Knowledge, Attitudes, and Practices on African Animal Trypanosomosis and its vectors Among Small-Holder Farmers Neighbouring Arabuko Sokoke Kenya: A Cross Sectional Study

Background: Animal African trypanosomosis (AAT) is a disease caused by several protozoan parasites transmitted by tsetse y (genus Glossina). AAT causes massive losses to farmers due to the high cost of control measures and treatment of sick animals. This study aimed to assess the knowledge, attitudes and practices of the local community on tsetse ies and trypanosomosis and evaluate demographic factors associated with adequate knowledge of tsetse ies along Kenya coast. Methods: We conducted a cross-sectional study between November and December 2017 among 404 cattle farmers living near the Arabuko-Sokoke Forest Reserve in Kili County. Structured questionnaires were used to collect data from randomly selected cattle rearing households. Descriptive statistics were used to analyze the local community's knowledge, attitudes, and practices regarding tsetse ies and trypanosomosis, and demographic factors associated with knowledge of tsetse ies were investigated using a logistic regression model. Results: Of 404 study participants, 214 (53%) were female, and 123 (30%) were elderly (>55 years old). Majority (N=312, 77%) were married, and 326 (81%) had only primary education or none at all. Although majority of farmers (N=353, 89%) knew the local name for tsetse y and that they are about the size of a housey (N=307, 77%), only 65 (16%) knew they transmit diseases to livestock, with 55/65 (85%) identifying they transmit Nagana disease. Overall, only 16% of the study participants had adequate knowledge on tsetse ies and trypanosomosis. Being 15 to 24 years old (aOR 2.86 (95% CI 1.10–7.45)) compared to those aged >55 years, secondary education (aOR 2.45 (95% CI 1.42–4.21)) and tertiary education level (aOR 3.78 (95% CI 1.53–9.31)) compared to no education were signicantly associated with higher odds of tsetse y adequate knowledge. Unemployment (aOR 0.15 (95% CI 0.10–0.23)) was signicantly associated with lower odds of tsetse y adequate knowledge when compared to self-employment. Conclusion: ndings suggest that farmers have poor understanding of trypanosomosis and its biological transmission by tsetse ies. Thus, training of livestock farmers will contribute to improved livestock health through better understanding of key livestock diseases, their vectors in transmission, and control.


Introduction
African trypanosomosis is a neglected tropical disease that is transmitted biologically by tsetse ies, (genus Glossina) and is found in thirty eight (38) sub-Saharan African countries including Kenya (1). Many studies have demonstrated the ability of biting ies such as Tabanids and Stomoxys to mechanically transmit Trypanosoma evansi and Trypanosoma vivax infections in horses, mules, camels, and cattle in Asia and South America (2,3). Indeed, T. evansi is no longer capable of infecting the tsetse y due to partial loss of mitochondrial DNA (4). Previous studies also showed that Trypanosoma congolense could be transmitted through mechanical means by the African tabanid Atylotus agrestis (5).
African trypanosomes are characterized by species-speci c distinct cell morphometry and motility patterns (6) and have exceptionally broad host range causing infections in both humans (causing sleeping sickness or human African trypanosomosis, HAT) and in domestic and wild animals (nagana or surra in camels) (7). HAT is caused by two subspecies of Trypanosoma brucei, namely Trypanosoma brucei gambiense and T. b. rhodesiense. It is estimated that 70 million people are at risk of HAT infection in Africa (8). HAT is rare in Kenya, but AAT is prevalent in 38 of the 47 counties. Kenya is endemic for tsetse ies and nagana (1).
AAT is mainly caused by T. b brucei, T. evansi, T. congolense and T. equiperdum (9,10). Trypanosoma vivax and T. congolense infect cattle, but can also cause disease in sheep, goats, pigs, horses, camels and dogs, among other wild animals like giraffe and waterbuck (9). Thus, AAT is one of the greatest impediment of livestock farming in Africa and gradually weakens the animals and reduces their draught power needed for cultivation of food and cash crops (11). Furthermore, it affects the livelihoods of farmers as it is one of the greatest causes of animal deaths leading to hunger and poverty (12). Further, the disease reduces the "living banks" reserves thus leading to inability to meet human social obligations such as dowry (13,14). Due to AAT, Kenya incurs huge agricultural losses of up to approximately 20 billion Kenya shillings in lost revenue and costs for treatment or death of animals annually (15). The capital channeled towards procurement of veterinary drugs for treatment and management of livestock trypanosomosis could otherwise have served to improve the living standards of the people in endemic areas.
Arabuko-Sokoke Forest Reserve (ASFR) is the largest and most intact coastal forest in East Africa with an area of 420 km 2 (16). This forest is infested by four species of tsetse ies namely Glossinabrevipalpis. G. longipennis, G. pallidipesand G. austeni (1). The human encroachment to formerly forested areas such as ASFR in search of arable land and pastures for their livestock has led to frequent exposures of humans and livestock with tsetse ies leading to transmission of diseases (17). Better understanding of livestock trypanosomosis and its transmitting vectors by communities living near tsetse belts have been identi ed as important barriers in disease control and prevention (18). Our study therefore sought to determine the status of community's knowledge, attitudes and practices towards tsetse ies and trypanosomosis commonly reported around the ASFR. area was purposely selected from three sub-Counties, namely Ganze, Kili North and Malindi, under which Arabuko-Sokoke Forest Reserve falls. The neighbouring area of the forest is mainly inhabited by the Giriama subtribe of the Mijikenda. Eight locations were selected as shown in Figure 1.

Study population
Participants sampled were aged ≥15 years and have lived or are living around the Arabuko-Sokoke Forest Reserve in Kili County for at least previous one year. These participants keep or kept livestock in the past near the forest.

Study Design and Data collection
We conducted a community based cross-sectional study including randomly selected livestock farmers between November and December 2017. Before the study commenced, sensitization and research engagement meetings were held in presence of the County administrators including chiefs and local leaders and livestock keepers in order to inform them of the proposed research. In each of the sub counties, sub-locations whose boundaries were less than 5 km from ASFR boundary were included in the survey. All the 32 geographical locations bordering the Arabuko-Sokoke were considered for inclusion but included eight locations randomly selected. The eight locations included were: Dida, Nyari, Chumani, Matsangoni, Mida-Majaoni, Mijomboni, Mongotini and Ngerenya as shown in Figure 1. In each of the eight locations selected, households to be included were randomly selected proportional to population size of the location.
Where a household was selected and did not rear livestock, it was substituted with the next household. Within a household, the questionnaire was administered to the household head or where not available, the most knowledgeable member of the household including older adolescents (≥15 years) who mostly look after the livestock in this region. Study data were collected using structured questionnaire administered by the trained research assistants supervised by the lead investigator and in language understandable by the study participants.

Study size
In a small study of only 70 households in Tanzania, the prevalence of knowledge of tsetse ies was 90% (21). With this 90% prevalence knowledge of tsetse ies, level of signi cance of 0.05 and precision of ±0.05, a sample size of 140 would be required. However, in our context we felt the 90% prevalence knowledge of tsetse ies was rather too high and thus estimated the study size using the 50% default prevalence adopted where there is no reliable data. With 50% prevalence, level of signi cance of 0.05 and precision of ±0.05, a sample size of at least 385 would be required.

Statistical analysis
The data were analyzed using STATA software version 15.1 (StataCorp, College Station, TX, USA). Data about knowledge, attitudes and practices of the local community on tsetse ies and trypanosomosis disease were analyzed using descriptive statistics and reported frequencies and percentages. To examine demographic factors associated with knowledge of the tsetse ies, we used logistic regression analysis. Farmers who could: a) correctly identify the local name for tsetse ies, b) provide correct description of the tsetse ies and c) were able to identify that tsetse ies transmits disease to livestock were classi ed as having adequate knowledge of the tsetse ies (the dependent variable). We used backwards stepwise binary logistic regression analysis to determine demographic factors associated with knowledge of tsetse y, retaining variables with P<0.1in the multivariable model but reported both crude and adjusted odds ratios and their respective 95% con dence intervals.

Ethical consideration
Permission to conduct the study was granted by the local leaders, livestock keepers and administration o cers. Ethical approval to conduct the study was granted by Pwani University Ethics Review Committee (ref no: ERC/PhD/015/2016). Prior to commencement of the study, formal verbal and written informed consent were obtained from eligible participants and from the parents/guardians of the children under 18 years who participated in the study.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Demographic characteristics
A total of 404 participants were recruited in this study. Two hundred and fourteen (53%) participants were female and approximately one-third (N = 123, 30%) were above 55 years. Only 47 (12%) were single, the rest were either married or were previously married but divorced, separated or widowed. There were 134 (33%) participants with no education, 192 (48%) with primary, 61 (15%) with secondary and 16 (4.5) with tertiary level of education. Only 41 (10%) were employed, the rest were either self-employed in their farms or unemployed. The number of participants from each location are shown in Table 1.

Knowledge Of Disease
Three hundred and ninety-seven (98%) participants reported to know a tsetse y however, it is 353 (90%) who could identify `Imbu' as its local name. Majority (N = 340, 84%) participants reported tsetse ies live in bush areas while 74 (18%) reported they lived in grassland. Three hundred and seven (76%) reported tsetse ies are approximate the size of a house y while 161 (40%) reported tsetse ies have long mouthparts for biting. There were 16 (4.0%) participants who did not know where livestock get into contact with tsetse ies, but 368 (91%) identi ed forest as where livestock get into contact with tsetse ies. Two hundred and eighty-nine (72%) reported tsetse ies' densities was highest during rainy season, 32 (7.9%) during dry season, 28 (6.9%) throughout the year and 48 (12%) did not know. Thirty-two (7.9%) did not know problems caused by tsetse ies to livestock. Only 65 (16%) knew tsetse ies transmit diseases to livestock, of which only 4 (6.1%) identi ed Nagana as the disease transmitted by tsetse ies. Health problems caused by the disease transmitted by tsetse reported by the 65 participant who knew it transmitted disease were: physical weakness (N = 43, 66%)), cows looking slim (N = 21, 32%), death (N = 16, 25%) and low milk production (N = 6, 9.2%) (Fig. 2). The source of information about tsetse ies and nagana disease was from practical knowledge through observation of tsetse ies among 50 (12%) participants (Table 2).

Practices And Attitudes
One hundred and nineteen (30%) participants employed tsetse control techniques in order to prevent tsetse ies from infesting livestock ( Table 3). The major tsetse control methods deployed included spraying livestock herds with insecticides (N = 61, 51%) and use of drugs (N = 42, 35%). Some 55 (14%) of the participants did not know reasons for persistence of diseases transmitted by tsetse ies, whereas 36 (8.9%) of the participants thought the reasons for persistence of tsetse-transmitted diseases was as a result of living near Arabuko-Sokoke Forest Reserve (Table 3). Three hundred and sixteen (78%) participants did not know if recommended tsetse y control methods are easy to use, 316 (78%) effective and 320 (79%) affordable (Table 3).  (Table 4). Further, in comparison to participants with no formal education, both secondary and tertiary education levels were associated with higher odds of tsetse ies' knowledge; crude odds ratios 3.90 (95% CI 1.47-10.32) and 5.58 (95% CI 2.01-15.50) respectively. Compared to self-employed, unemployed was signi cantly associated with lower odds of tsetse ies' knowledge (crude odds ratio 0.23 (95%CI 0.09-0.59)) ( Table 4). In the multivariable regression analysis, being 15 to 24 years old (aOR 2.86 (95%CI 1.10-7.45)) compared to those aged > 55 years, secondary education (aOR 2.45 (95%CI 1.42-4.21)) and tertiary education level (aOR 3.78 (95%CI 1.53-9.31)) compared to no education were signi cantly associated with higher odds of tsetse y knowledge. Compared to self-employed, unemployed (aOR 0.15 (95%CI 0.10-0.23)) was associated with signi cantly lower odds tsetse y knowledge (Table 4). No other demographic factors were associated with tsetse y knowledge. Discussion disease and, in this case, the vector. Understanding how AAT affects cattle owners and how they handle the disease is vital to developing effective, locallyadapted control programmes (22). Community participation in tsetse control activities ensures sustainability of the programmes especially if the community members and extension o cers are involved in design of the programmes (23).
Overall, we found only 16% of the livestock-keeping communities living near Arabuko Sokoke Forest Reserve had adequate knowledge on tsetse ies and animal African trypanosomosis. Similar studies in Ethiopia and Tanzania around Baro-Akobo and Gojeb river basins, and Serengeti National Reserve, respectively, found out that the communities around the reserve had limited knowledge of tsetse ies and trypanosomosis (24) (21). This is in contrast to other studies where 23.1% of the farmers could associate the causal association between tsetse ies and trypanosomosis among communities surrounding Shimba Hills National Reserve (25). In other studies in Tanzania, 90% of the respondents interviewed in communities surrounding Serengeti national park (27) (21) had adequate knowledge of trypanosomosis. However, earlier studies showed that people living around Serengeti National Park in Tanzania were knowledgeable about tsetse ies and human African trypanosomosis (27) and the tsetse ies (26) while those in Mateu district in Tanzania were more knowledgeable in AAT than HAT (28). The reason for the higher level of knowledge for in Serengeti may be attributed to the high number of respondents (45%) who had attained secondary education compared to 19% in this study for those with secondary and post-secondary education while the participants without formal education together with their counterparts with primary level education comprising of 81% of the respondents. People who are familiar with the clinical signs of the disease are more likely to seek assistance from the veterinary o cers because they can identify the sick animals early before their body conditions deteriorate. Furthermore, unlike other studies in which male predominance is high, particularly in the active age groups (21)(29), this study was predominated by women (53%) and people above 55 years (44%).
The huge difference in knowledge gap noted in this study could be attributed to the fact that studies by (21) occurred just after recent outbreaks which were followed by awareness campaigns. The other reason for the difference in knowledge gap could be the population sampled. Another study in western part of Serengeti reported 95% of the respondents were knowledgeable on tsetse ies and trypanosomosis (26). The study participants in Serengeti were predominantly male unlike the present study which was dominated by women, who dominate livestock keeping among the Giriama community. The higher number of older people sampled in this study may be due to their preference to live in the villages, unlike the younger educated generation with preference to live and work in cities. (26) Furthermore, the limited knowledge by the participants in this study, on the vector and the disease transmitted, in uenced their attitude towards prevention and control measures. The vast majority of the respondents (98%) assumed that they knew tsetse ies, but most of them were unable to describe tsetse y or could not distinguish it from other biting ies. The respondents' inability to describe a tsetse y raises questions about their certainty that tsetse y bites cause disease. They did, however, realized that contact with tsetse ies occurs primarily in the forest while the animals are grazing. Secondary and tertiary education were associated with higher knowledge of AAT compared to primary education and no formal education. This could be attributed to the knowledge acquired in post primary education. In one section of Dida location, where the wild animals watering point is at the edge of the forest near the village, the respondents noted that contact with tsetse ies occurred when the animals, especially elephants go to the watering point to drink water. Indeed, knowledge of tsetse ies and measures of trypanosomosis control have been shown as important factors in uencing the community's willingness to contribute resources and participate in control activities (23). Indeed, communities with a low level of education are principally vulnerable, and could be di cult to integrate them into control programmes (30).
Regarding the description of the disease transmitted by tsetse ies, most of the respondents could not describe the clinical signs of an animal suffering from trypanosomosis. Among the Giriama community, any form of sickness is locally referred to as "homa" which means fever. Many in this community believe that tsetse y bites cause "homa" in both humans and livestock. They also noted that hybrid animals such as Friesian and Ayrshire are the most affected, while a cross breed with the Zebu and Borana cattle (local breeds) are more tolerant to bites by tsetse ies. Indeed the Maasai Zebu and the Orma Boran breeds have been shown to be more resistant to trypanosomosis over other local breeds such as the Galana Boran and the exotic breeds (31). Other African local cattle breeds such as the Baoulé, Zebu, Ndama and Ndama/Baoulé crosses have various degrees of resistance to African trypanosomosis (32).
Compared with the unemployed and the uneducated, educated and employed farmers were found to be implementing the recommended methods of tsetse control. The attitude of the participants without formal education further in uenced their use of control methods for tsetse and trypanosomosis. Most uneducated people did not know the recommended methods for controlling tsetse and trypanosomosis and therefore could not use them. This is in agreement with other studies where knowledge of disease in uenced preventive practices (33). Other studies, on the other hand, have shown that disease knowledge does not always correlate to practice, with vector management strategies being lower than disease knowledge (30). Some of the participants in this study were aware on the control measures but did not believe it was worthwhile to use them. Others were aware of the control measures but could not afford to implement them. We also noticed that some of the farmers were dipping or spraying their livestock, but they were unaware that acaricides could prevent tsetse ies from biting livestock. While this study did not investigate the types of acaricides used by farmers living around Arabuko Sokoke Forest reserve, other studies in Uganda have shown that farmers may use acaricides that are resistant to tsetse ies but only effective against ticks (34). Although the County Government of Kili has dips, most of them were non-functional due to challenges such as water shortages and overdependence on the government to run the dips. Subsequently, tsetse control groups were formed in all sub-Counties and were supplied with pumps and pyrethroid-based acaricides that are effective against ticks and tsetse to supplement the dipping service.
In comparison to the reference group of 15-24 year olds, the majority of people, particularly those over 55 years old, did not know whether the recommended tsetse control measures were effective or worth using at all. As the head of homesteads, they also felt that the control measures were costly. This could be attributed to the fact that this group supports families and thus contributes money for tsetse control compared to the youth between 15 to 24 years. Furthermore, household heads preferred keeping goats over cattle because the cost of keeping cattle is high due to the requirement of regular dipping whereas goats are more resistant to trypanosomosis. The fact that goats are more resistant to trypanosomosis may explain the low prevalence of trypanosomosis among caprines (35) (36).

Study Strengths And Limitations
To the best of our knowledge, this is the rst study to assess the knowledge, attitudes and practices of communities surrounding Arabuko Sokoke Forest Reserve on tsetse ies and animal African trypanosomosis. KAP surveys help to identify gaps in community tsetse control measures, that, if lled, will make it easier for them to participate in AAT control activities and, in turn, increase tsetse control coverage and e cacy. This study forms a baseline for future studies.
The study relied on reported data from the farmers which could be subject to reporting bias. Some farmers from selected households did not keep livestock and were excluded in this study, however, these were very few. The cross-sectional survey measured knowledge, attitude and practice at the time the study, which may have changed over time.

Conclusion And Recommendations
The gap identi ed in this study in the knowledge of tsetse ies and trypanosomosis among the community surrounding ASRF, especially women who are the majority of livestock keepers, should inform prioritization of this group in control tsetse and trypanosomosis control programs. Public sensitization on tsetse control may be done through the county agricultural extension services or organizations such as the 'Friends of Arabuko Sokoke Forest'.  Health problems caused by the disease transmitted by tsetse ies in animals (N=65).

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.