Study Site
Abu Arish governorate is located at the heart of Jazan in the southwest of the Kingdom of Saudi Arabia in a fertile agricultural plain where vegetation is closely related to that of Yemen and East African countries such as Ethiopia and Eritrea (33)( Figure 1). It is strategically located between high mountainous areas in the east and the Red Sea coast in the west, at the intersection of longitude 16°58′08″N with latitude 42°49′57″E (an area of approximately 900 km2) (33). The altitude is about 67m above sea level and approximately 30km from the Red Sea coast. The average temperature in Abu Arish is 29.9 °C (min. temperature 21°C to max. temperature 38 °C), and precipitation averages 93mm with humidity of 50% (34, 35).
The governorate consists of the main city Abu Arish, which is surrounded by 56 small villages (the number of inhabitants in the villages range from 50 to 150) connected by a network of paved road and semi–paved roads (9). The population size is 197,112 (2010 census) with a growth rate 2.24 (36). The locals are mainly engaged in agricultural activities including animal husbandry, growing fill flowers, vegetables, lemons, guava, and mangoes in about 652 small– to medium size farm (average number of 20 farm per village with five workers per farm in average) (37). The completion of Jazan valley dam and established irrigation network with a length of 50km attracted a significant number of international migrant workers, including irregular migrants to the area (8, 9).
Plasmodium falciparum dominates malaria transmission (38, 39). The primary malaria vector is An. arabiensis, the only positive vector for malaria in Saudi Arabia with a sporozoite rate (0–0.7%) (34, 38) and the population presents in low density during autumn and spring and peaks during winter following the rainy season(40). The natural habitat of An. arabiensis are wadi and rain pools, but it also can be found in domestic water containers (shallow water with pH7–9). Adult biting–behavior is both endophagic and exophagic. An. arabiensis is found resting mainly in human dwellings (endophilic), and both anthropophagic and zoophagic (41)
Formative Assessment
Development of International Migrant Laborer Profile
The lack of clear definition for MMPs, and the fact that they are an extremely heterogenous population with no standard definitions, indicates that a clear definition and sub–group profile needs to be created for the purpose of this study to be used for future intervention. Rather than classifying the migrants based on their legal status, documented or undocumented, a more descriptive set of definitions (Table 1) was adapted from language used by the General Authority of Statistics in Saudi Arabia along with many international organizations like ILO, IOM, and UNHCR to better categorize the distinct differences in the various migrant populations at the border area.
Modified Peer Navigators Sampling
The peer navigator (PNs) recruitment method uses a chain–referral approach (Snowball Sampling) (32, 42, 43). Based on the information collected from the formative assessment, the study team recruited non–randomly selected peer navigators (seven migrants; only three continue the work) from different ethnic backgrounds to ensure that they were representative of the international migrant workers in the area and well–connected to the targeted population. The peer navigators underwent 12 hours of training that covered the aims of the study, filed work protocols, electronic data collection, informed consent, ethics, GPS point recording and interview skills. PNs recruited and interviewed there in the farms and community settings. To ensure wide coverage of the population so our sample size was as representative as possible, we also tried to reach marginalized subpopulations on each farm that was visited. Every eligible individual present at the time of interview was asked to participate in the survey. Additionally, GPS coordinates were recorded, and the total number of individuals present. If the participant were working in another location, he would be asked to provide his working location or GPS coordinates.
This study was an effort to collect new information about MMPs (e.g., international labor worker) in the Jazan region and was conceived as an experimental study; no formal sample size calculation was performed.
Study instrument
The research instrument for this study was a standardized questionnaire that was developed from earlier studies’ questionnaires related to malaria and MMPs studies (27, 30, 44, 45). The questionnaire included sections on socio–demographic status, work history, housing conditions, social network extent and depth, access to malaria health message, history of malaria infection, knowledge about malaria, malaria prevention activities and access and use of preventive measures (bed net ownership and use of bed net), treatment–seeking behavior, and knowledge about malaria and vector control program services in the area.
Data collection and management
A cross–sectional survey was conducted in seven villages in Abu Arish governorate starting November 2019 to mid–January 2020 among agricultural workers working in farming, animal husbandry, and daily laborer. All individuals who were males 10 years or older who worked, or intended to work, in the study area at the time of the survey were eligible for inclusion. Due to the sensitive nature of the population singed consent was not obtained; however, the consent form was read to them and were given a copy to read. All questions were addressed prior to oral consent from the participant was obtained. Survey data were electronically collected from consented participants using ODK–Collect software; each participant was asked to respond to a pre–tested questionnaire that was available in Arabic and English. Also, over the phone translation was provided in Amharic and Tigrine to those who needed it. Then data were loaded on secure cloud servers using the Android–based Moto (G7 and Moto G6) smartphone. Data quality was checked for completion and reconciled after each collection session with an off–sight data manager through a communication system, and any field notes and feedback from the data collectors were recorded.
Measurement
Socio–demographic variables
Educational status of the participants was categorized as no formal education, less than primary education (can write and read with no formal schooling or did not complete primary school), primary, and secondary or above. The languages that participant speaks and read. Country of birth. Age was collected as continuous variables, and then categorized into groups using five–year intervals. A simple wealth index was created based on personal assets possession and housing conditions using principal components analysis (PCA) (46, 47). The patient population was then categorized into wealth quartiles. The duration of residency was collected by asking the participants to choose one of the following categories: two years and more; less than two years and more than six months; or less than six months. A variable was created to classify international laborers (according to Saudi Arabia General Authority of Statistics) and to classify the agricultural worker as “temporary” or “permanent.” Current job and previous job before coming to Jazan.
Extent of social network
To measure the extent of a participant’s social network, the participants were asked the following questions: how many people like them aged 15 and above currently residing in the area; how many people they know by the first name working in the area; among who they know by the first name, how many did they met in the past 30 days; and how many have been recruited into the study. All answers were recorded as integers.
Self–reported malaria.
Participants were asked a yes or no question if they had malaria infection in the last six months, and another dichotomous question if they had received malaria treatment in the last six months. If a family member, coworkers, or a neighbor had malaria, the answer was recorded as “yes,” “no,” or “I do not know.”
Knowledge on malaria
Questions were asked related to different aspects of malaria to measure knowledge. They were ranging from knowledge of malaria transmission, knowledge of malaria vector, signs and symptoms of malaria, treatment, and prevention measures. These common principles will be used to measure knowledge (48-52). The respondent correctly reporting the three following statements defined “correct knowledge”: malaria parasites are transmitted by mosquitoes; both children and adults are at risk; and ITNs can be used to prevent malaria parasite infection(48).
Respondents who stated all three correctly were dichotomized as having the correct malaria knowledge, and the remaining as not having correct malaria knowledge.
Attitudes (perception) towards malaria variables
Perception of beliefs on susceptibility, seriousness, and threat of malaria was measured using a series of 11 Likert–scale statements (four negative and seven positive statement) were presented to the study participants[1]. The responses were captured using a scale of five, ranging from completely agree to completely disagree.
In the analysis to calculate the overall perception statements score the following steps was taken: First, each statement was coded as −2 “completely disagree,” −1 “disagree,” 0 “neutral”, 1 “agree,” or 2 “completely agree.” Second, the invert of all negative statements was taken. Third, the response to eleven statements were added-up and divided by 11 to generate mean attitude (perception) levels for everyone. Last, based on the mean score, respondents were further categorized as having acceptable perceptions when their score was ≤ 0; having positive perception when their score range was between 0.01 and 1.0; and, having very positive perception when their score was between 1.01 and 2.0(53-55).
Exposure to malaria health message
Participants were asked if in the past six months they had heard or seen any messages or information about malaria the answer was coded as “yes,” “no,” or “I do not know.” Participants who received health messages were asked about the nature of the health message. A list of topics (that the local malaria program promotes) was given to the respondents so they could provide more than one response, as well as the source of the message. All the participants were asked what malaria health messages they would like to receive in the future and what are the sources that they prefer all in multiple responses.
Practice of malaria prevention, visit health facility, and use of personal protection.
Participants were asked eight questions to measure routine activities and actions of the participants used to prevent malaria. Participants were asked about bed net ownership, and their responses was captured as “yes” or “no.” Among participants who owned bed nets, they were asked if the bed net was treated with insecticide, and if they slept under the bed net the night before they were surveyed with a “yes” or “no” question.
All the participants were asked about mosquito avoidance practices and seeking medical care in case of having a fever. The answer was recorded in scale of three ranging from “never,” “sometimes,” and “always.”
Local malaria program outreach activities
Participants were asked if they heard about malaria and vector control centers in their area, their response was captured as “yes,” “no”, or “I do not know.” Also, the participants were asked if they received a personal visit in the last six months from malaria health workers, their response was captured as “yes” or “no.” The number of personal visits was recorded as an integer. Participants were asked if in the last six months their village or working place receive a visit from the malaria health worker, their response was captured as “yes” or “no.” During the visit a list of activities was given a multiple response was recorded.
Statistical Analysis
Analysis was carried out in STATA v16 (College Station, Texas). To explore the information on malaria correct knowledge, malaria treatment and prevention access, malaria infection and bed net ownership and use. descriptive statistics was used to summarize survey data primary outcomes. Chi–square and Fisher’s exact test statistics was used to assess differences in bivariate outcome. A wealth index was developed based on a principal component’s analysis of personal possession index. The raw factor scores will be categorized at the median value into lower and higher asset categories.
[1] Appendix C shows the 11-perception statement used to capture the attitudes toward malaria disease and related outcome.