|
nif:isString
|
-
The study acquired both national and the World Health Organization (WHO) ethical clearance referenced P70/02/2013 and Protocol ID B20278 respectively. All participants were of consenting age (18 years and above) and were required to give written consent before engaging in any research activity. For illiterate participants, the researcher read out the consent form details and allowed participants to consent through provision of a thumb print instead of a signature.
The study took place in Baringo County’s Central, North and Marigat sub-counties. The research team classified the study site into four zones namely the highland, midland, lowland and riverine based on altitude. The highland has an altitude of >1500m above sea level (asl), midland >1000m-1500m asl, the lowland and riverine zones at <1000m asl. (Fig 1). The riverine zone is the area on the extreme left while the lowland zone is on the extreme right of the study site map, (Fig 1). The Tugen, a sub-tribe of the Kalenjin community, mainly inhabit the highland, midland and riverine zones while the Ilchamus who are a sub-tribe of the Maa community, are found in the lowlands. Both communities practice agriculture but the Tugen engaged in both crop and livestock farming while the Ilchamus are mainly livestock keepers. The 2006–2007 RVF outbreak occurred only in the lowland zone where the Ilchamus are found.
Figure data removed from full text. Figure identifier and caption: 10.1371/journal.pntd.0005582.g001 Map of Baringo County showing the study site. Study design, sampling and data collection: The study utilized a cross-sectional research design in which a Knowledge, Attitudes and Practices (KAP) survey and focus group discussions (FGDs) were conducted sequentially. Survey respondents and focus group discussants were mutually exclusive. Qualitative data on livestock production and livestock disease management practices was collected first and comprised of 26 FGDs. Due to differences in zones, sex and community distribution, iterations resulted in 26 FGDs (13 male only and 13 female only) with a total of 231 discussants. For triangulation of male and female views, four FGDs were conducted per zone, in the highland, midland and riverine areas among the Tugen. In the lowland zone, 10 focus group discussion were conducted among the Ilchamus and 4 in a rural town that had mixed communities. Purposive sampling technique was used to select FGD discussants. To qualify as a discussant, an individual had to be 18 years old and above, have lived in the area for at least one year, be a current livestock keeper or from a livestock keeping household or consumer of livestock products with previous experience in livestock keeping. The KAP survey, whose questionnaire was informed by the FGD’s findings, targeted 560 individuals drawn from the four ecological zones. The sample size was determined through the proportion to size sampling methodology for a finite population which resulted in 383 respondents (from 20 clusters, 5 clusters per zone). A further 5% was added to cover for possible incomplete questionnaires resulting in a sample of 400 (rounded figure). Owing to a desire to increase the external validity of the survey findings, the researchers proportionately increased the sample size by 8 respondents per cluster leading to a total sample size of 560. In each zone, clusters were selected from areas with at least 30 households. Thereafter, survey respondents were identified through simple random sampling ensuring that both men and women were proportionately represented. The zoning of the study site was used to assess whether there were any differences in knowledge and practices on livestock keeping and handling of animal products. The FGD guide was pretested in 2 separate FGDs comprising exclusively of men or women to check for its suitability. Similarly, the survey questionnaire was pretested with 40 respondents in three areas which shared similar characteristics with the sampled sites. These sites were consequently excluded from the main survey. Final adjustments were made to the FGD and survey tools prior to data collection. Only Tugen and Ilchamus speaking enumerators participated in data collection.
Survey data was analyzed in SPSS version 23 (IBM SPSS Statistics, Armonk, New York) after importation from CSPRO version 6.1 (United States Census Bureau, Washington DC) where it was entered and cleaned. Besides summary statistics, independent t-tests, one way ANOVA and Chi square tests were conducted to determine the relationship between different variables. Missing values were excluded from the analyses. The measure of statistical significance for this study was set at a p-value of 0.05. A binary logistic regression was also fitted to assess the association between respondents’ level of risk to RVF infection and demographic characteristics which comprised of zone of residence, sex, age, education level, marital status, household headship type, number of children, individual scores of knowledge of RVF transmission modes and livestock (cattle, sheep and goats) quantities measured in Tropical Livestock Units (TLUs) as guided by Cholinda and Otte [31]. The overall knowledge of possible RVF transmission routes was determined through eight questions on contact with sick animals, animal tissue, secretions and consumption of products from sick animals in the KAP survey. Each question had a Likert scale type of response where those who wholly disagreed, somewhat disagreed or neither agreed nor disagreed with a possible RVF transmission route were classified as not knowledgeable and scored a zero while those who somewhat or wholly agreed were classified as being knowledgeable and awarded a score of 1. The eight answers that a respondent gave were used to generate a cumulative score on knowledge ranging from 0–8. The level of risk of exposure to the RVF virus was determined through 23 KAP survey questions with Likert scale responses ranging from never engaging in a given practice, or engaging very few times, sometimes, most of the times or always. The questions addressed community practices on: handling and consumption of milk, meat and blood; disposal of dead livestock; management of animals that abort; foetus disposal; and handling and treatment of sick livestock. For each question, respondents that carried out good practice were awarded a score of 1 while those that did not got 0. Individual outcomes were summed to give the total score per respondent. Respondents with scores below or equal to the mean were classified as high risk and above as low risk. The binary categorization of risk was used as the dependent variable in a binary logistic regression where the high risk category was coded as 0 and the low risk as 1. The model’s goodness of fit was tested using the Hosmer-Lemeshow test (χ2 = 0.617, df = 8, p = 1.000) and the omnibus-corpus test (χ2 = 180.799, df = 21, p<0.001) owing to the contested credibility of the Hosmer-Lemeshow test [32]. In each FGD, data was captured through note taking and audio recording. Audio files from the Tugen and Ilchamus were later transcribed directly verbatim into English by native speakers fluent in English and Swahili. Each script was verified through comparison of content with its recorded audio file and corresponding notes. Cleaned FGD data was coded into salient themes in Nvivo 10 (QSR international, Melbourne) and analyzed using the content analysis method. The emergent themes are presented together with the survey data in the results section.
|
![[This page as RDF]](http://data.gesis.org/somesci/static/rdf-icon.gif){kind=icon}