Study design and sampling
This study was a descriptive community based cross-sectional type conducted in Northern Red Sea region. Four subregions namely Afabet, Foro, Gindae and Shieb were selected for the study in the region. These four sub-regions were selected considering they are more densely populated with livestock species than the remained six sub regions. Selection of study villages in these subregions was done by simple random sampling technique, after developing a list of all villages keeping livestock species. Finally, participant households were selected by systematic random sampling technique from each administrative area and villages. For each selected house hold, the head was responsible to be interviewed with the questionnaire and give blood sample for brucellosis testing.
Sample size calculation
There was no reliable published literature which indicates the prevalence level of brucellosis in the study area. To obtain the optimum possible sample size for the study, various related standard inputs and sample size adjustment factors were considered as follows: First, an initial sample size was calculated using Epitools online software (https://epitools.ausvet.com.au/prevalencess) assuming imperfect sensitivity and specificity for Rose Bengal plate test (RBPT) and competitive enzyme linked immunosorbent assay (c–ELISA) tests. In view this, the initial sample size was calculated considering 30% expected prevalence, 95% confidence level, 0.05 desired precision, 88% combined sensitivity and 99% specificity for RBPT and c-ELISA. The combined sensitive and specificity was calculated according to author [9] and the values for RBPT and c-ELISA was obtained from the literature [10]. With this inputs, the epitool software generated an initial sample size of 402 (n1). Considering 1.5 between cluster variance the size increased to 603 (n2) and assuming a non-respondent rate of 0.05 the total sample size (Ts) was enlarged to 634. Secondly a two-stage cluster sampling design was followed as describe by the author [11]. Thus, number of villages to be sampled were calculated using “fixed sample size and unknown clusters” for large populations. Then the sample size was further adjusted using finite population adjustments formula as follows:
g =1.962TsVc
d2Ts –1.962Pexp (1-Pexp)
Where:
g = number of clusters (villages) to be sampled, Pexp = 30%, d = 0.05, Ts = 634, Vc = 0.04 (Thrusfield, 2005). Using the above formula and inputs, 124 villages were estimated to be sampled in case of large/infinite population size. However, considering the existing small number of eligible villages in the study area, the number was further adjusted as follows:
g-adj.= Gxg , 124 x 38 = 29 villages
G+g 124+38
Finally, the number of livestock owners to be sampled from each selected village (n) was calculated as: n=Ts/g-adj, 634/29 = 22. Thus, a total of 638 (22*29) livestock owners were sampled for blood sample collection and be interviewed using structured questionnaires. Nine, seven, eight and five villages were selected randomly from Afabet, Foro, Gindae and Shieb sub- regions respectively.
Study population and setting
Northern Red Sea region has an area size of 34,236 Sq. Km. It has 10 subregions divided into 104 Administrative areas with a total of 289 villages and a total population of 491,657 [12]. Foro sub-region has a total population of 44,444 and 14 administrative area [12]. Gindae sub-region has a population of 58,000 with 5,644 households and Afabet has a population of 91,813 with 14 administrative area and 56 villages. Shieb subregion has a population of 42,600 and 8 administrative area with 15 villages [12]. For this study, a total of 637 individuals were selected and blood samples and questionnaires were collected from each participant.
Inclusion and exclusion criteria
Farmers who live in the selected community and can properly respond to the questionnaire were included in the study. Farmers aged less than 18 years were excluded from the study. Inhabitants who live in the cities of Foro, Afabet, Shieb and Gindae who live away from farming and grazing animal were excluded from the study, as they are expected to have less contact with animals, their environment and products.
Data collection
Data and blood collection was conducted from October 22, 2020 to November 10, 2020. Blood was collected by selected and trained laboratory staffs in 10 ml vacutainer tubes from the selected livestock owners. Then serum was separated and stored in refrigerators (0°C – 4°C) and transported in ice boxes covered with cool ice packs to the National Animal and Plant Health Laboratory (NAPHL), Ministry of Agriculture, Asmara for analysis.
Interviewer administered face to face questionnaire was used to evaluate the knowledge and practice of the community on brucellosis. The questionnaire was pretested in the field. The data collection tool was partly adapted from a study by Elisabeth Lindahl et al 2015[13], with slight modification and adjustment to fit the objectives and context of our study. Finally it was translated to local language `Tigrigna` for the convenience of the data collectors and retranslated back to English to retain its originality. Data collectors were selected based on their linguistic ability as most of the population in the study sub-regions are Tigre, Saho and Afar ethnic groups. Pilot study was done in neighboring village in order to validate the data collection tool and made necessary adjustment before starting the research in the selected villages.
Blood Sample testing
Serum samples were tested for Brucella antibodies using RBPT technique as a screening test and all positive results were confirmed using c–ELISA technique. Samples that tested positive on both RBPT and c-ELISA were considered to be positive for Brucella. All these laboratory tests were conducted at the National Animal and Plant Health Laboratory, Ministry of Agriculture, Asmara, Eritrea.
Data analysis and interpretation
Data was entered in CSPro7.2 and analyzed by SPSS software version 21. Results were presented in percent and tables. Statistical analysis with p-value <0.05 was used to determine the association of the variables. There were a total of 9 questions of knowledge and 11 questions of practice. Those who respond correctly for each question were given one and zero for those with wrong response. Then results were summed and interpreted as percent. Those who respond greater than 70% correctly were considered as having good knowledge and practice and those scored with less than 70 % as having poor knowledge and practice respectively.
Ethical consideration
Ethical clearance was obtained from the Ministry of Health Research and Ethical Review Committee. Confidentiality of the study for participants was secured and written informed consent was requested from each study participants. Local and respected administrators, Ministry of agriculture and Ministry of Health of Northern Red Sea region were also informed before starting the research.