Malaria knowledge and experiences with community health workers among recently pregnant women in Malawi

The survey was conducted in the districts of Ntcheu (pop—270,903), Central Region, and Nkhata Bay (pop—206,670), Northern Region, Malawi (Fig. 1). These districts were purposively selected from among the 10 districts in Malawi where the U.S. President’s Malaria Initiative (PMI) is supporting malaria control activities (out of 28 districts in Malawi) to ensure representation of two different regions. Both Ntcheu and Nkhata Bay are rural districts where most families survive on subsistence farming, with a small tourism economy serving Nkhata Bay district as well.

Ntcheu has almost twice as many health centres and ANC providers as Nkhata Bay (40 vs 22 and 55 vs 29, respectively), but only about 30% more pregnant women (13,544 vs 9905 annually) (Table 1). Nkhata Bay has more CHWs than Ntcheu; thus, Ntcheu has fewer CHWs per facility and each CHW, on average, serves a larger population. Although Ntcheu has fewer CHWs per facility, there are additional volunteers in the community called Secret Mothers. These volunteers receive minimal training and their primary function is to encourage women to attend ANC in the first trimester. The median age of marriage for women in both districts is 18 years old, while the median age at first birth is 19 years old. Women in both districts are equally likely to deliver at a health facility and with a nurse or midwife present [8].

The population of interest was recently pregnant women, defined as women between the ages of 16–49 years who had a pregnancy resulting in a live birth in the previous 12 months. A three-stage cluster sampling procedure was used to select survey respondents. After excluding District Hospitals, non-governmental facilities, facilities that do not provide ANC, and facilities accessible only by boat, 10 health centres were randomly selected in each district (20 in total). First, the catchment area of each health centre was designated as a cluster. Depending on geographical size, each cluster contained 3–18 enumeration areas (EAs). EAs are administrative data collection units, demarcated by the National Statistics Office, with an average of 250 households or 1000 people. In the second stage, a single EA was randomly selected from each cluster. Finally, all households within the selected EA were listed, noting households with recently pregnant women. Simple random selection was used to select 20 households with recently pregnant women and 20 households without recently pregnant women. If a selected household had more than one recently pregnant woman living in it, all women who met the inclusion criteria were included. EpiSample (PATH, MACEPA Developer Products, Seattle, WA) was used to map the households and sample participants.

It was estimated that a minimum of 300 recently pregnant women, 150 per arm, would be needed to achieve 80% power to detect a 30-percentage point increase in IPTp3 coverage from 30% at baseline to 60% at endline, with 0.05 significance level, assuming an intra-cluster correlation of 0.2, using a two-sided Z test (unpooled).

Twenty-eight (28) enumerators took part in the baseline household survey. All enumerators participated in a 2-day training followed by 1 day of piloting the survey in one of the non-selected EAs, and then a day of debriefing, during which minor updates were made to several questions to improve the readability and clarity.

The topics covered in the training included basic information about malaria and detailed instructions on how to map and administer the survey, including obtaining consent. The questionnaire was reviewed in both English and Chichewa (the local language) to ensure that the enumerators had good understanding of the overall organization, structure, and purpose of the questionnaire and to verify that the Chichewa questions were true reflections of their English counterparts. The survey was administered verbally as a structured questionnaire and questions were predominantly open response, allowing for multiple responses to be recorded. Whenever possible, ANC cards were used to validate women’s recall on the number of ANC visits attended and doses of SP received during their most recent pregnancy. Data were collected from December 1–16, 2017.

Surveys were conducted using Open Data Kit (ODK) Forms hosted on the SurveyCTO (Dobility, Inc., Cambridge, MA) platform. Files from SurveyCTO were exported into SAS V9.4 (SAS, Cary, NC) and STATA (SATACorp, College Station, TX) for analysis. The database was programmed with range checks and key fields were required. Data were checked for completeness by field supervisors before uploading. In addition, data checks for key variables were run on data downloaded from the SurveyCTO platform.

Descriptive summary statistics including means, proportions, and corresponding 95% confidence intervals (CI) for social-demographic attributes, out of pocket expenditures, and time required for ANC attendance for women in the two districts were calculated. Households were grouped into wealth terciles based on reported monthly incomes; women with no monthly income were considered “poor” while women with any income were grouped into middle and high categories based on the median income. A malaria knowledge index was created to categorize women’s overall knowledge. Women were awarded one point for each of the following: correctly identifying (1) two or more health problems in pregnancy, (2) two or more consequences of malaria in pregnancy, (3) mosquito bites as the primary way to get malaria (with no incorrect methods, i.e., dirty water, identified), and (4) both IPTp and insecticide-treated nets (ITNs) as malaria prevention in pregnancy. Women with a score of 2 or less were categorized as having poor knowledge, while 3 or more was considered adequate knowledge. A logistic regression model using generalized estimating equations (proc genmod) accounting for the effects of district, age, gravidity, and facility-level clustering was used to assess the impact of knowledge on IPTp uptake. Statistical significance was assessed using t-tests for differences and means, and Chi square tests for proportions between the districts, for normally distributed data. Non-parametric tests were used for skewed data. P-values less than 0.05 were considered statistically significant. All analyses accounted for selection weight and clustering of households at facility level. Analyses were done using SAS V9.4 (Cary, NC). A comparison of the two districts was undertaken to explore heterogeneity between the districts.

The protocol was reviewed and approved by the Malawi College of Medicine Research Ethics Committee (COMREC); the Centers for Disease Control and Prevention Human Subjects Office determined that CDC staff were not engaged in human subjects research. Representatives from the Government of Malawi Ministry of Health were involved throughout the design and implementation of the survey, and permission was obtained from the officials in each District Health Office prior to initiating the survey. In each EA, permission was obtained from the village leaders. Written informed consent was obtained from each respondent before data collection; participants were told that this survey was focused on antenatal care and malaria prevention in pregnancy.

A total of 370 recently pregnant women responded to the survey; 179 in Nkhata Bay and 191 in Ntcheu. There were no refusals. Overall, the women in the two districts were similar across socio-demographic features, attitudes, perceptions, and ANC utilization. The median age of surveyed women was 23 (range: 16–44); 33.0% of respondents were primigravid, 25.6% were secundigravid, and 41.3% were multigravid. Only a quarter of women had received any secondary education; women in Nkhata Bay were more likely to have received secondary education than in Ntcheu (30.7% vs 21.5%, p = 0.05). The majority of women in both districts were considered poor based on household incomes (Table 2).

Malaria was the most frequent response given by recently pregnant women when asked to list serious problems in pregnancy (43.4%), with women in Nkhata Bay being more likely to consider it a serious problem (51.2% vs 41.2%, p = 0.02). Women in Nkhata Bay were also more likely to acknowledge maternal death and abortion/miscarriage as possible effects of malaria in pregnancy (55.4% vs 42.6%, p < 0.001 and 48.3% vs 29.1% p < 0.0001, respectively). A majority of women in both districts cited sleeping under an ITN as a way to prevent malaria, with no difference between districts (85.6% vs 85.2%, p = 0.92), while a minority cited IPTp, with women in Nkhata Bay more likely than women in Ntcheu to cite taking IPTp as a method of malaria prevention during pregnancy (41.0% vs 24.8%, p = 0.02) (Table 3).

Women in Nkhata Bay were more likely to cite mosquitos as a cause of malaria and to know that both ITNs and SP are a prevention strategy for malaria in pregnancy (89.9% vs 76.6% p < 0.001 and 29.3% vs 15.5% p < 0.0001, respectively) (Table 3). However, women in Ntcheu were more likely to know that mosquito bites are the only way to get malaria and less likely to attribute malaria to a dirty environment (70.6% vs 62.0% p = 0.03 and 11.3% vs 21.7% p = 0.005, respectively). Despite this, women in Ntcheu were not any less likely to report sleeping under a bed net every night compared to women in Nkhata Bay (94.8% in both districts, p = 0.99). Women’s malaria knowledge overall, based on the composite malaria score, was low and did not vary significantly between the two districts (only 36.1% and 33.1% had adequate knowledge, p = 0.64). Adjusting for district, mother’s age, and gravidity, women who had adequate knowledge were more likely to receive IPTp3 than women with poor knowledge, though this was not statistically significant (adjusted odds ratio = 1.50, 95% confidence interval 0.97–2.32, p-value 0.066)). A similar proportion of women in the two districts reported having had at least one case of malaria during their most recent pregnancy (36.3% vs 27.6%, p = 0.06) (Table 4).

The majority of recently pregnant women in both districts found the CHWs in their communities to be generally helpful (77.9%). In both districts, the most frequent responses as to why women liked the CHWs were because they provide good care and are easily available (46.4% and 32.2%, respectively), but more women in Nkhata Bay reported liking CHWs because they provide good care (60.3% vs 43.4% p = 0.02). In both districts, CHW accessibility and supply availability were the most commonly cited areas needing improvement (40.6% and 15.3%, respectively) (Table 5).

Although few women (8.5%) mentioned discussing pregnancy as a health topic with the CHW in their community (Table 5), one-third of the women (35.7%) reported talking to a CHW during their most recent pregnancy. Among these, half (49.3%) talked to their CHW two or more times during their most recent pregnancy. In Ntcheu, conversations with CHWs were most likely to take place during a home visit (47.7% vs 19.9% p = 0.002), while women in Nkhata Bay were most likely to speak with CHWs at the health facility (47.5% vs 16.3% p = 0.02). Overall, the three main topics discussed during interactions with CHWs were: diet during pregnancy (38.6%), the plan for delivery (36.1%), and when to attend ANC (35.4%). Sleeping under an ITN was the most common advice given to pregnant women during CHW visits (92.0%); only 28% of women reported being advised by the CHW to take SP during pregnancy (Table 6). Less than one quarter (21.9%) of women in Ntcheu and 42.7% of women in Nkhata Bay reported that they or their family had received assistance from a CHW for malaria (p = 0.01) (Table 5).

Essentially all women attended at least one ANC visit (99.6%) and half (49.8%) attended four or more ANC visits; 77.5% attended three or more visits in the second and third trimesters based on a review of health passports. According to self-report, only 15.6% of women initiated ANC within the first 12 weeks of pregnancy, as recommended by the WHO. Ninety-four percent of women first disclosed their pregnancy in their first trimester, with their husband being the most common confidant (82.9%) (Table 4).

Women in Nkhata Bay were more likely to receive three or more doses of SP than women in Ntcheu (59.2% vs 41.8% p = 0.0002) (Table 4). Among those who attended at least one ANC visit, the vast majority had positive experiences, with 91.5% rating the care received at ANC as good, very good, or excellent (Table 7).

One-third of women (35.3%) travelled less than 1 h to reach ANC, another one-third (31.5%) travelled from 1 to 2 h, 19.1% travelled 2–3 h, and 14% travelled three hours or more. The majority (86.8%) of the women surveyed reported zero cost associated with traveling to ANC; 78.9% reported that they walked and 8.6% rode a bicycle. The majority of women surveyed (65%) reported at least one obstacle to receiving malaria treatment while pregnant. The most commonly reported obstacles were stock-outs of malaria tests and treatment at health facilities and the distance to treatment (15.9% and 15.0%, respectively).

As with most surveys, there are limitations due to potential recall and social desirability biases. Participants may not have recalled exact dates or interactions with health workers. Further, because the respondent knew that this survey was being administered in preparation for a malaria intervention, it is possible that responses were exaggerated or altered if the respondent perceived that this would increase their chances of receiving a benefit from the subsequent study. It is also possible that women responded with what they perceived were the “correct” responses. In addition, because the survey focused on recently pregnant women and did not include ANC providers, it is not possible to fully understand the causes of the gap between ANC attendance and IPTp coverage. The results presented here may not reflect the situation in other districts in Malawi. However, there is no specific reason to suspect that the opinions of the community about CHWs in other areas would vary substantially.