Quantitative data collection.

The study involved a quantitative health facility assessment based on adapted “health facility service availability and readiness assessments” [13]. This assessment had three different modules: one for anaemia, one for postpartum haemorrhage management, and one for blood transfusion—the one for anaemia is presented here. This module asked a comprehensive set of questions to determine the availability of appropriate standards or protocols, drugs, equipment, trained staff, and infrastructure to support anaemia prevention and management. Certain practices that are documented—or were expected to be documented—in patient files were also captured, for example, the measurement of haemoglobin at the time of labour. For such measures, we typically assessed all relevant patient files in the preceding three months. Within each health facility, different questions within the module were administered to maternal health care providers, laboratory technicians, or pharmacists as relevant, and where needed, observation (e.g. physically viewing a protocol) took place. No processes of care, however, were observed.

Qualitative data collection.

Qualitative data were derived from community focus group discussions (FGDs), in-depth interviews (IDIs), and key informant interviews (KIIs). FGDs were held in each district with: currently pregnant or recently delivered women; male partners of currently pregnant or recently delivered women; community leaders and elders (referred to hereafter as “community influencers”); and blood donors. FGD participants were recruited from communities within the catchment area of study sites in the health facility assessment. Study information was shared with community leaders, who then advertised it and helped identify prospective participants meeting study inclusion criteria in their communities to ensure 6–10 eligible persons would be available per focus group. Other than the provision of refreshments and transport reimbursements, participants were not unduly incentivised to participate. FGDs explored perceptions about anaemia (how it is locally understood, what causes it, what its consequences are, how it can be prevented or managed) and ANC (why women do or do not attend, what occurs during ANC, its importance, and its role in anaemia prevention and management). IDIs were conducted with maternity in-charges and the heads of the participating health facilities to understand perceived facilitators and barriers to good practice in anaemia prevention and management. Laboratory technicians were also involved in IDIs to understand testing for anaemia and ordering blood for transfusion. There were typically only one or two people fulfilling each role in each participating facility, and we identified and interviewed them where possible. KIIs were carried out with key district- or national-level informants who play a role in supporting maternal and/or blood transfusion services in each participating district, including district health officers, district heads of maternity/reproductive health services, district heads of laboratory services, and regional or zonal blood transfusion leads. IDIs and KIIs were complementary to the health facility assessments, serving also to explain any gaps identified. All participants were purposively sampled based on their role in their community, health facility, or district and their lived experiences. Each participant participated in only one interview or group discussion.

FGDs took 60–90 minutes to complete, and IDIs and KIIs took 20–60 minutes to complete and were audio recorded. FGDs took place in convenient locations in each community where participants were derived from. To minimize the ask on participant time and travel, IDIs took place in the health faciltiies where the assessments were carried out with the relevant members of staff and KIIs were carried out in the offices of the key informants. All IDIs and KIIs were carried out in English, whilst FGDs were carried out in local languages as appropriate (Twi or Ga in Ghana and Runyankore in Uganda). All qualitative data were collected by skilled research assistants. They were registered nurses trained in maternal health, and in Ghana, they were also supported by staff from the National Blood Service for aspects pertaining to blood transfusion. They had at least three-years’ of experience conducting health research within the study districts.

Our study team included clinicians specialising in obstetrics, haematology and blood transfusion, nurses, and social scientists, all with extensive experience in low- and middle-income country health services research. We developed our data collection instruments to explore some of the themes identified in the qualitative review of postpartum haemorrhage prevention by Finlayson et al [14] and the general processes described in the protocol by Akter et al [15]. We tailored them to the purpose and context of our study. Since tools needed minimal revision following piloting, we included data from the pilots in our analysis. Our data collection instruments can be found in “S1 Data Collection Instruments”.

Analysis.

Quantitative data were analysed in Excel to generate basic descriptive statistics (counts, averages, percentages). Qualitative data were read and re-read for familiarity. Framework analysis [16] was used to generate higher-level categories around specific pre-determined aspects of anaemia in pregnancy prevention and management. Data within each category were analysed thematically, being coded inductively, line-by-line in NVivo. These codes were grouped into increasingly higher-level codes to develop sub-themes to reflect key findings.

Quantitative data from the health facility assessments and qualitative data were then triangulated to present an overall picture of the key strengths and barriers to anaemia in pregnancy prevention and management across our study sites in Ghana and Uganda.

Ethical considerations.

We obtained ethics approvals from the Ghana Health Service Ethics Review Committee and, in Uganda, from the Research Ethics Committee of Mbarara University of Science and Technology and the Uganda National Council for Science and Technology. In the UK, research ethics approval was obtained from the Liverpool School of Tropical Medicine.

Permission to carry out the research was obtained from district health offices and the head of each participating health facility. All participants provided written informed consent before proceeding with data collection.

Anaemia is understood according to its symptoms.

In Ghana, some community FGD participants had a good understanding of the symptoms of anaemia. In Uganda, the term anaemia was recognised but was again understood in terms of its symptoms. These included being pale, having low energy, and dizziness. Anaemia was generally understood to mean a “low amount of blood” or “low blood volume” and was typically thought to be caused by not eating and sleeping well. In Uganda, community FGD participants also regularly talked about swelling, and in the FGD with pregnant or recently delivered women, anaemia was also linked to weight loss and loss of appetite.

Respondent 4: I know anaemia is when you don’t have enough blood, or your blood level is low.

Respondent 3: Please I don’t know.

Respondent 2: I haven’t heard of anaemia before.

Respondent 1: I haven’t also heard it before.

Respondent 5: Same.

Respondent 6: Same.

Interviewer: What are some of the symptoms of anaemia?

Respondent 6: Sometimes you feel dizzy and not feel okay.

Respondent 2: Your heart beats faster and you easily get tired.

(FGD, pregnant or recently delivered women, Ghana)

Anaemia was seen as dangerous by participants in all community FGDs, especially to the developing fetus, leading to foetal death or premature childbirth. Some participants also noted it was potentially life-threatening to the mother, though this was given less focus. In one FGD in Uganda, a male partner specifically noted anaemia means “less oxygen and nutrients for the baby.”

Most FGD participants were aware of haemoglobin testing as a way to detect “blood volume”—i.e. anaemia. Male partners in Ghana and community influencers in Uganda also noted that there is a normal haemoglobin range or a “cut-off” and that falling below these indicates anaemia.

Reflecting on what was noted by facility- and district-level participants, some women in Uganda also described how a clinical assessment for anaemia is done during ANC.

The doctors instructs you to look up and looks into your eyes and looks into your hands to see if you have blood. (FGD, pregnant or recently delivered women, Uganda)

There is understanding of local foods associated with “improving blood”.

In Ghana, participants across community FGDs could describe many local foods that were useful in combatting anaemia to “help their blood”, including leafy greens, legumes, plantains, cabbage, milk, dried fish, meat, and eggs. Tomato paste and coke were also mentioned. Community influencers in Ghana also made a link between anaemia and vitamin C deficiency, noting the importance of fruit for preventing anaemia. Pregnant or recently delivered women also mentioned multivitamins and “iron 3 [a hematinic oral solution]—it’s a blood tonic” (FGD, pregnant women, Ghana).

Key informants in Ghana described several successful programmes to educate pregnant women and community members about anaemia in pregnancy. These include pregnancy schools—which are educational sessions run outside of routine antenatal appointments—and dietary education provided through two community-based nutrition programmes.

… we used to have anaemia quite bad in pregnant women, so we put in two major programs. Food demonstration and food bazaars. (KII, district official, Ghana)

In Uganda, local foods widely understood to prevent anaemia or to be consumed by women with anaemia in pregnancy are fruits (mango, banana, watermelon, pawpaw, pineapple, avocado), vegetables (eggplant, amaranth leaves and other greens), maize meal porridge, ground nuts, beans, and meat.

Perceptions towards ANC are very positive.

Across all community FGD participants, ANC was seen as important.

There are many diseases that come up in the pregnancy period…the baby can even die at the point of birth if these things are not taken care of, so it is necessary to go

for ANC. (FGD, male partners, Ghana)

We are firm in knowing we won’t meet any complications because we have been attending ANC. (FGD, male partners, Uganda)

Across community FGDs in Ghana, services referred to as part of ANC included: checking the baby’s growth and position; checking the mother’s weight, blood pressure, and temperature; as well as doing tests for infections and providing treatment where necessary. The provision of counselling, especially around nutrition and hygiene, was also noted frequently. A few participants in Ghana also mentioned that screening for Rhesus factor and sickle cell disease would be done. Male partners in Ghana were divided into those who were unsure about what ANC entails and those who had more knowledge as they had participated in ANC or asked their wives about it and reviewed ANC booklets. Only community influencers suggested that couples receive preferential treatment if a male partner accompanies a woman to ANC. However, mandatory HIV testing during ANC was mentioned as a deterrent to male partner involvement.

I have no idea…no, because I am not allowed inside. (FGD, male partners, Ghana)

My wife said they tested her [haemoglobin] and was told her blood level is low, so I called the nurse and asked her because I didn’t see it in the antenatal book, so the nurse confirmed and said yes, she was told, but they didn’t write it in the book. (FGD, male partners, Uganda)

Community participants in Uganda were quite specific about what ANC covers, including measuring the “quantity of your blood” (to test for anaemia), screening for infections like HIV and syphilis and treating these, providing an expected due date, checking for blood pressure and blood sugar, counselling on nutrition and hygiene, the position and health of the baby, the amount of amniotic fluid, counselling on the place of childbirth, and education on understanding what different pains in pregnancy mean. Some also mentioned “mama kits” that include labour and childbirth supplies as well as newborn necessities that are given depending on the health facility they visited—it was expressed that they would like to see this given at all public facilities.

ANC timing and frequency are not always understood.

Guidelines in both contexts indicate that there should be eight ANC visits throughout pregnancy and that the first visit should be initiated within the first trimester to ensure that interventions are offered at the appropriate time. FGDs with pregnant or recently delivered women in Ghana highlighted variable knowledge about the timing and number of ANC visits, with only one participant noting that there should be “seven or eight” ANC visits throughout pregnancy. Timing of ANC initiation varied from two weeks to six months, though of the 23 pregnant women participating across three FGDs, 16/23 (70%) noted starting ANC in the first trimester. Women who sought care very early (within the first or second month of pregnancy) often commented that they had felt particularly unwell or that they had a history of miscarriage, which prompted them to seek care.

This variability in knowledge about ANC visits was also reflected in male partner FGDs in Ghana, but there seemed to be a stronger sense that going early was not required if there were no issues and that most women went after six months. However, when asked when women should go for ANC, there was consensus that starting at around three months, though a handful of participants noted that ANC attendance should begin as soon as the pregnancy is identified. When asked how often ANC occurs, there was a wide range of responses, from every two or three weeks to monthly, for a total number of visits ranging from 6–18 over the course of pregnancy. There was the sense amongst participants that ANC attendance is increasing, primarily due to policies supporting free care.

In FGDs with community influencers in Ghana, there was broad agreement around the start of ANC being from two-to-three months up until the pregnancy is visible. They expressed that women might delay going so that the pregnancy is more advanced, so they only have to pay for one ultrasound.

Interviewer: How many times [should a pregnant woman attend ANC] …? Number 3.

Respondent 3: 3 times.

Interviewer: 3 times.

Number 4 how many times …?

Respondent 4: It can be 6 times.

Interviewer: 6 times? Number 6 how many times?

Respondent 6: 9 times. (FGD, community influencers, Ghana)

Across all community FGD participants in Ghana, there was the perception that strong women and those with healthy pregnancies do not need ANC, and those who are “weaker” or with obvious health issues are the ones who need ANC earlier.

It depends on the individual. Some people are physically strong when they are pregnant. Others aren’t that strong, so they visit the hospital in their first week. (FGD, community influencers, Ghana)

In Uganda, most participants in the FGD with pregnant or recently delivered women spoke about the timing of the first ANC visit as 1–3 months, and it was clear that “eight ANC visits” was correctly understood as being the desired number of visits. However, in the FGD with male partners and both FGDs with community influencers, this ranged from four-to-eight. There was a shared understanding that most women would be likely to seek out ANC as soon as they have a positive pregnancy test but that there would be women who might go later in the absence of complications.

They go for antenatal on the sixth month of pregnancy…because until then there have not been any complications, so there would not be a need to go for antenatal. (FGD, male partners, Uganda)

However, community influencers suggested that most women start care from three-to-five months’ gestation and associated women coming later with a lack of responsibility.

[Coming at five months for the first ANC] is for a mother that has delayed or…doesn’t care…at three months, a responsible mother should be coming from ANC. (FGD, community influencers, Ghana)

Doctors taught them [women in communities] that when you get pregnant, you start coming for ANC. (FGD, community influencers, Uganda)

Persistent barriers to ANC uptake exist.

All community FGD participants in both countries described recurring barriers to ANC care. These largely centred distance to the health facility, difficulty finding and paying for transportation—especially in remote areas—and other financial barriers, including “hidden costs” for equipment or medication that are expected to be free of charge. These barriers are exacerbated when repeated visits are required.

Some pregnant or recently delivered women in Ghana—especially those who already have children—suggested that expenses are increasing. There is a small charge for each visit (10 cedis (~1.7 USD at the time of data collection)) and two cedis (~0.34 USD) for some medications, even with health insurance. Male partners in Ghana agreed that there are many costs associated with ANC and pregnancy, mainly linked to drugs and transportation, and estimated total expenses throughout pregnancy and childbirth as high as 1500 cedis (~255 USD). Although Ghana’s National Health Insurance should cover costs, community-based participants regularly cited cost-related barriers. For example, if medication is required, as they explained, there are hidden costs that can influence the decision to seek care. In Uganda, it was widely agreed across community participants that ANC in public facilities is free, though payment is expected in private facilities. However, costs may still be incurred, as women might be sent to another facility if certain equipment is unavailable (e.g. an ultrasound machine), or they may be asked to buy medications.

They also like to take money too much, whether you have insurance or not. Medicine that will cost 20 cedis you end up spending 100 cedis. (FGD, community influencers, Ghana)

Respondent 1: It comes down to a money issue and fatigue.

Respondent 3: I want to buttress that point. If you go today to do laboratory work, you would be expected to pay. After three days’ time, when you are scheduled to come back, you would be asked to go and do another. So if that continues, then you find out that your money is getting finished…if you have money, then you go and give birth. They say it’s free, but it’s actually not. (FGD, pregnant or recently delivered women, Ghana).

Many participants in both countries reiterated that the cost of ANC sometimes results in delays in women seeking and accessing anaemia in pregnancy diagnosis and management.

When they go early they will do so many scans, so if you go late in the last trimester that means you will do like just one scan and you leave. So, I think it’s the financial obligation that makes them stay longer at home. (FGD, community influencers, Ghana)

Poor attitudes of ANC staff were extensively highlighted in both countries by pregnant women and male partners, as well as long wait times in Ghana. These prevented women from taking up ANC and also from giving birth in a health facility. Community influencers also noted that healthcare providers do not consistently provide good quality care and can be unhelpful and unkind. However, some women stated they were satisfied with the care provided and the relationship with their midwives. It was understood that this may affect the reputation of the health facility in positive or negative ways as women share information about their experiences.

If I brought my wife here and if she wasn’t treated well—or even if she was treated well—now they are advising [other women]…which will cause a change of heart on the other person. (FGD, male partners, Uganda)

Local perceptions of having safely had children previously without seeking ANC or skilled attendance at birth may reinforce this idea among those women or women close to them.

Some people have given birth safely without antenatal, so they don’t see the need. (FGD, pregnant or recently delivered women, Ghana)

In Uganda, the perception of having lower socioeconomic status was seen as a barrier to some women. Lack of permission or support from husbands or mothers-in-law was also repeatedly cited across community participants.

Lack of proper clothes to wear to come to the hospital…they fear being seen in torn clothes and decided to keep home. (FGD, pregnant or recently delivered women, Uganda)

The use of traditional care was frequently mentioned by community-based participants in Ghana. Traditional care was appreciated as it lacked many of the barriers of formal ANC, such as cost and distance. Traditional medicine was also sometimes viewed as more effective.

Respondent: It is because sometimes traditional/herbal medicine is more potent than orthodox medicine…that is also a contributing reason. (FGD, pregnant or recently delivered women, Ghana)

Facility-based participants and key informants explained that inadequate staffing is sometimes a barrier to adequate ANC and, therefore, to anaemia in pregnancy diagnosis and management, as laboratory personnel were in particularly short supply.

Interviewer: So currently staff is a challenge?

Respondent: It’s a huge problem, it’s a huge problem. (KII, district official, Ghana)

Staff training on anaemia in pregnancy appeared to be infrequent in both countries, though there was training before qualifying and on-the-job. Facility-based participants communicated gaps in training, particularly for training on attitude and conduct.

Apart from training on certain topics, ours have been informal discussions when we go on rounds, we talk a lot and try to educate them on transfusion, that has been my approach, I don’t normally run a lot of formal workshops and those things, we see a patient we just discuss it. (IDI, midwife, Ghana)

Some district-level key informants in Uganda also noted the importance of continuing medical education for anaemia in pregnancy.

Preventing and managing anaemia are well-understood by healthcare providers and less well understood by community members.

In IDIs with maternity staff in both countries, alongside haemoglobin measurement, routine practices during ANC—including preventing and treating malaria, providing deworming medication, giving iron and folate supplements, and providing nutritional counselling—were described as important for preventing and managing anaemia in pregnancy. The use of insecticide-treated bednets was also mentioned by maternity staff and community-based participants. In Ghana, maternity staff reported that women are supervised in taking malaria medication to increase adherence. Women across FGDs in both countries noted that many women did not always like taking antimalarials or iron and folate supplements due to side effects.

Yeah, we are doing it actually in antenatal, we start from antenatal, we first do haemoglobin, then on subsequent visits, we do it, and then that’s why we give them ferrous and folic, whatever we have in place, even the minister of health, that’s what she is advocating for, even the [intermittent preventative treatment for malaria] we are giving, if the woman does not have malaria, if she is dewormed and worms are not sucking blood, if she is given these to supplement her blood, then there is nothing that should stop her from fighting anaemia. Then also advising her on what to eat, nutritional status. (IDI, matron in-charge, Uganda)

I want to find out why they give us the malaria drug. When I went, they ask us to take the medicine right there. But when I came home I had adverse reactions; my mouth area was itching and when I touched it, rashes developed. (FGD, pregnant or recently delivered women, Ghana)

These findings are echoed in facility assessments, where intermittent malaria prophylaxis is routinely given to all women across all participating facilities in both countries and iron and folate supplements are provided to all women during ANC. In Ghana, the provision of anti-helminth medications was offered to all women in ANC in 5/9 (56%) facilities and only to women in higher-risk areas in 4/9 (44%) facilities. In Uganda, these medications were offered to all women attending ANC (Table 3).

Participants in the FGDs with pregnant and recently delivered women in Ghana linked iron and folate supplements to the development of the baby’s blood and bones rather than with their own anaemia. “I think folic acid supports blood cells…about iron? I have no idea” (FGD, male partners, Ghana). In Uganda, pregnant or recently delivered women knew pills could be given for anaemia but did not refer to these by name.

They give us those red tablets to increase our blood levels…and…add on our blood levels and [ensure we] have enough blood to take care of the baby in the womb. (FGD pregnant women, Uganda)

Iron and folate supplements were known by a few male partners and community influencers in Uganda to help women “gain blood”. However, it was mostly understood as being important for health in general and not linked to resolving anaemia.

Although malaria and helminth infections were understood as bad for the baby and mother, they were not associated with anaemia in FGDs with pregnant or recently delivered women in Ghana. Only one pregnant or recently delivered participant in Uganda noted, “deworming prevents [pregnant women] from being anaemic”. However, some male partners and community leaders in Ghana did note the association between malaria and “reducing blood”, and one community influencer in Uganda noted malaria causes “blood loss and complications” and was therefore linked to anaemia. Only one male partner in Uganda noted that worms “take nutrients”, resulting in the mother becoming malnourished and then anaemic (Male partner FGD, Uganda).

Blood transfusion was sparingly mentioned in both countries by community participants as a treatment for severe anaemia.