Examining the impact of WHO’s Focused Antenatal Care policy on early access, underutilisation and quality of antenatal care services in Malawi: a retrospective study

Malawi is classified as a low-income country with a Gross Domestic Product (GDP) per capita as low as US$274 in 2014. This translates into US$0.75 to spend per day for the average individual [11]. Given the low GDP, the government has a limited tax base and faces dramatic trade-offs in its policy decisions while faced with considerable need [13]. These challenges are further exacerbated by a health system with poor infrastructure, lack of equipment and qualified human resources and weak management [12].

The Malawian healthcare delivery system is a three-tier system. The lowest level of care is the primary healthcare or community care system, which consists of community initiatives (village clinics), health posts, dispensaries, maternity units, health centres, community and rural hospitals. District hospitals constitute the secondary level and provide specialised services to patients referred from the primary healthcare level through outpatient and inpatient services and community health services. These services are enhanced by the provision of specialised supportive services, such as laboratory, diagnostic, blood bank, rehabilitation and physiotherapy services. Finally, tertiary healthcare consists of highly specialised services, and is provided by central and other specialist hospitals. These different levels are linked through an elaborate referral system [14].

Only 65% of facilities in Malawi offer antenatal care services [15]. A range of providers offers ANC, including government, non-profit and private providers [15]. In government facilities, the provision of ANC services is integrated with under-five clinics, family planning, postnatal care and other reproductive health services and it is provided free of charge [15]. For-profit and non-profit providers (such as the Christian Health Association of Malawi) require user fees at the point of use. As of 2010, 73% of ANC services were provided at primary health facility level on a daily basis while 27% were provided at secondary and tertiary levels [15]. In Malawi, nurses and midwives are the main providers of antenatal care services (80%), with the rest receiving care from doctors, clinical officers, or maternal-child health aides [15].

According to a Ministry of health report [5], Malawi adopted the FANC policy in 2003. FANC requires health facilities to have adequate infrastructure, clinical skills, essential equipment, drugs and laboratory supplies [9]. However, the government of Malawi did not take necessary measures to invest in the resources required for successful implementation of FANC. In 2010, annual monitoring results on the implementation of FANC showed that only one of the four central hospitals and four of the 24 district hospitals in Malawi met the WHO standards for delivering FANC [16]. Moreover, as of 2014, WHO reported that for every 10,000 people, there were only 0.2 doctors and 3.4 nurses and midwives in Malawi [12].

The study is a retrospective study, which uses three Malawi Demographic and Health Survey (MDHS) datasets conducted in 2000, 2004/2005 and 2010. We chose the 3 years based on comparability. The MDHS provides detailed health information for women of reproductive ages between 15 and 49 and their children. The survey uses a multi-stage cluster sampling design to select households for participation based on the Malawi Population censuses of 1998 and 2008. The first stage comprised a random sampling of the enumeration areas and household listing operation followed by a random sampling of the households in the second stage. This study uses the women data file of the MDHS, which contains data on, among others: demographics, household socio-economic status, maternal health care and ANC utilisation practices. Information on ANC services utilisation and components of care is reported on women of reproductive ages 15–49 who had a live birth during the 5 years before each survey. The response rate for each survey was above 95%.

The main independent variable of interest in this study is the FANC policy dummy, which captures the year when FANC was implemented in Malawi (see Table 1 for details). We created the FANC policy dummy using mother’s year of delivery or the child’s year of birth. All women who gave birth after 2003 were categorised to be in the post-FANC period and those who delivered prior to 2003 are in the pre-FANC period. Our total initial pooled sample was 28,763 women; however, we limited the analysis to women who delivered 3 years before and after FANC adoption, excluding women who gave birth in 2003. This restriction reduces the final sample to 8545 women. Limiting the analysis to the years that were closest to the launch of the policy enabled us to avoid the influence of policies introduced prior to 2000 or after 2006, including the 2007 traditional birth attendant (TBA) ban [5].

Our research considers the impact of FANC on three outcomes: early access, inadequate use/ underutilisation and quality of care.

In this study, we use timing of first ANC visit to measure early access. This variable ranges from 0 to 9 months. A Weibull hazard model (an example of a survival-analysis model) was used to model the gestational age at which the mother enters the ANC system. Weibull models were initially developed to consider the survival of machine components, so that interpretation is sometimes counterintuitive: a higher likelihood of earlier ‘component failure’ here is equivalent to a higher likelihood of an ANC visit at an earlier gestational age. Survival-analysis models are common in the health sciences [17‐19]. We study differences in early access to care before and after FANC was implemented.

FANC recommends a minimum of four ANC visits for women with uncomplicated pregnancies, therefore, in this context, a woman with less than the minimum number of four visits has underutilised the services.

For our model, we sought a more precise definition of underutilisation that does not overlap with that of early access. We therefore limit the sample to women who initiated their first ANC visit by 16 weeks of pregnancy as required by FANC to avoid duplication and overlap with the early access indicator. By limiting the analysis to women who accessed ANC early, we avoid a ‘double count’ problem where underutilisation may merely be another manifestation of late access. The definition also excludes non-users, who are captured in the early access indicator.

Underutilisation is a binary variable defined as 1 if a pregnant woman initiated her first visit in the first trimester of pregnancy but did not make the recommended number of four visits, and 0 otherwise. For the model, we considered cluster averages of underutilisation for the subset of women who initiated care early. Underutilisation is thus the likelihood of women in a specific cluster underutilising ANC service provided that they had a first visit as prescribed by FANC. We analyse this outcome with Ordinary Least Squares (OLS) regressions.

The aim of FANC is not only to achieve a minimum number of four visits, but also compliance with FANC protocols. We therefore also track whether health care workers complied in conducting eight key ANC tests or examinations. These include routinely conducted diagnostics (taking blood and urine samples), physical examination (measuring blood pressure and weight), and other preventive procedures (administration of tetanus toxoid, prophylaxis, iron and folic supplements and establishing complication readiness). These questions were asked in each of the DHSes and are here interpreted as proximate indicators of the quality of ANC services.

As was the case for underutilisation, the analysis sample is also limited to women who initiated ANC early. The first visit recommended by FANC is critical for HIV infected pregnant women, to ensure access to antiretroviral (ARV) prophylaxis. In addition, pregnant women can access early interventions such as syphilis screening and treatment, provision of ferrous iron supplements and malaria prevention and treatment [20]. When a woman accesses her first ANC visit later than 16 weeks of pregnancy, she risks missing some of the early ANC interventions crucial for early detection of pregnancy complications. In such a case, missing some required interventions would therefore not represent non-compliance with FANC protocols, but result from late access.

Using the identified eight key measures, an ANC composite-compliance index was constructed using multiple correspondence analysis (MCA) [21, 22]. Table 2 lists the variables underpinning the composite index, with categories and weights for each variable. The weights are identified from the first dimension of the MCA. This dimension explained about 85% of the total inertia. Positive (negative) weights reflect the higher (lower) quality of care.

The ANC-compliance index was then used as the main outcome variable in our OLS regression to estimate the statistical impact of FANC on quality of ANC services. Again, we aggregate at the community level. This shifts the attention from the demand side (women’s individual experiences) to the supply side (what we can expect a community clinic to provide on aggregate).

We are concerned about distinguishing the effects of FANC from two other shifts that occurred during this period: the expansion of education; the rise in HIV prevalence and launch of the prevention of mother-to-child transmission of HIV (PMTCT).

We used interrupted time series analysis (ITSA) applied to OLS or Weibull models to estimate the changes in the outcome variables of interest between the two periods. ITSA allows us to capture both changes in levels and changes in trends. The design is particularly suited to interventions introduced at a population level over a clearly defined time period that target population-level health outcomes [32‐34]. In standard ITSA, the following segmented-regression model is used:

Where Y_it is one of the outcomes of interest (in our case, quality of ANC services index; timing of first ANC visit; proportion of women with inadequate visits/underutilisation). i indexes either the individual or geographic cluster. t indexes time or year of delivery. β₀ represents the baseline level of the outcome at t = 0, β₁ is interpreted as the change in outcome associated with a time-unit increase (representing the underlying pre-intervention trend), β₂ captures any discrete changes following the intervention,  X is the FANC dummy variable indicating the pre-intervention period (coded 0) or the post-intervention period (coded 1). T measures the time before (negative) or after (positive) the intervention, and is included as a linear trend. β₃ indicates the slope change following the intervention (using the interaction between time and intervention: TX). Z is a vector of controls for confounding factors and μ is the error term. Our interest lies in assessing whether the time progression has been altered by FANC, so that is our coefficient of interest. All statistical analyses were conducted in Stata 14.

Descriptive statistics in Table 3 illustrate changes between the pre- and post-FANC periods in some of the variables used in our analysis. The results show significant increases in early access and quality of ANC services in the post-FANC period. However, the average number of ANC visits declined to just below the minimum number of four visits required by FANC in the post-FANC period.

The proportion of women with no education is lower in the post FANC period at 21% compared to the pre-FANC period at 27%. The proportion of women who reported having been tested for HIV between the two periods jumped from 15% before FANC to 41% after FANC. We see a statistically significant increase in two of the pregnancy risk factors: the proportion of women with a history of caesarean delivery increased from 3 to 5% after FANC and miscarriage increased from 11 to 13% after FANC. Lastly, there is also a significant increase in media access through radio and television in the years following FANC implementation.

Figure 1, shows changes in the timing of the first ANC visit between pre- and post-FANC periods. The y-axis represents the month after conception during which a woman initiated her first ANC visit and the x-axis represents the number of years from the implementation of FANC. Negative (positive) values are the number of years before (after) FANC implementation, while zero indicates the time of implementation in 2003. Similar to the descriptive statistics, we notice a significant reduction in late access following FANC implementation, over and above the existing downward trend that emerged before FANC.

Table 4 presents the Weibull regression results for the relationship between FANC adoption and early access to care. We find that FANC had decreased the likelihood of early access when we look merely at the levels, but this effect is overwhelmed by the positive trend effect. It is encouraging to see that there is a positive coefficient on the post-FANC trend and that the implementation of this policy has increased early access.

Furthermore, having at least primary-level education, access to radio, having been tested for HIV, a history of miscarriage or caesarean delivery, and urban residence are all associated with a higher likelihood of early access to care.

Figure 2, shows an increase in the community-level underutilisation of ANC services after FANC adoption. Table 5 confirms that FANC has had the unintentional consequence of increasing underutilisation. We see that before FANC, women were progressively less likely to underutilise ANC services over time. After FANC was adopted underutilisation increased by eight percentage points per year (after accounting for the confounding variables), relative to the initial downward trend.

Very few of the control variables and the proxies for confounding factors were statistically significant in these regressions. Those respondents who had access to the radio at least once a week are more likely to underutilise care.

Table 6 reports descriptive results on individual level quality of care. The findings show an overall improvement in compliance of individual ANC contents after FANC. Women who reported to have received all eight components increased from 9 to 11% in the post FANC period. There are significant increases in taking of blood and samples, communication about complications, dispensing of iron tables and dispensing of malaria prophylaxis.

In Fig. 3, we show the change in overall quality of ANC services at the community level, as measured by the MCA index. In the years immediately after FANC implementation, there was a decline in quality, which was then followed by a period of stagnation before a substantial improvement. This could represent a one-year lag or initial teething problems with the implementation of this policy.

The coefficients reported in Table 7 show no evidence of FANC improving the quality of care. Based on the descriptive results we consider the hypothesis that there may have been teething problems with implementation or a lag in implementation, changing the policy cut-off point to 2004, but this specification also does not provide any evidence of an improvement in the quality of care.

Urban location, education and HIV-testing prevalence rates are all positively associated with quality of care. In addition, skilled health providers (doctors and nurses) are more likely than unskilled providers (TBAs) to comply with the set of protocol measures in our quality index.

Our results show a positive association between FANC and early access. The findings demonstrate a significant increase of about 21 percentage points per year (relative to baseline) in early access after FANC adoption. FANC has helped to promote early access in Malawi. We find the results plausible because FANC required the first contact to occur by 16 weeks of pregnancy [2]. However, the descriptive results show that 3 years after FANC implementation only 29% of women in our sample initiated ANC by 16 weeks of pregnancy. At a national level, the MDHS reports that only 24% of women initiated ANC early in 2015 [35]. Many women delay access to care beyond the first trimester and this is worrying given that HIV transmission remains a concern.

Furthermore, our results show that women who have at least primary education are more likely to initiate their first ANC visit in the first trimester. This finding is similar to that of Yaya et al. [36] who argued that educated women are more knowledgeable about the importance of ANC visits. Similarly, earlier studies conducted by Furuta and Salway [37] in Nepal suggest that education promotes new values and attitudes that are favourable to the use of modern health care, thereby increasing the likelihoods of women accessing skilled care. Consequently, policies aimed at improving female education are likely to also increase ANC usage [28]. This is particularly important in Malawi, where only 1 in 4 women have at least some secondary education and only 3% of women have tertiary education [35].

Women with a history of miscarriage and/or caesarean deliveries are also more likely to initiate ANC early, confirming the results found by Pell et al. [38]. This finding is of particular importance because it shows that women are responding to FANC’s emphasis on complications and danger signs to watch for in pregnancy. Our study also finds that women who have access to a radio at least once a week are more likely to initiate their first ANC visit early. This finding is consistent with findings by Bbaale [39]. They concluded that it was necessary to focus on educating girls beyond the secondary level, but also to increase media penetration for the purpose of spreading information about the importance of ANC visits.

The type of residence a woman reports to come from has a significant impact on early access in our study. Women who reside in urban communities were more likely to initiate their first ANC visit early compared to women from rural communities; this finding is consistent with previous studies [40‐42]. Supply factors such as the quality of ANC services provided and the distance of the healthcare facility have been reported to play a significant role [43]. Malawi has promoted FANC via free public provision, however, the health facility distribution favours urban areas [5].

Other factors influencing the timing of first ANC visits, as reported by qualitative studies in Malawi, include women’s beliefs [43] and user fees in CHAM (Christian association of Malawi) hospitals [3]. Evidence shows that mothers often seek care or pregnancy lessons from traditional health workers and elderly women before seeing medical personnel at a facility in Mangochi, Malawi [43]. This practice stems in part from the belief that it is not necessary to attend clinics or hospitals unless one is physically ill. Furthermore, while FANC health services are free at the point of use in public facilities, in communities where the majority of facilities are managed by CHAM, they pay user fees at the point of use of 2000 Kwacha [3 US$] on average to access FANC services [3]. Many rural and poor women in Malawi cannot afford the cost to access private FANC services.

The results show that FANC has been associated with unintended increases in underutilisation of ANC services. These results are consistent with one previous Zambian study where the researcher reported a significant decline in the number of visits to an average of three following FANC adoption [44]. A qualitative study in Malawi reported that some FANC requirements served as barriers to accessing and adequately utilising FANC services in Mangochi [3]. In particular, FANC encourages men to engage in reproductive health and preparation for birth. The study reports that most of the health facilities in the district have by-laws formulated by health officials and community leaders stating that a pregnant woman can only be attended to if her spouse accompanies her. This led to many women not utilising FANC services. Furthermore, due to the requirement for birth preparedness as suggested by FANC guidelines, health workers demand that pregnant women bring with them cloth wraps for the baby (traditional “chitenje”) at their initial FANC visit [3]. All these requirements reportedly make FANC services expensive and therefore less accessible.

Another interesting result is the negative significant coefficient between underutilisation and access to radio. These results are consistent with those found by Bbaale [39]. Media access, especially access to radio, has been shown to be instrumental in changing behaviour regarding maternal health services utilisation [36, 42].

There was an increase in blood testing, but 3 years after the implementation of FANC only 47% of women received blood tests. This level of access is disappointingly low given that blood sample testing is one of the essential factors in reducing HIV transmission from the mother to the unborn child. Under reporting is a distinct possibility due to potential recall bias or the sensitive nature of this question.

Only 22% of women, who gave birth after FANC was implemented, reported having been given a urine test during the ANC visit. Lungu et al. [9] found that there were shortages of essential laboratory supplies at one urban clinic in Malawi, which made it difficult to conduct the essential laboratory tests. Nyarko et al. [7] reported similar findings in Ghana. Our results also show that only 11% of users in our sample received all ANC components after FANC implementation. In Nigeria, only 5% of users received the optimal level of care (all 10 components) after FANC adoption [27].

Our regression results show no significant improvement in the quality of ANC services even after accounting for a one-year lag in adequate implementation of the policy. This is similar to the experience in South Africa, where FANC did not improve performance in maternal health delivery [8]. A multi-country level study conducted in selected low and middle-income countries found very low compliance rates to FANC standards among women who went for their required four ANC visits [45]. The results of these studies show that low and middle-income countries struggled to implement FANC and as such, the policy did not translate into improved quality in ANC services.

Our results did however show a significant positive association between HIV-testing share and quality index, suggesting that PMTCT may have had an impact on quality. We also find that the type of health worker providing ANC plays a vital role in determining the quality of ANC received by women. Women were more likely to receive good quality ANC care if skilled providers (doctors or nurses) attended to them compared with those who were relatively less skilled (patient attendant/ health surveillance assistants and traditional birth attendants). Similar results were also reported in Nigeria [27] and Nepal [28]. However, a qualitative study analysing women’s perspectives on quality of care in Mzimba, Malawi reported that health worker attitudes towards clients had negative impacts on the clients’ perceptions of quality received [46]. As such, women preferred traditional birth attendants over skilled personnel.

The average education level of mothers is also positively associated with higher-quality ANC services. As previously explained, educated mothers are more aware of their rights and more knowledgeable about health education and what is expected at the ANC clinics. Similar findings were reported in Nigeria [27], Nepal [28] and Uganda [39]. In line with previous studies, we find urban communities have better quality ANC services in comparison to rural-based clinics. A study in Nigeria, also reports similar results [27], where women in urban areas had higher odds of receiving good quality ANC compared to those in rural areas. This may be because in most cases, urban clinics are better endowed in terms of resources and budget allocations [5]. There is therefore a need to improve the quality of ANC services delivery in rural areas.