The use of intermittent preventive treatment in pregnancy and insecticide-treated bed nets for malaria prevention by women of child-bearing age in eight districts in Malawi

The household survey has been described in detail elsewhere [8]. Briefly, we conducted a two-stage cluster-sample cross-sectional household survey April 16–30, 2009 at the end of the long rains, in the middle of the high malaria transmission season. The study was conducted in both urban and rural communities in eight of the 28 districts in Malawi (Lilongwe, Blantyre, Mwanza, Chiradzulu, Phalombe, Rumphi, Nkhotakota, and Karonga), which were selected in collaboration with the Ministry of Health to serve as sentinel districts to monitor malaria burden and coverage of malaria control interventions—but did not constitute a nationally representative probability sample. These districts contain approximately 33 % of the entire population of Malawi and are dispersed in the north, central, and south regions (Fig. 1). A two-stage cluster sampling design was used. The first stage was composed of selecting enumeration areas (EAs). Altogether, 30 EAs per district were chosen using systematic random sampling with selection probability proportional to estimated size using the 1998 census. In the second stage, the EAs were divided into segments of roughly 30–60 households and a segment was randomly selected using a personal digital assistant (PDA; Dell Axim X50s, Dell Inc., Austin, TX, USA) with a specially-designed programme for random segment selection developed by the Centers for Disease Control and Prevention, USA. All households or a randomly selected subset of households in a selected segment were invited to participate in the survey. Informed consent was obtained from the head of household or other adult household resident. All the minors between 7 and 17 years were asked to provide an assent to participate in the study. All household members were asked to participate. A household listing was created which included all people who usually live in the household as well as guests of the household who stayed in the household the previous night. Information on the number, and visual inspection on the integrity of the bed nets in the household was collected by creating a bed net roster, which included information on whether nets were ITNs or conventional nets. Using the bed net roster and household listing, we recorded which household member slept under which bed net, allowing us to collect detailed, linked information on household members and bed nets in order to explore useful information on the ITN ownership and use. Women of child-bearing age (WOCBA) were defined as all women aged 15–49 years. The use of IPTp was defined as the uptake of at least two doses of sulfadoxine–pyrimethamine (SP) with at least one received at an ANC visit under direct observed therapy.

Ethics approval was obtained from the Malawi College of Medicine Research and Ethics Committee (COMREC) in Blantyre and Centers for Disease Control and Prevention (CDC) Institutional Review Board in Atlanta, GA, USA.

Analyses were performed using SAS version 9.2 (SAS Institute, Cary, NC, USA) using the proc survey procedures, which use the Taylor expansion method to account for cluster sampling and unequal selection probabilities. Analyses were weighted and weights equaled the inverse of the exact probability of selection. Percentages reported in this report reflect this weighting unless otherwise noted. Results were defined as statistically significance if the p value was <0.05.

Multivariate logistic regression was used to assess predictors of IPTp and ITN use. The association of district, age (categorical in 5 year increments), education (categorical variable including categories for no education, some primary school, completed primary school, some secondary school, and completed secondary school or above), socioeconomic status by wealth quintile assessed by wealth asset index, and where applicable, pregnancy category (currently pregnant, completed pregnancy in past 2 years, not pregnant in past 2 years), on the outcomes of interest was assessed.

A total of 7407 households containing 29,806 persons were surveyed. Of these, 6985 (23.4 %) were WOCBA, and 3213 of WOCBA (45.9 %) reported having completed a pregnancy in the 2 years prior to the survey. The mean age of all WOCBA was 28.1 years (95 % CI 27.8–28.3) (Table 1). Overall, 34 % of WOCBA completed a primary education or higher, but this varied by district with 18 % of WOCBA in Nkhotakhota District and up to 65 % of WOCBA in Rumphi District having completed a primary education or higher. Overall, 53 % of women reported sleeping under an ITN the night before the survey.

Among the women who reported having completed a pregnancy in the 2 years prior to the survey, ANC use was high in all districts with 96 % of women attending ANC clinic at least once during their last pregnancy (Table 2). Overall 72 % received two or more doses of IPTp with 57 % of women in Nkhotakota and up to 83 % of women in Rumphi having received two or more doses of IPTp.

After adjusting for age, educational level, socioeconomic status, and use of ITNs or untreated bed nets, women from Lilongwe [adjusted odds ratio (aOR) 2.0, 95 % confidence interval (CI) 1.2–3.1], Phalombe (aOR 2.5, 95 % CI 1.5–4.4), and Rumphi (aOR 2.5,95 % CI 1.5–4.3) Districts were more likely to receive two or more doses of IPTp than women in Blantyre District (Table 3). Women who had completed primary school (aOR 1.6, 95 % CI 1.0–2.6), some secondary school (aOR 1.9, 95 % CI 1.1–3.3), and secondary school or above (aOR 4.1, 95 % CI 1.9–8.7) were more likely to receive two or more doses of IPTp than women who had no education. Women in the poorest wealth quintile were significantly less likely to receive two or more doses of IPTp than women in the least poor quintile (aOR 0.68, 95 % CI 0.48–0.97). Use of either an ITN (aOR 1.2, 95 % CI 0.97–1.60) or untreated bed net (aOR 1.4, 95 % CI 0.85–2.4) the night before the survey was not associated with receiving two or more doses of IPTp.

In multivariate models adjusting for district, age, education level, socioeconomic status, and pregnancy status, women who resided in Nkhotakhota (aOR 0.52,95 % CI 0.39–0.69) and Phalombe (aOR 0.67, 95 % CI 0.47–0.95) Districts were significantly less likely to use an ITN the previous night than women who resided in Blantyre District (Table 4). Moreover, women aged 15–19 years (aOR 0.61, 95 % CI 0.45–0.83) were significantly less likely to use an ITN than women aged 35–49 years. Neither education level nor socioeconomic status were significantly associated with ITN use. Women who were currently pregnant (aOR 1.5, 95 % CI 1.2–2.0) or who had completed a pregnancy in the past 2 years (aOR 2.4, 95 % CI 2.1–2.8) were significantly more likely to use an ITN compared to women who had not been pregnant in the past 2 years.

Although the areas surveyed are geographically diverse, and contain 33 % of the population of Malawi, due to the purposive rather than probability-based selection of the 8 districts, the findings are not intended to be generalizable to all of Malawi. In addition, recall and social desirability bias may have influenced the responses to survey questions. However, where possible, responses to survey questions (i.e. ownership of assets including ITN) were confirmed by visual inspection. In addition, the survey team reviewed the ANC card to verify the receipt of SP during pregnancy. Where the ANC card was not available, interviewers probed to see if the woman took three big, white tablets at the health facility (indicative of SP). This survey did not assess the number of ANC visits made by each woman, nor the timing of these visits, inhibiting our ability to assess whether low IPTp uptake was influenced by inadequate ANC attendance.