Compared to the isolated maternal and pregnancy hospitalization outcomes, there was a high proportion of unfavorable hospitalization outcomes for the mother-pregnancy entity in the two hospitals of the “Cascades” region. This comparison was possible thanks to the definition of the aggregate indicator "unfavorable outcome of hospitalization for the mother-pregnancy entity”. The factors associated with these unfavorable outcomes were the quality of anemia prevention, the stage of anemia, gestational age, mode of admission, and residence.
The limitations of such a study were inherent in the retrospective nature of data collection from hospital records, which were summarily filled out and sometimes poorly archived. However, the extension of the period covered by the data collection made it possible to have a representative number of subjects included in the analyses despite the extent of the missing data.
A selection bias may have been introduced by including only records with a positive diagnosis of anemia. This shortcoming was mitigated by checking the similarity of the results with the national references and those in the literature.
The aggregation of outcomes had the particularity to include referrals/evacuations and discharges without medical advice. This allowed for a more comprehensive assessment of hospital performance in providing care for anemia in pregnant women.
Thus, the referrals and evacuations accounted for the largest part of unfavorable outcomes. More than one in five patients were referred or evacuated to a higher level of the care system. One-third of the cases received from referral/evacuation were again evacuated to a higher level, even though regional hospitals were supposed to take care of the majority of cases received, and refer only a small proportion (Ministry of Health. National Guide to Referral and Counter-referral. 2005. Unpublished document). The proportion of discharges against or without medical advice was high, but the study design did not allow to explore their motivations. In 2012, 5.12% of patients were discharged without or against medical advice in Burkina Faso hospitals (13). Exploring these motivations could help reduce the frequency of these unplanned hospital discharges. In studies conducted in African hospitals, most authors have found high proportions of discharges without or against medical advice. The main reasons for these discharges were the lack of improvement in health, financial difficulties, poor hospital conditions and disagreements with caregivers (14).
The maternal mortality ratio in our sample was very high (3,732/100,000 live births). It thus reached 26 times the in-hospital maternal mortality of 144 per 100,000 live births reported in Burkina Faso in 2012 (13). The maternal mortality in our study was rather close to the 2,800/100,000 live births reported among cases of anemia in pregnancy in the same regional hospital in 2014 by Savadogo et al (7). It was higher than that of the study by Ouédraogo et al done at the hospital of the “Nord” region of Burkina Faso in 2019; they reported 1,581 maternal deaths per 100,000 live births among cases of anemia in pregnancy (6). According to Bailey et al, the case fatality of anemia in pregnancy reaches 2.3% in low-income countries (15). In our study, the proportion of maternal deaths showed a downward trend over the triennium: 3.5% in 2009, then 0.9% in 2010 and 1.8% in 2011.
The unfavorable hospital discharges for pregnancy was marked by the high proportions of abortions and stillbirths. The proportion of abortions of 127‰ in our study was higher than that reported by overall health facilities of Burkina Faso in 2012 (13). According to a 2011 estimate by Sedgh et al in Burkina Faso, 25‰ pregnancies ended in abortion (16). This rate reported by the same author was 31‰ for all West African countries (17). But, these estimates did not target only cases of anemia and took into account non-hospital data. Nonetheless, abortion is identified in the literature as a cause and not a consequence of anemia in pregnancy, particularly through its complications (hemorrhage, sepsis) (18).
The proportion of stillbirths reflected that observed in whole hospitals (6.5%) in Burkina Faso in 2012 (13). It was close to that of Savadogo et al, who reported a stillbirth rate of 9.4% in 2014 among cases of severe anemia in pregnancy at the regional hospital of “Cascades” (7). In contrast, Ouédraogo et al found a higher stillbirth rate (13.76%) at the regional hospital of the “Nord” region of Burkina Faso in 2019 (6). In a systematic review published in 2017 by Bailey et al, the in-hospital stillbirths ranged from 0.58–11.65% in low-income countries (15). We did not find an association between stillbirth and stage of anemia, although the association between maternal anemia and stillbirth is well affirmed in the literature. On this subject, Rahman et al in a meta-analysis published in 2016, attributed 18% of stillbirth (RR = 1.51; CI95%: 1.30–1.76) to maternal anemia.
For ectopic pregnancies, the frequency in our study was lower than that of the whole country (1.38% of pregnant women hospitalized), but close to that of the “Cascades” region (1.04%) in 2018 (13). In studies conducted in West Africa, this frequency was between 1 and 2% (19–21). Of course, ectopic pregnancy is not a consequence of anemia, but it could be the etiology, especially in cases of rupture (22).
Within the limits of our knowledge, the published literature had not yet documented the aggregate indicator of hospitalization outcome for the mother-pregnancy entity. This lack of information limited the discussion of our result without altering its relevance. The high proportion of unfavorable hospital discharge for anemia cases indicated the extent of unmet need for anemia care in hospitals in Burkina Faso. Several factors explained these unfavorable outcomes. The most important were severe anemia, admission from evacuation, poor quality of anemia prevention, and 1st or 2nd trimester of pregnancy.
Severe anemia was the most frequent reason for evacuation to the next level of care system, and was associated with maternal death as documented by several authors (2, 3, 7). The hospitalization for severe anemia in the 2nd trimester of pregnancy ended in referral/evacuation in nearly two-thirds of cases, with anemia as the reason in more than half of the cases. The care system was therefore sufficiently alert to detect but not to manage the severe anemia, particularly in the 2nd trimester of pregnancy. This excess of "repetitive" referral/evacuation of cases of severe anemia was explained by the low availability of blood for transfusion and the non-use of injectable iron in Burkina Faso hospitals. It is known that severe cases of anemia near the end of pregnancy respond poorly to oral iron treatment (5). This issue also explained the high case fatality of severe anemia in pregnancy.
The unfavorable hospital discharges were also associated with marital status and the determining parameter was the high frequency of abortion (91.4%) among women not in-couple. In a multi-site study in Ghana and Mozambique published in 2018, Dickson et al identified an association between non-union status of women and high abortion frequency (23). Further, Klutsey and Ankomah showed that women who were not in union were at greatest risk of induced abortion (24).
Also, multiparity was associated with a higher frequency of unfavorable hospital discharge, with a higher proportion of stillbirths (64.5%) and maternal deaths (64.3%) in multiparous women. This observation has also been made by several authors, especially in West Africa and South Asia, but also elsewhere (25–27).
Beside these factors, rural residence was associated with unfavorable hospital discharge of cases of anemia in pregnancy. This association was consistent across all logistic models. The majority of referrals/evacuations, severe anemia cases and maternal deaths were patients residing in rural areas. In an historical cohort (2008–2010) conducted in Niger in 2017, the authors made the same conclusion that the majority (74.9%) of maternal deaths occurred in patients from rural areas (28). In a "three delays model" analysis of maternal death, Sombié et al summarized the main factors contributing to maternal mortality, including poor access to emergency obstetric care, especially in rural areas (29, 30).
The low quality of anemia prevention was due to inadequate iron/folate supplementation, malaria preventive treatment and obstetric coverage. In a previous study conducted in the “Cascades” region in 2012, we observed in addition, insufficient knowledge of measures to prevent anemia in pregnancy. This insufficiency concerned 75% of health care providers and 66% of community health workers of the sample. (31). Anemia was still the second most common reason for hospitalization of pregnant women in Burkina Faso in the health statistics yearbook published in 2018 (32).
In summary, the residence in a rural area, the status of woman not in union, admissions from medical evacuation, the first two trimesters of pregnancy, multiparity, moderate or severe stage of anemia and the low quality of anemia prevention were the determining factors of the unfavorable hospital discharge of pregnant women with anemia in the “Cascades” region between 2009 and 2011. Thus, surprisingly, the technical platform of the hospital and the income level of the patients were not decisive in the results of hospitalization of cases of anemia in pregnancy.