms which were vacant due to school holidays and/or closures due to the epidemic. Interviewers and data entry clerks were trained in research ethics and confidentiality. Paper records were returned daily to the study coordinator and held in a locked cabinet in a locked office throughout. Electronic data were password-protected, access limited and anonymised using unique household and participant codes. Laboratory samples were identified only with anonymised codes. Code breaking sheets were only accessible to the study coordinator. Data analysis was performed on anonymised data.

Ethics

All participants gave individual, written, informed consent with parents or guardians giving consent for those aged under 18 years, and assent sought from children over 12 years. Participants were provided with contact details of the Principal Investigators and the Sierra Leone Ethics Committee in case of further questions. Ethical approval and permission to perform the study was granted by the Sierra Leone Ethics and Scientific Review Committee (19 May 2015). and the Ethics Committee of the London School of Hygiene & Tropical Medicine (Ref. No. 9866‑2).

Keywords

Ebola Virus disease, Ebola Virus Infection, Sero-prevalence, Risk factors, Exposure, Transmission Patterns, Household and Individual Characteristics, Asymptomatic infection

Language of written material

English

Additional Information

1. Exposure measurement

The table below contains the exposure levels that we developed based on the literature and discussion with staff from the Ebola Treatment Centre. We predefined these levels so that we could record only the highest level and not probe for details for possible lower levels. Our hierarchy of exposure appears to be accurate; we found strong correlations between EVD risk and each increase in exposure level.

Classification of level of exposure to Ebola Virus Disease cases

2. Variables

3. Missing data

Missing data is almost completely accounted for by persons who were absent at interview. For the 45 people who were absent at interview but still alive, basic demographics (sex, age, relationships, household characteristics) and outcome were recorded in the inventory. For 17 of the 45, information on exposure and symptoms was gleaned from the narratives of other household members. For the remaining 28 absentees information on individual exposure or symptoms is missing. Swab data is missing for 56 household members comprising 45 alive but absent, four who had died since the household had Ebola, and seven who refused. Of those who refused swabs, only two refused to continue with the interview. Six participants who did give oral swabs do not have results due to laboratory anomalies.

4.Rights and access

Data may be requested from the corresponding author, indicating the intended use.

Project information

Associated Roles

File description