Communication for behavioural impact in enhancing utilization of insecticide-treated bed nets among mothers of under-five children in rural North Sudan: an experimental study

A randomized community trial was conducted in rural areas of the Kosti Locality, White Nile State, Sudan from January 2013 to February 2014 to assess the effectiveness of the COMBI strategy in enhancing utilization of LLINs among the mothers of under-five children.

White Nile State is one of the Sudan’s 18 states. It lies in the South of the country, and bordered by Khartoum state from the north, South Sudan from the South, South Kordfan state from west, Blue Nile State from south east and from the east Senar, Gazera. This location makes the state susceptible to malaria transmission throughout the year as this area is an agricultural one. Primary malaria control and prevention activities in White Nile state are undertaken by the State Ministry of Health, the National Malaria Control Programme (NMCP) together with non-governmental organizations (NGOs). Kosti locality is one of White Nile state localities and it covers a total area of 1220 sq km with a total population of 179,800.

The study population comprised of mothers of under-five children who were living in rural areas (villages) of Kosti locality. Mothers having plans of leaving the area during the course of the study were excluded. The study sample was determined by using the following formula: n = Z2PQ/D2 [13] and was found to be 761 mothers of under-five children, where Z is the corresponding value to 95 % confidence limits = 1.96, P = utilization rate of LLINs (0.45), Q = 1 − P = (0.55), D = desired margin of error (absolute precision) = 0.05, and Deff = 2.

Out of a total of 163 villages in Kosti locality, 12 villages were selected according to the NMCP criteria for LLINs distribution and were used for this study. These criteria were villages near to White Nile river bank which were surrounded by farms and having a mosquito density all year round- four mosquitoes per room, 3.8 larvae per dip. Randomly, six villages located in North West region of the Kosti locality were assigned to the intervention group and six villages located in South East region of the Kosti locality were assigned to the comparison group. Both groups were assessed twice during the study period: pre-intervention and post-intervention. The calculated total sample size (761) was divided into comparison and intervention group proportionate to population size and hence 412 (54.1 %) mothers from a total of 1720 mothers in intervention villages and 349 (45.9 %) from a total of 1460 mothers in comparison villages were selected for the study. Systematic random sampling was used-every fourth mother was selected from the list of mothers prepared by the NMCP prior to distribution of LLINs.

The study included three phases, namely pre-intervention phase, intervention phase and post-intervention phase.

During pre-intervention phase, baseline data was collected in the last week of January 2013 from randomly selected mothers of under-five children using a structured questionnaire that included socio-demographic characteristics of households, KAP of mothers related to malaria prevention and villages’ environment, sleeping behaviour of mothers and general conditions and utilization of LLINs (see Additional file 1). In addition, baseline malaria parasitological data were collected during the malaria transmission season in the 1st week of February 2013 from both intervention and comparison groups. One child of less than 5 years of age was selected randomly from each household from where the mothers were included in the study sample. Two drops of blood were taken from each selected child, one drop was wiped off from the finger using a swap dipped into methylated sprit and the other drop was used for testing in the field using rapid diagnostic test to detect P. falciparum. Before baseline data collection, during the first 10 days of January 2013, a total numbers of 13,650 LLINs were distributed in both intervention and comparison villages. Different household received different numbers of LLINs depending on its population size- one net for two members.

COMBI activities were carried out for a period of 10 months between February and December 2013. COMBI activities were implemented from different points of view as summarized below:

Personal selling was done through village health promotion teams. For every ten mothers of under-five children in intervention villages, a volunteer made house to house visits with the frequency being once in every month for a period of 10 months to create awareness about the importance, use and maintenance of LLINs.

Focus group discussions were conducted by the locality committee and health promotion teams with all sampled mothers of under-five children (412) in the intervention villages. The FGDs were about malaria and its prevention modalities including the importance, use and maintenance of LLINs. Each FGD consisted of 8–10 mothers and was facilitated by the head of the health promotion team and one malaria prevention specialists from the locality committee. FGDs were repeated for reinforcement after six months of the initial FGD.

Sustained massive advertising: Distribution of health education materials, including posters, pamphlets and books, to mothers of under-five children in intervention villages in women public places, schools, primary health care centres and households having mothers of under-five children.

Health education sessions were conducted monthly through lectures, group discussions and workshops by health promotion teams regularly and by local committee occasionally to mothers of under-five children and school children in the intervention villages. The health education sessions were conducted following the manual developed by the federal ministry of health.

Community mobilization was done by health promotion teams through monthly mothers’ group meetings and home visits.

Administrative mobilization was done monthly by the locality committee with each health promotion team in intervention villages with a purpose of mobilizing the teams through meetings and discussions.

At the post-intervention phase, in January 2014, the same questionnaire which had been used for baseline data collection was used for data collection from intervention and comparison groups. While the end-line malaria parasitological data were collected during the malaria transmission season in the first week of February 2014 from both intervention and comparison groups. Like baseline data, one child of less than five years of age was selected randomly from each household from where the mothers were included in the study sample. The same group of data collectors, who did baseline data collection, collected end-line data. Both baseline and end-line survey was conducted during malaria season and in the same months, last week of January, in the year 2013 and 2014, respectively. In order to avoid bias in response to the questionnaire, data collectors were not involved in COMBI activities. Data collectors received training to avoid leading questions. A structured questionnaire was used and data collectors just read out the questions and recorded respondents’ response without providing any further clue.

Depending on the nature of the variables, descriptive statistics were used to tabulate and describe the data (frequency distribution, percentages, means and standard deviations), and inferential statistics (Chi Square and Fisher’s exact tests) was used to examine association between categorical variables. Odds Ratio (OR) and Eta Square (η²) were calculated to determine effect size of independent variables on dependent variables.

The data was analysed using the Statistical Software Package for Social Sciences (SPSS) version 18. For close-ended questions of the questionnaire, the replies of equivalent meaning were pooled into different categories during analysis and percentages were computed for different options. The tests were carried out at 95 % confidence interval, p values less than 0.05 were considered significant, and OR with 95 % confidence interval was used to assess the presence and degree of association between dependent and independent variables. Multiple logistic regression was performed using the Stata/SE 13.1.

Data collectors were trained on ethical issues- obtaining verbal informed consent, respecting mothers’ decision of not participating in the study, ensuring participants’ privacy and confidentiality of information provided. Verbal informed consent was obtained from each participant before enrolling them in the study. At the end of each day data collection, all hard copies of completed questionnaires were submitted to the health promotion team co-ordinators in each village and the researcher for daily review for consistency and quality of data, and to safeguard confidentiality of the information provided. All malaria positive cases detected during data collection were given, by a qualified nurse, an appropriate age-weight specific treatment course as per the national treatment guidelines.

Of the 761 participants in the study, in both intervention and comparison villages, almost half had only primary education (49 and 45.6 % respectively), while over a quarter (28.4 %) did not have any formal schooling. Majority of the respondents’ husbands were on irregular free job both in intervention (63.6 %) and comparison villages (62.2 %) and the mean household monthly income was 68.8 and 77.1 USD, respectively in intervention and comparison groups. In intervention and comparison group, respectively, the mean age of the mothers were 29.3 and 30.3 years; average family size was 5.3 and 5.6; and average number of under-five children was 1.6 and 1.7. No statistically significant difference was found between intervention and comparison group on socio-demographic variables (see Table 1).

Majority of mothers were found to have correct knowledge about malaria vector both before and after intervention. However, there was a significant increase in knowledge about malaria vector among mothers in intervention villages from 86.9 % before intervention to 97.3 % after intervention and hence the knowledge level increased by 11 % (Eta square η² = 0.11). No such significant increase was observed in comparison group of mothers.

In intervention villages, 82.5, 43.9 and 45.6 % mothers knew fever, joint pains, vomiting and diarrhoea respectively as the symptom of malaria before intervention, while after intervention the respective knowledge of mothers significantly improved to 98.3, 85.7 and 86.7, respectively (p < 0.05). 25 % mothers cited one of the household members had been infected by malaria before intervention that significantly reduced to 18 % after intervention (p < 0.05).

Also, the results revealed that 34 % mothers cited not using LLINs continuously as the major reason for getting malaria infection which significantly increased to 47.3 %, after intervention, while 9.7 % mothers cited never using LLINs as the second reason to get malaria infection which significantly increased to 32.4 %, after intervention (p < 0.05). This indicates that level of knowledge significantly increased by 22 % after intervention (η² = 0.22). With regard to knowledge about personal protective measures against malaria, 85 % of mothers stated that they used LLINs as personal protective measure, which significantly increased to 92.1 % after intervention (p < 0.05).

Related to recognizing the types of mosquito nets, most of mothers in intervention villages (84.3 %) recognized all types of mosquito nets than the mothers from comparison villages (11.2 %) (p < 0.05). This also indicates that level of knowledge increased by 64 % following intervention (η² = 0.64). Also, 47 % mothers recognized all types of ITNs before intervention which significantly increased to 56.8 % after intervention (Table 2).

Regarding malaria as a preventable disease, Table 3 depicts that 80.6 % of mothers perceived malaria as a preventable disease which significantly increased to 99.3 % after intervention, (p < 0.05), (η² = 0.30). 75.2 % mothers recognized malaria as a serious disease before intervention, which significantly increased to 98.8 % after intervention (p < 0.05). In intervention villages, the results showed that the negative attitude of mothers about LLINs with large holes (damaged LLIN) protecting from malaria significantly reduced from 32.5 to 12.4 %, after intervention (p < 0.05) (η² = 0.24).

56.8 % of the mothers had a positive perception about effectiveness of LLINs in preventing against mosquito bites that significantly increased to 92.2 % after the intervention (p < 0.05) (η² = 0.41). Also, 52.7 % mothers encouraged their family members to use LLINs before intervention which significantly increased to 93.7 % after intervention (p < 0.05). Likewise, 77.7 % mothers had positive perception about effectiveness of LLINs in prevention of malaria that significantly increased to 96.1 % after intervention (p < 0.05).

While in comparison villages, 54.2 % mothers’ perception about effectiveness of LLINs as preventive measure against mosquito biting at the time before intervention slightly increased to 59.3 % at the time after intervention (p > 0.05). 52.6 % mothers encouraged their family members to use LLINs at the time before intervention, which slightly increased to 56.2 % at the time of after intervention (p > 0.05). On the other hand, in the perception of mothers about effectiveness of LLINs in prevention of malaria, there was no significant change (p > 0.05).

Figure 1 presents data on LLINs utilization amongst mothers of under-five children in households that owned nets. The questions focused on daily use of LLINs. In intervention villages, before COMBI intervention 19.2 % mothers reported that they had slept under a bed net which significantly increased to 82.2 % after intervention, (p < 0.05), [OR = 4.6, CI = (3.72–5.72)]. This indicates 64 % improvement in LLIN utilization following COMBI activities (η² = 0.64). In comparison villages, the percentage of households using LLINs appeared to be slightly increased from 18.1 to 22.6 % at the time before intervention and after intervention, respectively (p > 0.05).

The results presented in Table 4 revealed that the LLINs utilization was more in households who have sufficient knowledge about malaria (3.4 times) [OR = 3.4, 95 % CI (1.82–6.34)], households who had members with formal education (3.6 times) [OR = 3.6, 95 % CI (2.40–5.40)], Government employees (3.7 times) [OR = 3.7, 95 % CI (0.87–15.8)] and also in those who had less than three children (3.5 times) [OR = 3.5, 95 % CI (0.89–13.80)]. However, income and family size did not affect LLINs utilization (p > 0.05).

Table 5 demonstrates that after controlling for demographic variables there is a significant association between being in the intervention or in the comparison groups and the utilization of LLINs following 10 months of COMBI activities. Being in the intervention group was associated with a 0.59 points increase in the log of odds of using LLINs in comparison to the comparison group. This table also presents that there is a significant association between education and LLINs utilization. Formal education was associated with a 0.16 point increase in the log of odds of using LLINs in comparison to informal education following COMBI. No significant association was found between the LLINs utilization and other socio-demographic variables- occupation, household monthly income, household population and number of children under 5 years in the household. This model could explain 37 % of the total variation (adjusted R squared 0.37).

Regarding overall knowledge about malaria among mothers in intervention villages there was improvement in knowledge from 3.8 to 34.2 % (p < 0.05) (η² = 0.64). This signifies that level of knowledge increased by 64 % in intervention villages. However in comparison villages, there was no significant improvement (p > 0.05).

With respect to overall attitude about malaria among mothers in intervention villages 75.2 % of mothers had a positive attitude towards malaria, which significantly increased to 99.3 % after the intervention (p < 0.05). However 24.8 % of them had a negative attitude towards malaria, which significantly reduced to mere 0.7 % after the intervention. Mothers after the intervention developed a positive attitude towards malaria more than two times higher than mothers in the comparison villages (OR = 2.25). However, there was no significant change in overall malaria KAP score in the comparison villages (p > 0.05).

About the overall score of practice of malaria prevention in intervention villages, 0.8 % of mothers of under-five children, had good preventive practices against malaria that significantly increased to 33.6 % after the intervention (p < 0.05) (η² = 0.79). While 32.4 % of them had poor preventive practices, that significantly reduced to zero after intervention (p < 0.05) (Table 6).

Table 7 shows that in intervention villages prevalence of malaria among children reduced from 4.7 % before intervention to 2.3 % after intervention (p < 0.001). No significant change was seen among the comparison villages. Infection was higher among children in comparison villages (4.8 %) compared to children in intervention villages (2.3 %). It is an indication that COMBI was found to have significant influence on reduction of malaria prevalence [OR = 1.4 CI (1.05 − 1.9)].