Across all the sites, respondents recognized that sleeping under an ITN was a way of preventing malaria. Even though respondents at all sites offered additional explanations (such as poor hygiene) for a bout of malaria (or the local illness that overlapped with biomedically defined malaria (see  for further details), mosquitoes were reported to be the main cause. The connections that respondents made between ITNs use and malaria prevention were therefore unsurprising. Malaria was also viewed as a common disease for pregnant women, and considered to be a cause (along with other contributing factors) of miscarriage .
In addition, negative attitudes towards ITNs were rare and there were no specific objections to their use during pregnancy. This finding, in varied social and cultural contexts with different mechanisms of ITN distribution, contrasts with several previous studies that have highlighted health concerns linked to ITN use during pregnancy, particularly with regard to the impact of the insecticide treatments on the unborn child [22–25]. These studies were carried out prior to the distribution of long-lasting ITNs and it was often the insecticide, and process of re-treatment, that provoked such concerns. One multi-site study in Kenya highlighted additional negative attitudes towards ITNs including complaints about distribution campaigns targeting pregnant women and infants . Such comments were attributed to a lack of information linking increased malaria vulnerability with ITN distribution . By contrast, respondents in this study often highlighted pregnant women’s increased risk malaria  and no such negative comments about the targeting of pregnant women were encountered.
The broader meanings associated with ITNs had varying implications for their use. In central Ghana, women viewed ITNs as a consumer good and therefore left them unused during pregnancy, hanging them up for the first time to mark the birth of the child. Kenyan respondents also valued ITNs as a household item with multiple uses, such as protecting crops from birds. However, in this context, the multiple uses of ITNs did not necessarily prevent pregnant women from sleeping under them. These findings are a reminder of how health technologies (in this case, intended by its designers for malaria prevention) can take on quite different meanings and usages. This is particularly the case in resource-poor contexts where material goods are scarce and this scarcity can foster innovation.
More practical issues also had implications for ITN use during pregnancy. In Kenya, occasionally, sleeping arrangements together with the prioritization of children sleeping under ITNs prevented pregnant women from using the ITN that they received at ANC. In northern Ghana, ITN use was seasonal, depending on the presence of mosquitoes and temperatures low enough to sleep indoors. In Malawi, shortages of ITNs at health facilities limited the number of women who received free ITNs and their use. However, when available, Malawian women reported sleeping under ITNs.
As previous research has also highlighted [9, 15, 27], knowledge of IPTp varies across different contexts. In central Ghana and Kenya, pregnant women did not usually associate IPTp with malaria, whereas, in Malawi (as has also been identified in previous research ) and northern Ghana, it was more often linked to malaria, but not always prevention. Reports of side effects linked specifically to IPTp were therefore more prominent at these sites and, in northern Ghana, vomiting was particularly associated with IPTp, and, therefore, seen as one negative aspect of ANC. However, vomiting did not lead directly to future non-compliance with IPTp or discourage ANC attendance: often, regardless of side effects and without supervision women took their malaria prevention, “Fansidar” or “white tablets” because it was a component of the ANC package, and in an effort to follow the instructions of healthcare staff – see  for more detail of ANC at the same sites. Similar confidence in healthcare staff’s instructions about IPTp has also been identified in The Gambia  and Uganda . Even so, the women who did not adhere to IPTp – and there were suggestions that this occurred even if administered under DOT – did so largely because of previous negative experiences. The instances of IPTp non-compliance underscore a need for further in-depth (observational) research on compliance in real-world settings to ensure that uptake of IPTp is not overestimated.
Optimal treatment of MiP was hindered by a lack of available malaria diagnostic tests and by negative test results being ignored. Pregnant women, along with other community members, generally viewed healthcare staff as the authoritative source of malaria diagnosis, particularly severe malaria. Moreover, malaria tests were generally valued as a way of confirming the diagnosis. However, even if available, the tests did not entirely dispel uncertainty around the presence or absence of MiP. The non-specific nature of its symptoms has prompted recent policy recommendations to test for malaria prior to administering treatment . Pregnancy however further complicates the clinical diagnosis of malaria because there is overlap between the symptoms of non-severe malaria and what women often consider to be normal pregnancy experiences  or related illness . Although a positive test result resolved any uncertainty for healthcare staff and pregnant clients, a negative result was not so conclusive: in Ghana and Malawi, observations and women’s reports suggested that there instances of treatment in spite of a negative result.
Ghanaian healthcare staff asserted that pregnant women were exceptions to the policy of testing prior to treatment and provided treatment based on symptoms even when a malaria test result was negative. This is however a misinterpretation of the relevant policy document, which states that, in the absence of laboratory diagnosis, pregnant women with clinical symptoms of malaria should not be denied anti-malarials because the risk of not treating far outweighs the risks associated with overtreatment . A lack of confidence in malaria tests and reliance on symptoms is however well-documented in other African contexts [30, 31]: for example, in northern Tanzania, overtreatment of malaria in general was linked to a range of factors, including healthcare staff members’ assumptions about malaria being the most important disease and patients’ expectations . Overtreatment of MiP in Ghana was linked to a (mis)interpretation of national policy, yet the findings offer little insight into how this came about. However, the emphasis placed on MiP and its deleterious impact on the health of mother and child could have contributed to the better-safe-than-sorry approach. The reports of treatment in Malawi were however, more indirect. Healthcare staff only referred to treatment after a positive result, even if such tests were not readily available. Pregnant women who suspected malaria but received a negative test result were disappointed when they did not receive anti-malarials, but the data offer no insight into whether this led healthcare staff to provide treatment.
With regard to malaria treatment, pregnant women generally reported following the advice of healthcare staff. However, although scattered, there were instances of women ignoring such instructions or self-treating for malaria without seeking diagnosis at a health facility. In Ghana, where there were four cases of women who did not complete their treatment course, side effects played a role, as did confusion about the advice from healthcare staff, particularly if it contradicted ideas about malaria causation. In spite of such cases, health messages tended to focus on the use of non-prescribed and non-biomedical treatment during pregnancy and little emphasis was placed on adherence to prescribed anti-malarial regimens. The instances encountered in Kenya of women self-treating with drugs from other sources stood out from the other sites. Yet, although similar findings have been made at another Kenyan site , it is unclear to what extent they were indicative of a more systematic trend of self-medication. At all the sites, women made use of paracetamol and other cold remedies to combat the mild symptoms associated with pregnancy and/or malaria. Such vague ideas of “malaria” medication (presumably linked to the breadth of the local illness that overlaps with malaria ), can contribute to inappropriate use of anti-malarials and other “malaria drugs” through self-treatment or non-adherence to prescribed treatment regimens.
A preference for anti-malarials that “suit” an individual was more prominent in Malawi than at the other sites. Such ideas about compatibility were perhaps the most extreme examples regarding the role that personal experience plays in attitudes and behaviours towards MiP interventions. Because pregnancy is generally considered to be a particularly vulnerable bodily state and one in which women experience a range of health complaints , it is unsurprising that women display a preference for health interventions that are viewed as not contributing further to the negative symptoms of pregnancy. Indeed, these experiences can – albeit rarely – override women’s typical adherence to the instructions that healthcare staff provide.
Strengths and limitations
The strengths of this study are intertwined with the anthropological approach: fieldwork over a one to two year period enabled observations to be carried out to triangulate the data that respondents shared with the research team and enabled women to be interviewed multiple times over the course of their pregnancy to develop rapport, cross-check previous responses and to monitor their experiences of care over the course of a pregnancy with a follow-up post-delivery. However, the findings, with regard to malaria interventions, are limited by several factors. Regarding ITN use, reported use could not be confirmed with observational data at all the sites: as a result of the organization of dwellings and the restrictions on researchers’ access to sleeping spaces, only in Kenya was it regularly possible to observe the presence of ITNs in women’s sleeping quarters. Also, by relying on women’s reported malaria treatment practices – because it was not feasible to carry out direct observations of women’s drug intake outside of the health facility – malaria self-treatment and non-compliance with anti-malarial treatment courses may have been underestimated across all the sites. The lack of systematic observations of drug intake makes the cases of self-treatment and non-adherence all the more striking and highlights a need for further systematic analysis.