Introduction

Approximately 30 million pregnancies occur each year in malaria endemic areas of sub-Saharan Africa [1]. Severe complications of malaria during pregnancy include cerebral malaria, maternal anemia, and maternal mortality, which tend to be more frequent during epidemics and among primigravid and/or immunocompromised pregnant women [2], [3]. Complications affecting the fetus or newborn may arise from either clinical malaria or asymptomatic parasitemia during pregnancy and include miscarriage, stillbirth, low birthweight, preterm delivery, and neonatal mortality [2], [3].

The World Health Organization (WHO) guidelines for the prevention of malaria during pregnancy include 1) use of intermittent preventive treatment during pregnancy (IPTp) with sulfadoxine/pyrimethamine (SP); and 2) sleeping under an insecticide-treated bed net (ITN) [4]. IPTp-SP is defined as provision of treatment doses of SP to asymptomatic individuals living in malaria endemic regions, regardless of malaria parasitemia status, and the current recommendation is that at least 2 doses of SP should be administered after the first trimester during antenatal care (ANC) [5].

Use of IPTp is estimated to reduce the occurrence of low birthweight by 42%, neonatal death by 38%, placental malaria by 65%, and antenatal parasitemia by 26% [6]. Even in areas where SP monotherapy for symptomatic malaria results in up to 25% treatment failures, 2 doses of IPTp with SP continued to provide considerable benefit to HIV-negative semi-immune pregnant women [7]. Despite the effectiveness of IPTp, and the nearly universal adoption of a national IPTp policy among malaria endemic countries [8], its use remains relatively uncommon in sub-Saharan Africa. Furthermore, data on IPTp coverage from national surveys remain limited. During 2007-2008, only 9 high-burden countries had national survey data on IPTp, resulting in only 20% of pregnant women who received 2 or more doses of IPTp [8]. In Uganda 37% of women reported receiving at least one dose of SP to prevent malaria during pregnancy, and only 18 percent received two or more doses [9].

Demographic factors associated with use of IPTp have been evaluated in several studies [10], [11], [12], [13], [14], although associations with education and age have been inconsistent across studies [10], [11], [14], [15]. A lack of association between IPTp use and socio-demographic factors household wealth, knowledge of malaria, travel times to the ANC clinic and the number of ANC visits has been reported [10], [11], [12], [14], [15]. Operational barriers affecting uptake of IPTp include: 1) imprecision in the estimation of gestational age leading to missed doses [16], [17]; 2) confusion regarding the timing of doses and/or the recommended number of SP doses [16], [17], [18]; 3) late or no antenatal care attendance [16], [17], [19]; and lack of potable water and/or drinking cups (for directly observed SP administration) [16], [17], [18].

Utilizing a population-based sample, we sought to describe the use of SP during pregnancy in one health district in Uganda and to identify determinants of use of IPTp. Additionally, we sought to describe the administration practices of SP during visits to the ANC clinic.

Methods

Results

Between November 2008 and January 2009 we visited 629 households to identify 500 eligible women (Figure 1), none of whom declined to participate. Seven of the index pregnancies ended in stillbirth. Use of ANC was nearly ubiquitous with 94.2% of women reporting 2 or more visits, and 90.8% reporting 2 or more ANC visits after the first trimester. Only 34.0% completed the recommended 4 or more visits during pregnancy. Most of the women delivered in a health facility (73.4%), and 16.2% delivered at home. Symptomatic malaria during pregnancy was self-reported by 66.8% of the participants. The characteristics of women with at least two qualifying ANC visits were similar to those of the entire cohort and are shown in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Eligibility and participation of study participants.

*Reflects completed pregnancies lasting at least until the beginning of the 3^rd trimester, regardless of pregnancy outcome. **Most recent pregnancy occurred more than 12 months from the interview date due to a lag time between the start of the census and the start of the study.

https://doi.org/10.1371/journal.pone.0015066.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of the study population.

https://doi.org/10.1371/journal.pone.0015066.t001

IPTp-SP use during pregnancy among women with 2 or more qualifying ANC visits

While the proportion of women who received 2 or more doses of SP was 44.8% (224/500), only 31.3% (86/275) of women received a full 2-dose course of IPTp-SP according to the Ugandan policy for the prevention of malaria (Table 2). A partial course of IPTp was used among 28.7% of women. The number of women receiving no doses of SP in the 2^nd and/or 3^rd trimesters despite qualifying ANC visits was 40.0%.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Doses of IPTp-SP taken during pregnancy among women with 2 or more qualifying ANC visits.

https://doi.org/10.1371/journal.pone.0015066.t002

Among the 168 women that were offered IPTp, 164 (97.6%) of them took the dose of SP. Only 2 women used SP for prevention despite never being offered it during an ANC visit, and only 4 women who were offered SP for prevention during an ANC visit chose not to use it.

Predictors of IPTp-SP among women with 2 or more qualifying ANC visits

Individual-level factors associated with receipt of a full-course of IPTp-SP compared to no doses were assessed among women with at least 2 qualifying ANC visits controlling for the total number of ANC visits (Table 3). Receipt of a full-course of IPTp-SP was relatively more common among women living in a rural village compared to a peri-urban area (RR: 2.73; 95% CI: 1.50, 4.99), those with the capacity to decide if SP should be used during pregnancy (RR: 2.28; 95% CI: 1.48, 3.49), and those who were less knowledgeable about the safety of SP use during pregnancy (RR: 1.87; 95% CI: 1.40, 2.49). Furthermore, women with lower educational attainment were more likely to receive a full-course (RR: 1.56; 95% CI: 1.03, 2.38), as were women living more than 30 minutes walking distance to the ANC clinic (RR: 2.06; 95% CI: 1.23, 3.46). No differences were found between other socio-demographic factors, pregnancy history, or socio-cultural factors.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Individual-level factors associated with full adherence to IPTp-SP recommendations among women with at least 2 qualifying ANC visits.

https://doi.org/10.1371/journal.pone.0015066.t003

Individual-level factors associated with receipt of a full-course compared to a partial-course of IPTp-SP were assessed among women with at least 2 qualifying ANC visits controlling for the total number of ANC visits. Women receiving a full-course of IPTp-SP had slightly less knowledge of malaria and SP compared to women receiving a partial-course (data not shown). No differences were found between these groups and the remaining socio-demographic factors, pregnancy history, and socio-cultural factors (data not shown). Among women who received a partial course of IPTp-SP, the main reasons given for not having taken a full-course of IPTp were “I didn't know about it” (49.3%), and “it wasn't offered” (34.7%).

Discussion

In the present study of 500 recently pregnant women in Jinja, Uganda, less than one-third of women received a full-course of IPTp-SP despite the high utilization of ANC in this cohort. While IPTp is a relatively simple intervention to administer, missed opportunities were common in our study. This is one of only a few studies to investigate correlates of IPTp use from the perspective of the user.

Our definition of IPTp-SP was restricted to preventive doses of SP administered after the first trimester of pregnancy, in accordance with the IPTp policy guidelines. Including treatment doses of SP taken after the first trimester in the analysis of IPTp would lead to an overestimate of the prevalence of IPTp due to the large proportion of women receiving SP for the treatment of malaria. The means of assessment of uptake of IPTp has varied across studies and disaggregation by presence or absence of symptomatic malaria has not been standard practice. Some studies explicitly stated that preventive doses of SP were presumptive (when the woman was not sick with malaria) [10], [30], [31], [32], although timing of doses were not specified. Other studies referred only to the number of doses of SP without reference to the womans' clinical state [11], [14], [33], [34], [35]. One study defined IPTp based on number of doses of SP given at ANC irrespective of signs or symptoms of malaria [18]. Appropriate IPTp has been defined in one study based on the timing of SP use during pregnancy (at least one dose in the 2^nd trimester, and at least one dose in the 3^rd trimester), although this study did not indicate if doses were restricted to preventive use [12], while another study defined IPTp based on 3 presumptive doses starting after 16 gestational weeks, with subsequent doses being indicated at intervals of 1 month or more [16]. Surveys such as the DHS ask, “During this pregnancy, did you take any drugs to keep you from getting malaria?”, and only if the women answer “yes” will they specifically be asked if they took SP, and the number of times. Standardizing the approach to defining IPTp with SP, or other drugs, would facilitate measurement of utilization across studies.

Among women in our study, the predominant factor predicting preventive use of SP during pregnancy was being offered IPTp during an ANC visit. Stock-outs of SP were seldom responsible for missed doses; only during 5.6% of ANC visits (61/1084) was a woman told that the clinic was out of stock of SP and they should purchase it on their own. The majority of women were either offered SP and used it, or SP was never mentioned during their visit and they did not use it. Small numbers precluded an assessment of individual-level factors associated with use of IPTp independent of being offered SP for prevention. Only 2 women used SP for prevention despite never being offered it during an ANC visit, and only 4 women who were offered SP for prevention during an ANC visit chose not to use it. Our results suggest that use of IPTp was low among women in our study because providers failed to offer SP. These findings highlight the importance that health workers play in delivering IPTp and the need to explore barriers to offering this therapy.

To our knowledge, this variable “being offered” has only been assessed in two studies, both of which reported not being offered SP during ANC was the reason for not completing a full-course of IPTp-SP [15], [36]. Confusion among health care workers regarding the timing of doses and/or the number of doses has been reported in several studies, and may be the underlying reason a dose of SP was not offered during a qualifying visit [16], [17], [18]. Improvements in the administration of IPTp following focused training of health care workers was demonstrated in one intervention study [18] supporting the notion that unclear guidelines are adversely contributing to the uptake of IPTp. Further uncertainty regarding administration of IPTp may also be related to an omission in the WHO IPTp guidelines regarding how treatment for symptomatic malaria (either presumptively treated or laboratory confirmed) might alter eligibility for or timing of the IPTp. The majority of women in our study had complex histories related to self-reported malaria, and use of antimalarials for treatment of malaria was high during the index pregnancy. This led to difficulties in classifying a woman's SP use during pregnancy as being consistent or inconsistent with the recommended guidelines. For example, if a woman is sick with malaria-like symptoms and receives SP as a treatment dose in the 2^nd or 3^rd trimester, the guidelines did not specify if this dose should be counted as IPTp. Or, if a woman is sick with malaria-like symptoms and treated for confirmed or presumed malaria with non-SP antimalarials, the guidelines did not specify when she should receive SP as presumptive treatment.

We were unable to evaluate why health care workers offered a full- or partial-course of IPTp-SP to some women, and failed to offer IPTp-SP to forty percent of the remaining participants with at least two qualifying visits. While we were able to explore individual-level predictors for having received a full-course of IPTp-SP compared to no doses among women with at least two qualifying ANC visits, these exposures were a poor proxy for health care worker behavior. In addition to ANC attendance, individual-level factors associated with having received a full-course of IPTp included living in a rural residence, being less knowledgeable about the safety of SP use during pregnancy, having less education, living further from the ANC clinic, and the respondent being the household member who decides if SP will be used during pregnancy. The findings across the various studies, including ours, are not consistent with regard to these potential influences on receipt of IPTp [10], [11], [12], [14].

Similar to our findings, the primary reasons for not completing a full 2-dose course of IPTp among women who had received only 1 dose included not being given IPTp from the ANC and lack of awareness about the 2-dose schedule [36]. The lack of individual-level factors strongly associated with use of IPTp-SP in the literature, combined with our finding of the importance of having been offered SP during an ANC visit, suggest that factors related to health-care provider behavior are more influential in the uptake of IPTp than most measurable individual-level factors. Evaluations of health care provider practices regarding administration of IPTp are needed to measure the extent of the problem and develop targeted interventions at improving access to IPTp-SP during ANC visits. However, in light of the growing drug resistance to SP [37], [38], its continued use for IPTp may eventually be replaced by more effective alternative drugs which may require alternative dosing schedules.

This study is potentially subject to the following types of bias and limitations. Women who know IPTp is desirable and recommended may be relatively more likely to falsely report taking SP for prevention when they did not. However, based on the very high concordance among self-reported SP and SP recorded on the ANC card, the magnitude of this sort of social desirability bias is likely to be minimal. Additionally, we did not have information on medical contraindications to SP, such as history of sulfa drug allergies, or daily use of cotrimoxazole for the prevention of HIV-associated infections. If these factors were highly prevalent in our populations, misclassification would lead to an overestimate of the prevalence of missed opportunities, and the prevalence of appropriate SP use would be underestimated. Furthermore, suspected malaria in this area is treated presumptively, despite information which suggests relatively few cases which present with malaria-like symptoms actually have clinical malaria [39], [40], [41], [42]. If a large proportion of the women misclassified their reason for using SP, then our measure of IPTp could be biased in either direction. Lastly, we were unable to evaluate why health care providers failed to deliver IPTp for apparently eligible women.

The strengths of this study included use of a population-based random sample, validation of self-reported use of SP and ANC visits through an assessment of the ANC cards, and use of several visual aids throughout the interview to reduce recall bias. The visual aids included a pregnancy history calendar in which each pregnancy was mapped out over time on the calendar to record episodes of self-reported malaria, any use of SP or other antimalarials during pregnancy, and ANC visits, and photographs of commercial packaging for the most common formulations of SP available in the area, and the corresponding tablets.

In summary, we evaluated the level and determinants of use of IPTp during pregnancy. Our results indicate that few women in the Jinja district of Uganda were receiving the full 2-dose course of IPTp according to the recommendations, and that women are willing to take preventive SP in pregnancy if it is provided during an ANC visit. Missed opportunities to administer SP during ANC were common in our study, as was administration of SP in a manner that is not directed by the Ugandan guidelines, suggesting provider level improvements are needed. Improving healthcare providers' knowledge about the proper use and administration of SP may be an effective intervention to improve uptake, as well as simplifying the guidelines to provide IPTp-SP at each scheduled ANC visit after quickening.

Acknowledgments

We would like to thank the women of Namizi and Kibibi parishes who participated in this study; the team working in the field: Juliet Naigwe, Teddy Naukenge, Sarah Kisakye, Sarah Kamali and Nasur; and the Uganda Malaria Surveillance Project.