Assessing supply-side barriers to uptake of intermittent preventive treatment for malaria in pregnancy: a qualitative study and document and record review in two regions of Uganda

For policies to be implemented effectively, they need to be communicated and interpreted in a consistent manner across the health system [18]. A review of national-level malaria in pregnancy documents in Uganda carried out in 2013 revealed that guidance with regard to timing and dosage of IPTp was often inconsistent, with some documents specifying the timing of doses in terms of weeks of pregnancy, while others use months and some do not specify timings at all [12]. It was also not clear whether the time windows specified for the provision of IPTp are mandatory or whether they merely indicate when IPTp should ideally be administered. Similarly, there was a lack of clarity as to whether the guidelines define two as the maximum or simply the recommended number of doses a woman should receive.

None of the district officials and health workers interviewed was aware of the existence of conflicting guidelines. Most stated that the guidelines are clear and easily understood. Only one district official implied that ambiguous or outdated guidelines may cause confusion among health workers over when to offer IPTp to pregnant women attending ANC:

District official 2: Then maybe one can receive that one who is at 36. What do I do? They told me 28. You are late. You have missed out.

Even though the majority of district officials and health workers were not aware of the issue, the existence of inconsistent IPTp guidelines is likely to contribute to missed opportunities for the provision of the treatment to pregnant women, as health workers may not provide IPTp to women they incorrectly believe not to be eligible or because they prefer not to provide it because they are unsure about whether a woman is eligible. Similarly, using guidelines that are not in line with current WHO recommendations means that women will not receive optimum protection from malaria in pregnancy. It has been shown that inconsistent, complex or outdated guidelines contributed to low IPTp coverage in Tanzania and Mali [19, 20].

The Republic of Uganda’s Ministry of Health aims to encourage male partners to accompany pregnant women to ANC. Many respondents across all types of respondents suggested that this can lead pregnant women to assume they will not receive services if they cannot attend with their partners. It was also reported by respondents from all target audiences that some women will delay attendance because their partners are unable or unwilling to accompany them:

Opinion leader 5: Sometimes when the pregnant women go without their partners, they are sent back and such a woman will refuse to go again unless the partner is there—and when the partners is far from her, the woman will not even step the health [facility] again.

Several opinion leaders reported that women who did not attend with their husbands had been denied ANC services by health workers. While none of the women interviewed confirmed this, several indicated that they had been treated differently depending on whether or not their partners had accompanied them. For example, they reported that women attending with their partners had been seen to first, while those attending without their partners had been made to wait:

Woman 5: The two visits I made with my husband we were attended quickly and we didn’t take long at the facility, but when I went alone, I took long, because those with partners are attended first.

Male involvement in maternal and child health is a crucial factor in improving maternal and child health services [21] and should be encouraged, but health workers should be discouraged from giving preferential treatment to women who attend with their partners, which may not be possible for many.

Despite established standard procedures, ANC service delivery appeared to differ from facility to facility in the units visited. While some health facilities were said to provide daily ANC services in line with national guidelines, health workers, women and opinion leaders reported or implied that other facilities only offer ANC on certain days of the week, focusing on other services like family planning or immunization on the remaining days:

Woman 16: They said each institution has different rules and regulations and at such a time they do other things, not to attend to mothers now.

District officials stated that this was particularly common in lower-level facilities, some of which reportedly do not offer ANC services at all due to a lack of funding or qualified staff.

Many opinion leaders pointed out that low staff levels frequently lead to midwives being unable to attend to all ANC clients on a given day and refusing services to some women:

Opinion leader 2: So these women from the village, they come footing up to the dispensary. Now, reaching there, they will get the nurses already selected twenty. The first twenty –

Opinion leader 1: Are the ones they handle.

Opinion leader 2: They will tell the majority, “Go back and come back next week.” […] Because now, for them they will always select the first twenty.

Several opinion leaders related that midwives sometimes refuse to attend to women they perceive to be inadequately dressed:

Opinion leader 2: You know like these village women, they come on their skirts, what – like that [indicates tight fitting skirt]. So these nurses want, “You come when you are putting on your maternity [refers to a wide, loose-fitting skirt conventionally worn by pregnant women]. You have a leso [a cloth wrapped around the waist], we will attend to you.”

It was suggested by both district officials and opinion leaders that some women may delay ANC attendance to avoid embarrassment or because they do not have the resources to buy a maternity dress, which may result in late or irregular ANC attendance and hence missed opportunities for the provision of IPTp. Non-availability or refusal of ANC services by health workers not only leads to women not receiving vital services, including IPTp, when they visit the health facility, but those practices are also likely to have a negative effect on uptake of ANC and IPTp in the longer term, as they will shape women’s and communities’ perception of ANC services.

Opinion leaders, women and district officials speculated that low job satisfaction among front-line health workers may contribute towards poor attitude and unprofessional behaviour among ANC staff. While district officials suggested that staffing levels were generally adequate, many health workers stated that there were too few staff in ANC and that midwives’ workload was immense. This is in line with reports of widespread dissatisfaction with working conditions among health workers in the literature [22, 23].

Many respondents from all target audiences stated that they had heard of cases of unprofessional behaviour on the part of health workers. They frequently commented on the effect of poor staff attitude on women’s perceptions of ANC, linking negative experiences to low uptake or late ANC attendance:

District official 1: Maybe also the other factor is—I cannot fear to mention this—is the attitude of the health workers—the ANC midwives. Their treatment to the – to the mothers. Also—possibly also can make them shy away.

Disrespectful and rude health worker behaviour was reported specifically in the case of women who miss ANC appointments, as well as for women who are perceived as initiating ANC too early or too late. Since those are particularly likely to miss out on the recommended number of doses of IPTp, health workers’ brusqueness with this type of client is worth noting. This reflects findings of previous studies conducted in Uganda [24], which cited instances of mistreatment of clients by health workers and found that clients’ satisfaction levels with health service are low. However, most of the women interviewed, regardless of whether or not they had been selected with the help of a health worker, stated that they had generally been treated politely and respectfully during their own ANC visits:

Woman 2: In this health centre, you are welcomed and treated well. They mind of your health and the wellbeing of pregnant women.

Many respondents referred to the provision of free mosquito nets, usually during a pregnant woman’s first visit, as a major incentive to attend ANC:

District official 5: But, you’ll realize that some of these indicators just improve—sometimes they just improve by themselves when a cer—especially when certain incentives are there. Like, recently, there is net—because of the net, the enrolment has started changing.

This was confirmed by many of the women interviewed:

Researcher: Is there anything else that encouraged you to go for ANC that first time?

Woman 3: Yes, I went to the hospital they gave me my net I came back to sleep under my net.

A number of respondents from one study district reported that mosquito nets for ANC had not been available for a number of months. The shortage was repeatedly linked to a universal net distribution campaign, which may have resulted in available nets being channelled away from ANC. Respondents expressed concern that this campaign could result in lower ANC uptake as receiving a net will no longer be seen as an incentive for attending ANC. Other respondents felt that the lack of nets for ANC could lead to reduced protection from malaria for some pregnant women, as the nets distributed through the campaign might be reserved for use by men in a context where pregnant women traditionally do not share a bed with their partners. Continuing the provision of incentives was one of the most frequently mentioned recommendations in terms of maintaining high levels of ANC attendance.

District health officials and health workers in both study regions reported that the supply chain had been significantly strengthened through the efforts of the Ministry of Health and implementation partners in recent years and, as a result, stock-outs of SP no longer occur. Women and opinion leaders also reported that drugs for IPTp, unlike many drugs for other purposes, were typically available in the health facilities. This was confirmed by the review of SP stock cards, which did not find stock-outs during the 6-month period under review in the public facilities visited. This is a very encouraging finding since shortages and stock-outs of SP had been identified as major bottlenecks to IPTp provision in many countries, including Uganda [11, 22, 25]. The findings from this research are in line with those of a recent comparative study in Ghana, Kenya and Malawi, which did not find evidence of drug shortages [26].

However, PNFP facilities do not seem to have benefited from supply chain strengthening efforts in equal measure. Under the current system, National Medical Stores, an autonomous government corporation tasked with ensuring drug supply in the public health sector, provides SP free of charge to public facilities, but private and PNFP facilities are required to procure SP from Joint Medical Stores, a not-for-profit wholesale enterprise that sells pharmaceuticals and equipment to the private health care sector. As funds to procure SP are often limited, frequent stock-outs of the drug were reported by health workers at the PNFP facility visited for this study. This was also confirmed more generally by district health officials for PNFP facilities in their district. Where SP is available, health workers at this facility described how they are forced to choose between two scenarios: either they require women to pay a fee for the provision of IPTp (against a national guideline which requires facilities to provide IPTp free of charge) or they provide IPTp for free to clients, but at a financial loss to the facility. Either scenario is likely to negatively affect IPTp uptake, particularly given the significant role of the private health care sector in Uganda [27].

A number of respondents emphasized the negative impact of non-availability of drugs on women’s and communities’ perceptions of ANC. In Eastern region, the term used for attending ANC in the local language literally translates as going to take the medicine, so in this context failure to provide medication may be equated with failure to provide good ANC in general:

Health worker 7: Because now our mothers are used – […] It motivates them that when you are going to a facility, there is that saying – they say that, “alosingo amat ekiya” [Ateso for ‘I am going to take medicine’], so if she goes away without any medicine, it will imply that her visit was not fruitful, so it demoralizes them.

District officials and health workers overwhelmingly reported that water, jerry cans, cups, and water purification tablets had been provided in sufficient quantities. The majority of the women interviewed in both study regions also reported that clean cups and water had been available when they were offered IPTp. This is another encouraging finding, as research carried out in Uganda in 2011 had found that, while 74 % of facilities had safe water ready for use in ANC, only 47 % had adequate cups [28]. Lack of safe water and cups has also been reported in studies in other African countries [22, 29].

Formal supervision or mentoring mechanisms for ANC staff were only reported by health workers at one of the eight facilities visited. More typically, health workers related that they can discuss complex cases with senior staff on an ad hoc basis if required. However, several health workers felt that such opportunities were rare:

Health worker 11: Actually I don’t know how I should say it has not really been so much there. It is just once in a while.

In terms of receiving formal classroom training on IPTp, health workers were evenly split: about one-third indicated they had received training on IPTp; one-third said they had not been trained on IPTp specifically, but IPTp had been covered as part of a broader training, for example on treatment of malaria; and, one-third stated they had never received any formal training on IPTp. There was a consensus among health workers and district officials that training is an important factor in ensuring quality of service delivery, as it enables health workers to refresh their knowledge and keep track of scientific progress. However, it was also pointed out that classroom training can be disruptive to service provision. A number of health workers expressed a preference for less disruptive training methods such as on-the-job training or mentoring:

Health worker 11: I think the training would be better on job here – not carrying people somewhere, but it would be better here just on job. Like it is an antenatal day, mothers are also there – we get mothers so – kind of mentorship I think it would be better.

Lack of supervision and training on IPTp is likely to contribute towards health workers’ poor understanding and knowledge of current IPTp guidelines. This confirms the results of several studies that identified a lack of effective training and individual supervision as barriers to the provision of IPTp [22, 30]. Conversely, training health workers has been shown to increase uptake of IPTp [31]. In Uganda, a review of relevant materials found that no pre-service education materials for midwives covering malaria in pregnancy or IPTp existed in 2013 [12].

When asked to outline the current guidelines for IPTp provision, it transpired that there was wide variation between health workers, particularly in terms of their knowledge and interpretation of the recommended timing of IPTp administration. Most health workers correctly specified that the first dose of IPTp should not be given in the first trimester, with 16 weeks mentioned by a majority of respondents as the lower limit. Week 20, 24, 26 and 28 were variously mentioned as the upper limit. There was similar variance in terms of the timing of IPT2, where the lower limits mentioned ranged from week 16 to week 34, with upper limits mentioned ranging from week 28 to 36. Very few health workers pointed out that IPTp can be given up to the time of delivery and two explicitly stated that IPTp should not be given during the later stages of a woman’s pregnancy.

Most health workers implied that they treat the time windows specified by the guidelines as recommendations only. However, several health workers suggested that the time windows are mandatory, which means opportunities for the provision of IPTp may be missed if women present for ANC at other times. Similarly, the majority of health workers appeared to interpret the current guidelines as specifying that a woman should not receive more than two doses. In light of current evidence that three or more doses of SP are associated with fewer low birth weight deliveries compared to the standard two-dose regimen [32], this means women do not currently receive optimum protection from the effects of malaria in pregnancy. Health workers’ confusion about timing and dosing has been cited as a barrier to IPTp provision by a number of studies in a range of African countries [33, 34]. Inconsistent and unclear policies and guidance [12] are likely to contribute to the inadequate knowledge and differing interpretations of the guidelines among health workers.

Observations made as part of the document and record review revealed that provision of IPTp appears to vary between facilities as well as between individual health workers. In some health facilities, pregnant women were generally not offered IPTp until after week 22 of their pregnancies even if they had initiated ANC earlier. Similar variations were found at the later stages of pregnancy, with women not being offered IPTp if they reported or returned after week 36. All the practices observed create missed opportunities for the provision of IPTp, which is considered safe from week 13 up to the time of delivery, despite previous concerns about the safety of administration of SP late in pregnancy [5].

All district officials and health workers demonstrated an understanding of malaria in pregnancy as a health risk to mother and child. They generally agreed that in order to maximize protection, pregnant women should take preventive medication in addition to practicing protective behaviour, such as sleeping under a mosquito net. While most district officials and health workers were broadly supportive of the concept of IPTp, some appeared to have reservations with regard to its efficacy and safety. This was based mainly on the observation that cases of malaria in pregnancy persist despite IPTp administration over several years:

Health worker 11: But about how efficient, I have some questions about it and am not really very sure if it is really efficient. As I told you before, much as mothers take it, they still suffer from malaria. So my concern is in that area how efficient it works.

A small number of health workers also expressed concerns about the side effects of SP, particularly if taken on an empty stomach. Several health workers mentioned that SP can also be used for the treatment of symptomatic malaria; however, a few pointed out that it is no longer used as a first-line anti-malarial. One interviewee indicated that the reasons behind discontinuing SP for the treatment of malaria while retaining it for IPTp were not clearly communicated, raising doubts with regard to the drug’s safety and efficacy. Two health workers in West Nile on the other hand indicated that they still use SP to treat symptomatic malaria because they feel it is more efficacious or better tolerated than other available anti-malarial drugs.

About half the health workers interviewed stated that they were not aware of or had never been confronted with safety issues or side effects related to IPTp administration. The other half indicated that women did occasionally experience side effects after receiving IPTp and there were also concerns over allergic reactions to SP. However, most of the health workers who voiced concern elaborated that side effects and allergic reactions were rare and typically mild:

Health worker 12: Well, there is this slight side effect of having maybe headache, dizziness—the dizziness is there. Maybe nausea—feeling like vomiting. But we have not got very serious side effects which is adverse effect—adverse —which is worse side effect. We have not got it. But this one of nausea—feeling like vomiting—abdominal pain—they complain. But it’s mild. It’s mild. Not very serious yet.

Only two midwives stated that the side effects can be serious, one of them because she herself had experienced them after taking SP.

In order to ensure health workers’ compliance with IPTp provision guidelines, it will be important to provide information about the efficacy and safety of SP. While none of the health workers interviewed stated that they do not provide IPTp because they doubt the drug’s efficacy and safety, it cannot be ruled out that this may contribute to low IPTp coverage.

All health workers interviewed reported that if SP is available and they have determined a client’s eligibility, they always offer IPTp. A majority of the women interviewed also indicated that they had been offered IPTp when they attended ANC. However, the three women who were interviewed because they had not completed a full course of IPTp related that they had not been offered the treatment by health workers on at least one occasion, despite regular ANC attendance. With regard to supply-side issues, this was the only noticeable difference between the responses given by the women interviewed during the two rounds of data collection. Health workers did not explain to those women who did not receive IPTp why the service had not been offered.

Researcher: What reasons were you told for not getting the tablets?

Woman 19: The nurse did not tell me why I didn’t receive the tablets.

Researcher: There was nothing you were told completely?

Woman 19: The nurse only examined me and there after she did not tell me the reason for not getting the tablets.

Note however, that self-reported administration of IPTp may be unreliable. A recent study in Uganda found that only about one-third of women who self-reported having used IPTp tested positive for SP in the blood stream [35].

Health workers and district officials described a number of scenarios where IPTp might not be provided. For example, it was suggested by one respondent that health workers might prioritize provision of IPT1 over provision of IPT2 when stock levels of SP are low. A district official speculated that some health workers might not provide IPTp if women initiate ANC late during their pregnancies—either to punish them for late attendance or because they think SP is not safe to be administered during the final weeks of pregnancy. Many health workers stated that they do not offer IPTp to women who report having ‘reacted’ to IPTp during previous pregnancies. However, from respondents’ examples, it appears they are referring to mild side effects rather than potentially life threatening allergic reactions:

Health worker 8: We got one who said in the previous pregnancy she took it and it really worsened her condition – as in she started vomiting – that thing – and you know, so we just – what we did was to give her the information – ok to give her the benefits of taking IPTp, but we didn’t force her to take.

In the context of poor knowledge of the IPTp guidelines and lack of adequate training, this might reflect a lack of knowledge of common side effects and contra-indications of IPTp, a challenge cited in a recent study in Ghana [36]. In order to maximize provision of IPTp, it will be important to educate health workers about differentiating between mild and serious side effects. This is particularly relevant as health workers frequently reported that women who are reluctant to take IPTp out of fear of side effects can be persuaded to accept the treatment if health workers provide information about the possible side effects and assure women that they are typically mild:

Health worker 9: Of course during health education talks they can ask you questions. In case when somebody develops side effects like rush and they—and in case somebody develop—feel sometimes nausea, but with the help of health information given to them, they accept.

The majority of health workers and opinion leaders reported that IPTp is routinely provided under the supervision of midwives as DOT. However, several health workers reported that they sometimes do not observe DOT, particularly if their workload is high or if women request to take the tablets home because they have not eaten and do not want to take the medication on an empty stomach:

Health worker 12: Well, practically what we do is when these people come for antenatal, we tell them to take it at DOT. […] But the challenge is some mothers—they have come minus having breakfast. So they fear taking the drugs minus food. We tell them, “No you can take and go home, then have food.” But some insists they are going to take it, they first want to eat then they take the drugs, which is also right.

This practice appears to be particularly widespread in West Nile. A district official and two facility in-charges in this region confirmed that they were aware that midwives sometimes ask women to take the tablets at home, particularly if they are busy. This practice is problematic as compliance with DOT has been shown to increase IPTp uptake [37]. Midwives should be educated about the benefits of DOT and should be reminded to strictly observe the guidelines.

Effective and consistent tracking and reporting of interventions helps to ensure policy makers, implementers, donors, and key stakeholders are kept informed of the progress towards achieving desired objectives, indicating the impact of current activities and identifying areas that require improvement [38]. While data relating to ANC and IPTp are routinely recorded in all facilities visited, the document and record review found that recording practices differed widely between facilities and individuals. For example, non-provision of IPTp is recorded using a variety of symbols and abbreviations, being variously denoted by a dash <->, a tick <✓>, <NT>, <ND>, or <N/A> in the IPTp column of the ANC register. Similarly, while <NA> stands for ‘not available’ in some cases, in others it is used to denote ‘not applicable’ or ‘IPTp completed’ (i.e., client previously received two doses of IPTp). Moreover, while a tick stands for ‘IPTp given at this visit’ in some facilities, in others it signifies ‘IPTp was given in the past’. Many health workers also reported that standard HMIS recording tools, such as maternal passports and ANC register books, are often out of stock, requiring staff to use poor photocopies or manually drawn registers instead, which may be inconsistent with the standard tools.

Several midwives indicated that their recording and reporting duties significantly add to their workload. However, most midwives interviewed commented that they view data recording and reporting as routine work, which does not pose any significant challenges:

Health worker 5: Yes—and also with our reporting. […] So it’s a daily thing—so it is something which is every day. So I think you cannot even forget it since you’ve started it and something which is in record that you are supposed to do it on daily basis whenever a mother—it’s a routine.

At the district level, aggregating facility-level data and reporting to the national level is the responsibility of an HMIS focal person or biostatistician. Facility-level data are entered into the District Health Information System (DHIS 2) online reporting tool, which connects district health offices with the Ministry of Health Resource Centre. The DHIS 2 data entry form automatically rejects certain entries as erroneous. For example, it does not allow the figure for provision of IPT1 to be greater than the one reported for ANC1 and the figure for provision of IPT2 must not be greater than the one reported for provision of IPT1 in any given month. However, all the scenarios rejected by the system are logically possible and do in fact occur quite frequently. Discussions with data managers at district level revealed that they resort to adjusting the reported figures to match the data entry system’s requirements when faced with this challenge, thereby falsifying the reports received from the facilities:

District official 7: Okay now, secondly, when we get it—like we—we do—we have ANC1—ANC first visit is 218. IPT1 is rated maybe 300. That one can’t happen. So we reduce that one to this [indicates act of replacing one figure for another].

Both district officials and facility health workers overwhelmingly stated that they believe facility and district-level data to be fairly accurate. There was, however, a clear mismatch between health workers’ perceptions and the results of the document and record review, which found that both under and over-reporting with regard to all indicators reviewed occurred in all facilities visited. In some cases, figures reported to the district varied from data recorded in the facility’s register by more than 500 %. For example, in July 2013, a health centre II recorded three ANC4 visits, but reported 25 (variance of 733 %). Similarly, in the same month, a hospital recorded five women who received IPT2, but reported 34 (variance of 580 %). While both under and over-reporting on a smaller scale of up to 25 % occurred very frequently, high variances in excess of 50 % tended to be cases where a much higher number was reported compared with what had been recorded in the register, maybe because health workers feel under pressure to report high uptake figures in an attempt to present their work in a positive light, as suggested by one district official:

District official 1: Then at the end of the month, you say, “Ah, what do I do?” That’s where I’m—the other—maybe the other issue of exaggerating data comes in. You can feel, “Ah, now, I need to be performing better. Well, my mothers have been—and the numbers have been ok. Then this time, what will they think, hm?” I think that’s what I—my mind takes me that maybe that’s where it sometimes—you get these figures interchanged.

Discrepancies between recorded and reported data tended to be greater at higher level than at lower level facilities, probably because of the larger data volumes handled by higher level units (see Fig. 3). In comparable facilities, discrepancies were found to be lower in Eastern than in West Nile, possibly as a result of the higher levels of implementation partners’ support in Eastern region. The poor accuracy and quality of data recording found through the document and record review reflect the conclusion of WHO’s assessment of health facility data in Uganda [15].

Where health workers acknowledged that, on occasion, data inaccuracies do occur, they typically ascribed them to high workload:

Health worker 3: Challenges when women are very many, you get tired and sometimes, recording, you need to record, you need to go for other things, examinations.

Poor data management skills and lack of training were also mentioned. Some respondents pointed out that non-compliance with DOT might obscure actual IPTp uptake figures, as women might not actually take SP, while health workers are likely to record provision of IPTp in the register. Many health workers noted that data accuracy did not appear to be seen as a priority, as the criteria used for assessing facilities’ monthly data reports at the district level are promptness and completeness. Districts do not commonly provide feedback to the facilities with regard to the accuracy of recorded data submitted.

The results presented above suggest that coverage rates recorded and reported should be treated with caution, as there appear to be sources of error and inaccuracy along the recording and reporting chain. In Uganda, poor data quality and reporting practices have been identified as important challenges to effective monitoring and evaluation of malaria interventions [11] and HMIS has been singled out as a priority area requiring improvement [39]. Furthermore, it should be noted that there is currently no system for collecting and integrating data from the private sector into HMIS [11].