Setting and Study Sites
Our study was conducted in June and July 2019 in N’zérékoré, prefecture. N’zérékoré is Forest Guinea’s most populated prefecture and is also the name of the prefecture’s capital city where the largest regional referral hospital of Forest Guinea is situated . Apart from the hospital, 19 health centres are distributed in urban and rural areas to ensure a wide coverage of primary care; numerous small health posts are found in the most remote areas and villages . These are mostly unequipped facilities that provide limited services to the population . As study sites we selected the regional hospital and all 19 main health centres of the prefecture - seven in the urban region and twelve in the rural sub-prefectures. Health posts were not included in our study.
In total, we interviewed 102 HCW. Respondents comprised physicians, nurses, midwives, technical assistants, laboratory staff, pharmacists, non-medical doctors and medical and nurse students. Respondents were selected based on a convenience sample of HCW present during our visits to healthcare facilities. Health centres were visited once for interviews; interviews at the regional hospital were conducted on several days. At each of the healthcare facilities, the director or acting director gave permission to interview HCW and recommended in which order to interview HCW from different services to leave daily routine as undisturbed as possible. We approached respondents based on those recommendations.
Instruments and Scoring System
The questionnaire was developed based on existing KAP studies of EVD from Nigeria and Guinea (view supplementary material). Sections on knowledge and practices were adapted from Oladimedij et al. . Questions on knowledge were adjusted to VHF in general rather than EVD only. The scoring system, which is well-explained elsewhere, was maintained . In short, if three or more out of five questions on incubation period, infectivity, transmission, symptoms and disposal of infected corpse were answered correctly or with a sufficient number of correct answers, knowledge on VHF was scored as “good”. In addition, a list of nine diseases was read out loud to respondents and they were asked to identify the ones that were VHF. This question was not included in the knowledge score. We also added an open-ended question in which respondents were asked which VHF suspect case definition they applied in their daily clinical practice.
Questions on practices were adjusted to better fit the Guinean post-Ebola context. Furthermore, the original binary structure of possible answers was changed to a five-point Likert scale to reflect frequency of practices. Three sections - each consisting of nine questions - measured the frequency of practices in general precautions, VHF suspect case identification and VHF suspect case management. The answer “never” received one point, “rarely” two points, “sometimes” three points, “often” four points and “always” five points. An additional possibility “not applicable” was added if HCW felt the demanded practice did not match their clinical responsibility. For questions involving practices with clinical materials, a possible answer was added for “material unavailable”. Our scoring system for practices deviated from the original version. The mean for each section was calculated and a minimum of four was considered as indicating good practice. Respondents had to reach a good practice score in each section to reach overall good practices. The items where the answers “not applicable” or “material unavailable” were given were excluded from the calculation of the practice score.
The section on attitudes was adapted from Jalloh et al. . Questions in this section were adjusted to VHF and to the clinical reality of HCW as respondents. Respondents were not scored in the attitudes section and we did not classify certain attitudes as “discriminatory”.
The language of the questionnaire was French. The original versions in English underwent a forward and back translation by native speakers acquainted with the necessary terminology. In contrast to the reference works of the source questionnaires, our questionnaire was fully interviewer-administered. This was due to the simple reason that overall literacy in French is estimated to be as low as 27% in the N’zérékoré prefecture, leaving even some university graduates with difficulties to read French . Two healthcare professionals from the N’zérékoré prefecture and three graduate students in medicine and pharmacology from N’zérékoré public university were trained in the research protocol and acted as interviewers.
Data was analysed using IBM SPSS 25. Descriptive statistics were generated and proportions were compared using Pearson’s-Chi² Test and Exact Fisher Test. Statistical significance was determined at p ≤ 0.05.
The study protocol was approved by the local health authorities, the Guinean National Committee for Research in Health (opinion number 82/CNERS/19) and the Ethics Committee for Medical Research at the Ludwig-Maximilians-Universität (LMU), Munich, Germany (opinion number 18-834). Written informed consent was obtained from all respondents.