Detecting and managing hypertensive disorders in pregnancy: a cross-sectional analysis of the quality of antenatal care in Nigeria

Following discussion with the Nigerian Federal Ministry of Health and the donor, seven states covering the country’s six geo-political zones were chosen as study sites – Sokoto, Bauchi, Katsina, Kogi, Cross River, Ebonyi, and Ondo (Fig. 1). In consultation with the states’ Ministries of Health, ninety-five public facilities with ANC units were purposively selected from within these states to represent diversity in facility type (PHCs = 56, secondary/tertiary hospitals = 39). Following an inventory assessment at 95 selected facilities, researchers conducted service provider interviews (n = 200) from a subset of 93 facilities; 2 facilities were excluded as no willing providers were present on the day of the inventory assessment. Researchers also observed 135 ANC consultations and conducted 135 client exit interviews from a subset of 26 facilities. The subset of facilities used for consultation observations and client interviews was chosen using convenience sampling, based on the location of the facility (considering difficult terrains), as well as the functionality of the facility (facility operating hours), while ensuring a representative sample across facility level and region.

The framework used for measuring quality of ANC was adapted from Warren et al. [17] and included many of the essential elements of antenatal care as informed by WHO guidelines [1]. It was further tailored to include indicators specifically pertaining to the capacity of the ANC to detect and manage HDPs.

The framework incorporated three elements of QOC as described by Donabedian [9], Bruce [18], and Hulton [19]: structure, process, and outcome. Structure encompassed factors related to inventory and infrastructure (facility guidelines, equipment, supplies, drugs, referral mechanisms, etc.), as well as human resources (provider knowledge and training). Process encompassed the range of services provided, information and documentation shared with the client, the practitioner’s competency in history taking and physical examinations, and their interpersonal skills. Outcome encompassed client experiences, satisfaction, and health comprehension. Table 1 shows the indicators used to create structure, process, and outcome scores. Scores were calculated as the average cumulative number of points received by each unit of observation (facility, provider, ANC consultation, or client interview). All scores are reported as a percent of the total possible score.

This paper uses cross sectional data collected during a landscape analysis that assessed the health facility and provider capacity to prevent, detect, and manage preeclampsia and eclampsia in seven states in Nigeria [13]. Data collection occurred over a three-month period, from June–August 2015.

We compared mean scores obtained from PHCs and hospitals in each QOC category using one-tailed independent sample t-tests (H₀: secondary facility = PHCs, H_A: secondary facility > PHC); variances were assumed unequal if Levene’s Test for Equality was significant. For all categorical variables, significant differences between facility type were determined using the Pearson chi-square test; Fisher’s Exact test was used in cases where individual cell counts were below 5.

We then conducted two multivariate linear regression models to identify independent variables associated with provider knowledge and outcome scores. Variables in model 1 included facility level, provider age, gender, and designation; variables in model 2 included facility level, client age, education, and socio-economic status (SES). Clustering within facility was adjusted for using a robust estimate of variance (sandwich estimator).

We developed a single variable for SES through principal component analysis of indicators pertaining to wealth, household facilities, and assets (including vehicles, land, and livestock owned). We classified SES as either low, medium or high based on whether the individual fell into the bottom, middle, or top tertile [20].

A significance threshold of p = 0.05 was used for all analyses. Individual observations with > 50% of data missing were excluded from analysis. Where appropriate, means (with standard deviations), coefficients (with 95% confidence intervals), and p-values are reported. Data were stored, cleaned, and analyzed using STATA v 15.1 [21]

Client and provider characteristics are described in Table 2. The average client was 27.78 years of age, with those visiting PHCs being significantly younger than those visiting hospitals (24.9 vs. 28.3 years; p = 0.001). Most women had completed at least secondary level of schooling (30.1%), and the level of education attained was significantly associated with facility type, with a much higher percentage of women at hospitals having attended postsecondary school (p < 0.001). Those identifying as Muslim were approximately three times more likely to attend a PHC than a hospital facility (p = 0.001), likely because the majority of PHCs sampled were in Northern states, which are predominantly Muslim. There were no significant differences between PHCs and hospitals in the client’s SES, marital status, or spending decision habits.

The mean age of providers (39.1 years) was similar at both PHCs and hospitals, and the majority were female (81.6%). In PHCs, most providers were Community Health Extension Workers (CHEWs) (64.8%), followed by nurse/midwives (18.1%); this was significantly different from hospitals, where the majority of provides were nurses/midwives (66.7%), followed by general practitioners (15.6%) (p < 0.001).

Table 3 shows the structural attributes related to facility inventory and infrastructure. Out of a total of 38 points, PHCs scored significantly lower than hospitals (47.5% vs. 60.0%, p < 0.001). PHCs also scored significantly lower in domains of general infrastructure (40.8% vs. 54.3%; p = 0.0116), general equipment (54.2% vs. 64.6%, p = 0.0304), referral mechanisms (54.9% vs. 67.3%; p = 0.0161) and capacity for HDP specific services (38.8% vs. 68.6%; p < 0.0001). There were no significant differences between facility type in scores for data collection tools (90.5%), HDP guidelines (33.9%), or availability of drugs associated with HDPs (38.5%).

Table 3 shows the structural attributes related to provider knowledge and training. Out of a total of 66 points, providers in PHCs scored significantly lower than those in hospitals (46.0% vs. 54.1%, p = 0.002). PHCs scored lower than hospitals in additional facets of provider knowledge including signs and symptoms of HDPs (60.1% vs. 67.8%; p = 0.019), how to manage severe pre-eclampsia (49.2% vs. 59.0%; p = 0.032) and eclampsia (49.7% vs. 65.2%), correctly diagnosing HDPs (58.0% vs. 73.6%; p = 0.000), correct use of magnesium sulfate (15.9% vs. 31.1%, p = 0.000), and use of antihypertensive drugs to manage mild HDP (13.6% vs. 23.5%, p = 0.001).

The difference observed in overall provider knowledge between PHCs and hospital facilities was no longer significant after regression adjustment for provider demographic variables including age, gender, type of provider, and length of time working at facility (Fig. 3). However, overall provider knowledge was significantly associated with provider type; using maternal health specialists as the reference (obstetricians/gynecologists), it was found that general practitioners, nurses/midwives, CHEWs, and CHOs scored lower by 12.0 points (p = 0.001), 15.4 points (p = 0.000), 23.8 points (p = 0.000) and 18.5 points (p = 0.000), respectively. Further adjustment for clustering by specific facility did not quantitatively or qualitatively change the results of the regression analysis.

Table 4 shows the process attributes related to provider technical skills. Out of a total of 37 points, PHCs scored significantly lower than hospitals (42.2% vs. 49.8%, p = 0.0017). There were also significant differences between facility type for skills in general history taking (42.2% vs. 61.0%; p < 0.001) and history taking for women with a previous pregnancy (39.6% vs. 58.8%; p = 0.0073). There were no significant differences in scores for physical examinations, lab tests, documentation, health promotion/prevention of disease, or encouragement of follow-up, but all categories except for documentation and encouragement of follow-up scored below 70% of potential points.

Table 4 also shows the process attributes related to provider interpersonal skills. Out of a total score of 11, PHCs and hospitals scored highly – 85.3 and 84.9%, respectively. There were no significant differences observed in the scores of interpersonal sub-categories, or in the cumulative interpersonal score between PHC and hospital care providers. However, individual indicators within the rapport-building and communications sub-categories were found to significantly differ (see Additional file 1). For example, PHCs more often used the client’s name (97.5% vs. 89.5%; p = 0.016) and greeted in a friendly manner (100% vs. 85.4%; p = 0.011), but were less likely to maintain audio privacy (67.5% vs. 92.6%; p < 0.001) and visual privacy (65.0% vs. 90.5%; p < 0.001), or inquire for need of other services (35.9% vs. 83.3%; p = 0.022).

Table 5 shows the outcome attributes related to client experiences. Out of a total of 16, PHCs scored significantly lower than hospitals (45.4% vs. 55.2%; p = 0.004). While perceptions of wait times, costs, and satisfaction were not significantly different between facility type, clients at PHCs had significantly lower health literacy than those at hospital health facilities (39.5% vs. 50.6%, p = 0.020). Wait-time and time spent with providers were comparable between primary and secondary facilities, but women attending ANC at hospitals reported willingness to pay a higher amount for services (1.28 USD vs. 0.43 USD; p = 0.044) (Table 6). More than half of all women (51.1%) could not recall any danger signs in pregnancy that would compel them to return to a health provider. Only 5.2% of clients could name at least 3 danger signs – most commonly severe headache, bleeding or fluids from vagina, and swelling of legs and feet (Table 6). Notably, satisfaction was very high for all facilities; 100% of respondents in PHCs and 89.4% of respondents in hospitals reported being satisfied with services received.

The difference observed in overall outcome score between PHCs and hospital facilities was no longer significant after regression adjustment for client age, level of education, and socioeconomic status (p = 0.189, [Additional file 2]). Outcome scores for women with a “high” SES were 1.42 (95% CI = 0.03, 2.80) points lower than scores for women with a “low” SES (p = 0.045). Further adjustment for clustering by specific facility did not quantitatively or qualitatively change the results of the regression analysis.

Facilities captured in this study obtained very low scores for inventory and infrastructure - below half of the tools, recent guidelines, equipment, drugs, and mechanisms deemed necessary to perform antenatal services. Within general equipment, only three quarters of facilities had operational sphygmomanometers for measuring blood pressure, and less than half had dipsticks for proteinuria testing. These numbers are concerning, especially regarding the important role they play in diagnosing pre-eclampsia and eclampsia [22].

PHCs scored significantly lower than hospitals in almost all aspects of inventory and infrastructure. Ademiluyi and Aluko-Arowolo [23] have written on the unique infrastructural challenges faced by PHCs, citing Nigeria’s colonial history and its implications on the evolution and distribution of healthcare resources as the root cause. To this day, the bulk of public healthcare dollars are spent in secondary facilities, as is government spending on general infrastructure [24]. While all levels of facilities must meet standards for accreditation, secondary level facilities have much stricter infrastructure requirements, including those for operating theatre, pharmacy, laboratory, and personnel capacities [25]. Despite their limited capacity to treat pregnant women, PHCs play a vital role in advising women on danger signs of obstetric complications, detecting these complications (including pre-eclampsia and eclampsia) early and referring women to general or specialist hospitals in a timely manner (often following initial stabilizing care, such as administration of loading dose of MgSO₄) [26]. However, PHCs likely face challenges in efficiently getting patients to hospitals in the case of emergencies, as this study reveals that only a quarter have any ambulatory transport systems.

Other research has found a similar lack of infrastructure in both primary and secondary facilities in Nigeria. Kress et al. [12] reported on a “general shortage of drugs and supplies” that likely are the result of segmented supply chains and financial constraints placed on Local Government Authorities (LGAs) who are responsible for allocating resources to PHCs. They report that as much as 95% of funding must be allocated to staff salaries at the expense of adequate drugs, supplies, and maintenance. Oyekale [11] similarly found that tools and drugs essential for ANC including sphygmomanometers, thermometers, stethoscopes, magnesium sulfate, folic acid, and calcium gluconate were unavailable or dysfunctional in as many as 87% of facilities studied.

Provider knowledge and training received the lowest overall scores, with PHCs scoring significantly lower than hospitals. This finding is not entirely unexpected given the presence of lower level staff at PHCs compared to secondary facilities. Indeed, regression analysis shows that provider knowledge scores are dependent on the type of provider, (with obstetricians/gynecologists scoring at least 12 points higher than the next best cadre – general practitioners). This is not to say that lower cadre health workers cannot be adequately trained on critical components of ANC, as has been demonstrated in previous research [27‐30]. Despite existing evidence that shows task-shifting and training of lower-cadre healthcare workers is an important strategy for low-income countries, only 12.2% of providers in this study reported having received any training in the past 3 years.

For both PHCs and hospitals, scores for providers’ skills were moderate to low. Providers were relatively successful in the realm of ‘follow-up/ continuation of care’ (most clients were observed being instructed to come for a follow-up appointment and told where to go for follow-up). However, a smaller percentage were given a written reminder of when to come or encouraged to return in the case of emergency. It is well-documented that Nigerian women attend fewer ANC consultations than is recommended for healthy pregnancy and childbirth [8]. Providing written reminders may be a simple strategy to improve rates of ANC attendance. Previous research in Zanzibar has shown the effects of follow-up reminders (via SMS) to be beneficial in not only the number of ANC visits attended, but also in the percentage of women with antenatal complications identified and referred [31].

In 2015, Fagbamigbe and Idemudia [16] studied ANC quality in Nigeria and reported on the eight nationally recommended ‘critical components’ of ANC (blood pressure, iron supplements, blood sample, urine sample, tetanus injections, danger signs, HIV tests, malaria prophylaxis). As in our study, the most common of these eight components practiced were blood pressure measurement and provision of iron/folic acid supplements. Compared with those in Fagbamigbe and Idemudia, the participants in our study reported higher frequency of HIV testing (78.5% vs. 41.7%) and provision of malaria prophylaxis (77.8% vs. 40.1%), and lower frequency of urine testing (28.1% vs. 81.9%). Furthermore, process scores may be an optimistic interpretation of the provider’s skill in these areas; although the data tells us that the actions were taken, it does not specify the degree to which they were correctly completed. For example, there are many sources of inaccuracy for blood pressure measurement including patient, device, procedure, and observer-related sources [32]. Previous research has identified blood pressure inaccuracy as a challenge in hospitals in developing nations, and especially among providers with lower qualifications such as nurses and CHOs [33, 34].

Provider interpersonal skills, as assessed by direct observation of their interaction with clients, scored highly. Audio and visual privacy was the only interpersonal indicator in which PHCs performed worse than secondary facilities; this is understandable given the relatively little space and busy atmosphere of PHCs. Babatunde et al. [35] found similarly positive results among PHC users, with the ideal interpersonal skills being displayed in 71–91% of healthcare interactions.

Low scores in health literacy (dissemination, communication, and explanation of consultation results) were reflected in clients’ poor recall of danger signs of pregnancy. Hospitals did significantly better in this regard than PHC, possibly because of better-trained staff, or more educated clients. Gaining sufficient and adequate information about one’s health and potential complications from a care provider is important during the antenatal period and is associated with improved ANC attendance and pregnancy outcomes [36‐38]. Health literacy is critical to patient empowerment, and accurate information from an informed provider is especially valuable when the alternative might be informal education, misrepresentation, or mythology.

Despite the low scores received for structural and process attributes, as well as moderate perceptions of wait time and cost, all facilities scored extremely well in terms of client satisfaction. This result is encouraging, as each new interaction with the health system presents an additional opportunity for HDPs to be detected and managed. However, perception of care must be interpreted cautiously, as low-income and low-education women can be uncritical of the healthcare they receive or base it on the provider’s interpersonal skills rather than their competence [17, 39, 40]. Only a quarter of clients attended an ANC visit before 15 weeks gestational age, despite the WHO’s recommendations for an initial contact within 12 weeks of pregnancy. This is especially important for first time mothers, and to ensure early detection of HDPs, HIV, and anemia [1].

In 2016, the WHO released “Standards for Improving Quality of Maternal and Newborn Care in Health Facilities” a document which outlines 8 domains of quality of care that should be assessed, improved, and monitored within the health system [41]. These range from the appropriate use of data, to effective and compassionate communication with women and their families, to appropriate physical environment and competent, motivated staff. Furthermore, quality antenatal care should result in a “positive pregnancy experience”, defined as not only maintaining a healthy pregnancy for the mother and baby, but “having an effective transition to positive labor and birth”, and “achieving positive motherhood (including maternal self-esteem, competence and autonomy)” [1]. In accordance with these standards, our study shows that improvements to ANC must be holistic and encompass not just the facility, but also provider knowledge and practices. It must also involve community mobilization to educate on the importance of ANC and improve maternal health literacy. In an environment where resources are limited and there are competing interests in terms of health priorities, solutions must be innovative and easy to adopt, especially for PHCs. This may include mobile training and decision-making tools, such as the m4Change application created as part of the Nigerian government’s Saving One Million Lives Initiative [29]. Policy that mandates regular training refreshers around ANC skills could help to close knowledge gaps. Improvements should be evidence-based and outcomes-oriented, meaning that they are measured using outcome indicators such as the number of women who have at least four ANC visits, tetanus protection at birth, or the proportion of women with a written birth and emergency plan at 37 weeks of pregnancy [42]. This research comes during the development of a new National Health Act, in which the Federal Government of Nigeria has committed to “reactivate” and “revitalize” the country’s PHCs. This has begun with the inauguration of an inter-agency Supply Chain Committee, which will act as the implementation partner for the PHC revitalization agenda [43]. This research may therefore serve as a framework for which major gaps in PHC quality, specifically as it pertains to ANC and the ability to manage HDPs, require immediate attention.