Access to and use gaps of insecticide-treated nets among communities in Jimma Zone, southwestern Ethiopia: baseline results from malaria education interventions

The data were obtained from a larger community based cross-sectional survey conducted during December 2013 to January 2014 in three districts (Mana, Goma and Kersa) of Oromia region. The three districts are parts of Jimma Zone in Oromia. The Zone is located on the geographic coordinates of 7° 40′ 0″ N latitude and 36° 50′ 0″ E longitude in Oromia, Southwestern Ethiopia. Figure 1 shows map of the study area. The data were collected to establish baseline indicators for malaria education interventions implemented through school students and religious leaders. The three districts were purposively selected based on the burden of malaria and absence of ongoing malaria interventions by Non-Governmental Organizations. Goma district is situated at an altitude that ranges from 1380 to 1680 m, and about 20.1 % of the landmass of the district is swampy. The total population of this district was 213,023 and only 5.9 % of its population were urban dwellers. Mana is a district neighboring to Goma and has a total population of 146,675 where females accounts for 49.1 % and only 3.0 % of the population were residing in urban areas. The altitude of the district ranges from 1470 to 2610 m above sea level. Likewise, Kersa district is located at altitude that ranges from 1740 to 2660 m. The population of the district was about 165,391, of whom 49.5 % were females, and 3.28 % of the population were living in urban areas. Oromo is the dominant ethnic group and the majority of inhabitants were Muslims in the three districts [12]. The three districts are found in similar geographical and ecological conditions. And, malaria is an endemic in the locality. In Ethiopia, which is also true to the study area, malaria transmission is mostly unstable and seasonal. Peak transmission generally occurs between September and December after the main rainy season (i.e. from June to August). However, in southwestern of Ethiopia where the study area is located, rainfall season often begins earlier in April and the area lacks clearly defined rainfall season. In addition, following a short rainy season in February and March, some minor malaria transmission also occurs in April, May and June. The country also experiences some inter-annual variation in rainfall and temperature leading to a persistent risk of malaria transmission patterns [5, 13, 14]. Comprehensive data are lacking on the pattern of malaria in the study area but one community based study in the adjust district reported a prevalence of 10.5 % where Plasmodium falciparum accounted for 40.9 % [15], and another retrospective study in one of the study district reported that Plasmodium falciparum accounts for 62.4 % of malaria cases [16].

This baseline study employed a community cross-sectional design triangulated with qualitative methods to assess households’ behavioral practices on malaria prevention and control in the three districts. However, this article reported LLIN access and use gaps in the study community. The sample size was determined using single population proportion formula (n = Z_1-α/2² p (1-p)/ d²) assuming the proportion of under five children who slept under LLIN during the previous night (55.4 %) [17], 5 % marginal error, 95 % confidence interval, design effect of 2 and 10 % non-response rate. This gave a sample size of 834 households. The households were selected randomly from 13 Gandas (the smallest administrative unit in Oromia, Ethiopia). The number of Gandas was decided considering logistic feasibility and resource available for the study. In each district, Gandas with high malaria risk were randomly selected. The stratification of malaria transmission as high risk, in the study districts, generally followed the national guideline which considered elevations below 2000 m above sea level as high risk area [2, 5]. In addition, some local factors such as malaria case load and availability of water collections for vector breeding were considered to define Gandas as having risk of malaria transmission. Consequently, the number of Gandas with risk of malaria transmission were 20, 16 and 26 in Kersa, Mana and Gomma respectively. Eight Gandas were randomly selected from Kersa and Mana (four each) and five Gandas from Goma district. First, based on probability proportional to size, the sample size was allocated to each district and within district; the sample size was proportionally allocated to each Ganda based on household size. Then, systematic sampling method was employed to select households for an interview. Household heads were considered for interview. However, the spouse was interviewed if household head was not present at the time of visit. Given that urban population approximately accounts for 15 % in Ethiopia [17], to have fair representation, about 15 % of the households were selected from urban settings. For the purpose of this study, urban was defined as an administrative town with municipality service. A total of six (i.e. two in each district) Focus Group Discussions (FGDs) with women and men in separate groups (three in each) were conducted. In addition, 11 key informant interviews, eight in Mana and Goma (four in each) and three interviews in Kersa, were also conducted with various categories of health workers, school teachers and religious leaders, to complement the quantitative data.

Standard questionnaire adapted from literatures was used to collect data onLLINaccess, ownership, and coverage [17, 18]. Household and population based indicators were used for assessment of universal coverage of LLIN [8, 10, 11].

The numerator consists of all households that own at least one insecticide mosquito net and the denominator is the total of number of sampled households.

The numerator contains all households where the ratio between numbers of LLIN owned, the number of household members is 0.5 or higher and the denominator is the total number of sampled households. This is the indicator of proportion of households having sufficient access to LLIN.

The numerator includes all household members in the sample who had access to LLIN assuming each LLIN was used by two people and the denominator is the population in the sample. The calculation of the numerator was performed based on Roll Back Malaria Monitoring and Evaluation Reference Group’s (MERG) procedures [11]: first, potential LLIN users were obtained by multiplying the number of LLIN in each household by the factor of two. Second, the access indicator was calculated by dividing the potential LLIN users by the number of members of each household and computing the overall sample mean of that fraction.

The numerator contains all members of the household identified as one of the users of an LLIN based on the listings of member and the denominator is the population included in the sample.

Additionally, two indicators were computed according to MERG’s recommendation [11] in order to facilitate interpretation of ownership and use gaps. They include Proportion of households with at least one LLIN for every two people among households owning any LLIN (P₅). This measures the saturation with LLIN for households with any LLIN. And, the inverse of P₅ (1-P₅) gives the intra-household ownership gap.

Proportion of population sleeping under an LLIN the previous night among those with access (P₆). The fact that method of calculating the access indicator does not allow allocation of access to specific individuals within the household, P₆ was calculated for the overall population included in the study by dividing the number of people who slept under LLIN the previous night to the total population who had access. The inverse of this indicator (1-P₆) gives the use gap.

The quantitative data were collected by experienced and trained field enumerators. The tool was translated into Afan Oromo (local language) and pretested on 5 % of the sample size in similar setting. The whole process of data collection was closely supervised by three academic staffs of Jimma University. The FGDs and the key informant interviews were conducted by experienced public health professionals who had master’s degree of qualifications.

The data were analyzed using SPSS version 17.0 software package. Proportions were computed for each indicator and presented in tables and figures. Binary logistic regression was applied to determine the association of background factors with LLIN use. A 95 % confidence interval and level of significance less than 0.05 were used during the analysis. The data from FGDs and the interviews were transcribed verbatim and then translated into English, and finally triangulated with quantitative findings.

The study was approved by the research ethics committee of Jimma University (Ref: RPGC/260/2013). The purpose of the study was explained to each respondent and informed verbal consent was obtained from all respondents.

A total of 798 households participated in the study giving a response rate of 95.7 % (Table 1). A total of 4107 persons were living in these households with average family size of 5.1 ± 2.1.

A total of 1174 functional LLINs were accessed in 798 households with mean1.5 ± 1.3 LLINs per household. Less than half (44.9 %) of the nets were distributed to households very recently (i.e. during the last six months). Overall, 70.9 % (95 % CI: 67.8–74.1 %) of the study households had at least one LLIN at the time of the survey (Table 2). However, of the households with at least one LLIN, the coverage of at least one LLIN for every two people was 88.9 % (95 % CI: 86.3–91.4 %).

At the time of interview, only 491(41.9 %) of the LLINs were hung over the sleeping areas while more than half 641(54.6 %) of the nets were kept folded. Table 3 shows LLIN access and use by district and place of residence. Nearly half, 51.9 % (95 % CI: 50.5–53.5 %), of the study population (2130/4107) had access to LLIN within the households, with mean access 54.7 % (95 % CI: 51.9–57.5 %). LLIN access was the highest in Kersa district and the lowest in Goma. Likewise, the rate of access to LLIN was higher in rural than urban settings.

Regarding LLIN use, 38.4 % (95 % CI:36.9–39.9 %) of the study participants slept under LLIN the previous night (Table 3). This net use coverage was higher in rural settings. Of those who had access to the LLIN, the proportion of people who used LLIN the previous night increased to 73.1 % (95 % CI:71.2–74.9 %) with little variations by district and place of residence. This gave an overall use to access ratio of 70.2 %. Despite low access to LLIN in Goma district, the prevalence of LLIN use was relatively higher among those who had access to it (Fig. 2).

This study identified two broad LLIN gaps (i.e. ownership and use gaps). Accordingly, 29.1 % (95 % CI:25.9–32.2 %) of the study households did not have LLIN at the time of the survey. Similarly, intra-household net gap was 11.1 % (95 % CI:8.5–13.6 %), meaning 11.1 % of the households with LLIN did not have sufficient access. Of those households who had any LLIN, 88.9 % (503/566) had sufficient LLIN for usual household members. However, 20.5 % (n =103) of these households had excess LLINs (i.e. more than one LLIN for every two people). On the other hand, of those population with sufficient access to LLIN, 26.9 % (95 % CI:25.0–28.7 %) did not actually use it. Table 4 shows reasons for not using LLIN.

Since the study was conducted during low malaria transmission season, the finding may not reflect year round situations of LLIN use among the populations. Furthermore, this study was limited to a small geographical area with similar seasonal and unstable malaria transmission and it may not reflect LLIN use in other eco-epidemiologic settings in Ethiopia.