Community knowledge, attitude and practices regarding zoonotic viral haemorrhagic fevers in five geo-ecological zones in Tanzania

This community-based cross-sectional study was carried out from April to November 2018. Respondents were selected using a multistage cluster sampling design. The country was divided into five ecological zones based on rainfall pattern, vegetation, land use pattern and altitude. Furthermore, these areas have previously been described as suitable habitat for primary vectors, which can contribute to the occurrence of mosquito- and tick-borne viral diseases [23]. Zone 1 comprised of the western parts of the country with tropical forest, unimodal rainfall pattern, and altitude < 2,300 m above sea level (a.s.l). In this zone, Buhigwe (in Kigoma Region) and Kalambo (in Rukwa Region) were selected as study districts. Zone 2 included the Southern highlands regions characterised by high precipitations, bimodal rainfall pattern, tropical forest and elevation > 2,300 m a.s.l. In this zone, Kyela district (in Mbeya Region) was selected. Zone 3 comprised of the north-eastern regions, with bimodal rainfall pattern and an elevation of < 1,800 m a.s.l. In this zone, Kinondoni (in Dar es Salaam Region) and Kilindi (in Tanga Region) districts were selected. Zone 4 covered the central regions of the country, characterised by moderate precipitation and unimodal rainfall pattern. In this zone, the selected districts were Mvomero (Morogoro Region) and Kondoa (Dodoma Region). Zone 5 comprised of the Lake Victoria zone, characterised by bimodal rainfall pattern; and Ukerewe (Mwanza Region) was the selected study district. In each district, three wards and nine villages were selected to account for local ecological biodiversity variations.

Based on the variations of the ecology in each zone, the number of individuals to be included in the study was calculated independently for each zone. Then population-weighted samples were used to split among the districts. The sample size for this study was calculated based on the conservative prevalence of peoples’ knowledge about VHF of 50%, a desired absolute precision of 5%, and a confidence level of 95%. The sample size was adjusted for clustering between districts by a design effect factor of 1.5. A contingency of 10% was considered to account for nonresponses, refusal or missingness. The minimum estimated sample size was 2,840 individuals. Within each district, individuals were evenly distributed between sampled wards and villages. Individuals over the age of 18 who voluntarily consented were interviewed.

A pre-tested semi-structured questionnaire installed in smartphones with digital data collection tool was used to collect sociodemographic and other relevant data [24]. The questionnaire consisted of four sections that addressed the (i) socio-demographic information; (ii) participants’ knowledge of the VHFs, source of information; (iii) attitudes towards VHF prevention; and (iv) exposure and risk practices related to VHFs. Trained research assistants administered the questionnaires in Kiswahili, the national language in Tanzania. To avoid stigma and improve participation, all interviews were organised privately.

Each response was scored 1 for the correct answer while 0 for the wrong or “don’t know” response. Knowledge about VHF was graded on 0 to 3, with 0 being the lowest and 3 being the highest. Participants’ scoring from 2 or 3 was considered “good knowledge” while 0 or 1 indicated “poor knowledge”. Attitude about VHF score was reversed graded on a scale of 0 to 3, with a score of 0 and 1 indicating “favourable attitude” and a score of 2 or 3 indicating “unfavourable attitude”. Finally, risk practice against VHF was measured on a scale of 0 to 22. Participants who received a score of 11 or higher were deemed to have optimal/good practice, whereas below 11 were considered to have at-risk/poor practice (Supplement 1).

Data analysis was performed using the Statistical Package for the Social Sciences (SPSS) version 27.0. Descriptive analysis was performed, and the results were reported as frequencies and percentages. Bivariable and multivariable logistic regression analyses were conducted to determine the relationship between socio-demographic characteristics and knowledge, attitude and practice levels. All variables with a p-value < 0.2 from the bivariable analysis were entered into the multivariable model. Possible associations were measured using an adjusted odds ratio (AOR) with 95% CI, and the p-value of less than 0.05 was considered statistically significant.

A total of 2,965 participants were interviewed from 24 wards and 48 villages of eight districts. The mean age was 35 (SD ± 18.9) years. Over half 1,726 (58.2%) of the study participants were females, and most respondents 1,826 (61.6%) had primary school education. The majority of the respondents were farmers. Half 1,803 (60.8%) were farmers (Table 1).

Table 2 shows the responses of participants on knowledge on VHF. Of the participants, 872 (29.4%) were familiar with at least one VHF and reported VHFs were EVD. RVF, CCHF, YF were least mentioned VHF diseases. Other diseases mentioned were dengue 87 (10%), chikungunya 43 (5%) and malaria 19(2.2%). According to 688 (79%) respondents, radio was the main source of information about VHFs. The majority of the respondents were of the (n = 714, 81.8%) opinion that VHF is preventable (Table 2).

Bivariable analysis revealed that education and occupation were highly associated with the knowledge index. The multivariable regression analysis indicated that occupation and education were the two independent predictors of the knowledge index. The employed participants (AOR: 5.82, 95% CI: 1.27–6.48) were more likely to have better knowledge than other participants. Respondents with highest education level had a great understanding of the VHFs compared to their counterparts (AOR: 2.03, 95% CI: 1.27–6.48). Respondents’ knowledge did not statistically vary between zones (Table 3).

The disease fatality was unknown to almost over half of the study participants. Three quarter of the respondents (= n 654, 75%) said they would either interact or take care of a person suffering from VHF. A large proportion of respondents (n = 637, 73%) said that they would visit regions known to experience VHF outbreaks (Table 4).

Age and sex were the two independent predictor variables of attitude towards VHF. Male participants (AOR: 2.72, 95% CI: 1.07–6.89) were more likely to have an unfavourable attitude than females. Participants aged 36 years and above had significantly less favourable attitude than their peers (AOR: 2.95, 95% CI: 1.25–8.27). There was no association between either education level, occupation or zone and attitude (Table 5).

About half of participants (n = 1,473, 49.7%) admitted to have had received a tick bite, removed a tick and/or crushed a tick with their bare hands during a period of six month before the study. About half of the respondents (1n = 420, 47.9%) reported having received a mosquito bite during the previous week. The majority of the participants (n = 1,897, 64%) were using mosquito nets to protect themselves from mosquito bites. Other exposure practices identified included keeping animals within the household 1,050 (35.4%), contact with wild animals or rodents 788 (26.6%), consumption of raw milk 726 (24.5%), frequent visits of outdoor recreation facilities 664 (22.4%) and contact with bats 471 (15.9%) (Fig. 1). Overall, 22% of the participants scored below the mean. The mean exposure practice score was 9.70 (SD = 1.17) out of a maximum of 22 points, indicating a low level of exposure practice.

Sex, age and residence of the respondents were significantly associated with the high-risk practices. There were increased odds of VHF exposure and risk practices among males (AOR: 2.95, 95% CI: 1.25– 6.27), participants aged 36–45 years (AOR: 3.58, 95% CI: 1.59– 7.82) and those who reside in Western, North-Eastern and Lake Victoria zone (AOR: 2.53, 95% CI: 1.07–6.66; AOR: 2.64, 95% CI: 1.13–7.58 AOR: 2.25, 95% CI: 1.07–3.84) than their peers in the other zones (Table 6).