Use of family planning and child health services in the private sector: an equity analysis of 12 DHS surveys

Disparities in access to quality healthcare, as well as achievement of health outcomes by wealth status and area of residence has been well documented [1–3]. Poorer individuals are more likely to go without care than wealthier individuals, and to spend proportionately more of their incomes on healthcare [4, 5]. To reduce inequities in health, many public and private health programs in developing countries explicitly aim to serve the poor, often defined as the poorest 40% of the total population or as those living below a given global poverty line. Equity in the utilization of healthcare services, as well as the differences in health outcomes based on socioeconomic status, is of interest to donor organizations, program leaders, researchers, policymakers and governments. Specifically, to ensure progress towards targets 1, 2, 7 and 8 of Sustainable Development Goal 3 of healthy lives and well-being for all, attention must be paid to reproductive, maternal and child health services [6]. Furthermore, the ability to access quality healthcare without financial hardship is a key element of universal health coverage (UHC) [7].

The Every Woman Every Child initiative, the World Health Organization’s (WHO) Partnership for Maternal, Neonatal and Child Health, and the objectives of the Global Financing Facility all focus on family planning (FP) and services for routine childhood illnesses [8]. Diarrheal disease, respiratory infections and malaria currently account for 4,756,000 deaths worldwide across all age groups, while maternal causes account for 230,000 deaths [9]. For children under 5, lower respiratory infections account for 13.06% of deaths, with malaria and diarrhea comprising 10.34% and 8.92% of deaths respectively [9]. Many of these deaths can be averted through the expansion of known interventions and services, such as FP [10], which save the lives of both mothers and children.

Family planning and treatment for diarrhea, malaria and respiratory infection are available at a wide range of health service providers in both the public and private health care sectors. Private providers deliver a significant portion of healthcare services in low- and middle-income countries (LMICs) in both rural and urban areas, for low socioeconomic groups and for the wealthy [4]. But research on the equity of private provision of health care is inconclusive [11]. Some analyses have shown that provision of care in the private sector is inherently inequitable, with significantly greater services accessed by the wealthy [5]. Others have shown that strategies such as contracting out services to the private sector can improve equity, or minimally, that a growing private sector does not harm service delivery equity [12, 13].

Use of the private sector varies by region and service type, highlighting the need to better understand where individuals from different wealth quintiles seek healthcare [14]. In a meta-analysis of Demographic and Health Surveys (DHS) data from 57 LMICs, Campbell and colleagues (2015) find that between 15% and 25% of FP users in the poorest quintile use the private sector across all geographic regions (Sub-Saharan Africa, Middle East/Europe, Asia, Latin America), and between 45% and 50% of those in the wealthiest quintile use the private sector [15]. Yet treatment for childhood illnesses differs from treatment location for FP. Globally, over half of all care for children seeking treatment for diarrhea, fever, and cough is provided by private providers, with more urban and wealthy people using the private sector than rural or poorer individuals [14].

While policymakers have often invested more in public sector service delivery, there is growing interest in how the private sector can complement the public sector [16]. For policymakers and program leaders who are designing and implementing interventions to address causes of maternal, neonatal and childhood mortality and morbidity, it is important to understand where individuals of different wealth and geographic residence seek care. For those with initiatives in multiple countries, it is also important to understand if countries behave similarly with regard to service provider choices. This paper explores use and source of family planning, and source of care for children with diarrhea, fever or cough by residence and wealth quintile to draw meaningful conclusions for private sector healthcare provision.

Source of data: This study is a secondary analysis of DHS data regarding source of care for FP services, and childhood diarrhea, fever or cough treatment. The DHS are nationally representative surveys conducted by national implementing agencies with technical assistance from MEASURE DHS and ICF Macro. Countries adapt individual surveys from a model questionnaire, and DHS data files are prepared in a structured way allowing for cross-country comparison [17]. DHS datasets were downloaded from www.measuredhs.com and analyzed using Stata 13/SE. Datasets are publicly available, so no ethical approval was sought.

Analysis methods: The women’s response file, including data from women of reproductive age (15–49 years), from the full DHS dataset was downloaded for the most recent survey from each country. This file contains key household level variables, including urban or rural residence and wealth quintile. The wealth quintile variable is constructed by using principal components analysis on a list of household possessions and housing characteristics, ranking all individuals by their relative wealth, and then dividing the ranking into equal fifths [20]. Quintile 1 refers to the poorest 20% of people in the country, and quintile 5 refers to the wealthiest 20% of people. The wealth quintile reports a measures of relative wealth, and as such, quintile 1 in one country does not indicate the same level of poverty as quintile 1 in another country. As a relative measure of wealth across countries, DHS data have been previously used in cross-country analyses of equity [14, 21–23].

For FP outcomes, analyses were restricted to current users of modern contraception, based upon the variable for current contraceptive method. The recoded source of most recent resupply was used in analyses [17]. These categories are government clinic/pharmacy, government home/community delivery, NGO, private clinic/delivery, pharmacy, shop/church/friend, other. We combined other with the shop/church/friend category for our analyses.

For outcomes of diarrhea, fever and cough, the DHS asks female respondents about illness in the prior two weeks for each of their children under the age of 5. Our analyses focused on the youngest child under age 5 (the first child listed). For diarrhea, respondents are asked if the child had diarrhea in the previous two weeks, “Did you seek advice or treatment for the diarrhea from any source?”, and “Where did you seek advice or treatment?”. Respondents are also asked if the child had a fever in the previous two weeks, if the child had an illness with a cough in the previous two weeks. If the child had a cough, the respondent is further probed as to whether the child “breathe[s] faster than usual with short rapid breaths or [has] difficulty breathing?”. Children whose cough is accompanied by rapid breathing, and those with fever, are asked about sources of care, if any. In the DHS questionnaire, questions about treatment provided for fever and cough are grouped together. Consequently, in our analyses, we have grouped these two outcomes together, rather than dropping observations for children who experienced both symptoms in the two weeks prior to the survey, or reporting on those responses separately. The symptom of cough with rapid breathing is indicative of an acute lower respiratory tract infection (ARI), and the term ARI is used to describe the data.

For each country, sources of care for diarrhea and fever/acute respiratory infection are entered in up to 24 categories. The categorization within a survey is the same for source of care questions for diarrhea and for fever/ARI. To facilitate analysis, these categories were recoded into the same eight groups for all 12 countries, and are presented as such. Children may seek care at more than one place for the same episode, so analyses presented allow for multiple responses. The groups are: government hospital; government health center; government community health workers, mobile clinics or other public sector sources; private sector hospital, clinic or doctor; private sector pharmacy; NGO or mission (faith based) facility; other private qualified provider; unqualified provider. Examples of country-specific sources of care re-categorized into “other private qualified provider” are “home of trained health worker”, “nurse/auxiliary”, and “private – fieldworker”. Examples re-categorized into “unqualified provider” are “market” “drug peddler”, “street vendor”, “friends/relatives”, and “traditional practitioner”.

Weighted univariate, bi-variate and stratified bi-variate tabulations were conducted for FP, for diarrhea and for fever/ARI. Covariates of interest were wealth quintile and urban/rural residence. To explore whether sources of FP and treatment for childhood illness overlapped, we limited our sample across all 12 countries to mother/child pairs who reported both a source of FP and a source of treatment for her child. We then collapsed the two public facility categories, and the “other” categories from the source of treatment for childhood illness, to make them comparable to the FP sources.

Analyses focused on patterns to identify if there were any trends in use of private sector services, which services or providers within the private sector were more favored, and if it was possible to assess these data by wealth quintile. Table 1 lists the sample sizes for each of the 12 countries for both FP and childhood diarrhea and fever/ARI services. Sample sizes varied greatly among countries, indicative of varying illness prevalence, as well as different survey requirements. For example, in Nigeria and Kenya, large samples were collected in order to have region and district-specific estimates for key indicators. Analyses by multiple sub-categories were limited by small sample sizes in some instances.

To measure progress towards Sustainable Development Goal 3, the global community has agreed upon a number of targets. Notably, we will dramatically reduce maternal mortality, end preventable deaths of children under five, and ensure universal access to reproductive health services, while attaining universal health coverage [6]. Given the already high reliance on the private sector for provision of health services in many countries, it is important to determine if support or promotion of the private sector will increase inequity in access to health services [24]. This concern exists because private services are not free, whereas public sector services in many countries are free or subsidized.

The source of care for FP varied significantly across countries when looking at our 12 country sample, with the private sector providing the majority of FP in Bangladesh, Cambodia, DRC, Haiti and Nigeria, and was the source of just under 50% of contraceptives in the Dominican Republic. The DRC, Nigeria, Mali and Senegal all exhibit a similar pattern of contraceptive use with low overall mCPR, and contraceptive use highly skewed towards the wealthiest women. However, our analyses provide further insight, demonstrating that even the wealthy women in Senegal, Mali and Zambia have limited reliance upon the private sector. In these three countries in particular, the private sector is not highly utilized for any health need examined in this article, indicating a generally small private sector footprint.

Limited use of the private sector for FP means that interventions that want to leverage these private providers successfully are non-traditional. One example from Mali is the creation of a public sector social franchise network, where selected primary health facilities receive branding, training and oversight from a social franchisor [25]. By contrast, in Nigeria Shelton and Finkle [26] advocate for expansion of traditional social franchising and social marketing channels, as well as task-sharing so that private patent medical vendors can deliver injectable contraceptives.

The introduction of a new contraceptive method is a good example of the utility of looking at FP use by wealth and source of supply across countries. In the DRC, the Dominican Republic, and Nigeria, private sector pharmacies are the largest provider of short-term methods. Although pharmacies are primarily providing oral contraceptives in these countries, it is likely that the role of pharmacies will expand with the introduction of Sayana® Press in DRC and Nigeria, as well as social marketing of Sayana Press in Nigeria, and significant purchases in Senegal. Of the 12 countries in this study, nine have received some doses of this new contraceptive between 2014 and 2017. Among the largest procurers are Senegal, DRC and Nigeria receiving 761,200 doses, 818,000 doses, and 2,508,000 doses respectively¹ [27]. The private sector will likely play a very large role in providing access to this product, and our data show that women across the socio-economic spectrum use the private sector to obtain their FP method.

When examining the treatment of childhood illnesses, it is important first to understand patterns of prevalence. Our data show that in the Dominican Republic, Ghana, Haiti, Kenya, Liberia and Nigeria, household wealth quintile and diarrhea prevalence have a high inverse correlation. However, only Ghana, Kenya and Liberia have a greater than two percentage point difference in prevalence between urban and rural areas. Combining these two findings, it is indicative of the fact that the poorest are primarily located in rural areas in these countries. Prevalence of fever/ARI is only consistently associated with household wealth in Kenya and Zambia, and in both of these countries, prevalence of fever/ARI is lower in the poorest quintile than in quintile 2. There is a greater than two percentage point difference in prevalence between urban and rural areas in Bangladesh, Ghana, Haiti, Kenya, Liberia, Zambia, Dominican Republic and Senegal, with the latter two countries having a higher prevalence in urban areas.

In our 12 country sample, the Asian countries and Nigeria had substantially greater use of the private sector for treatment of both diarrhea and fever/ARI, consistent with trends seen across all LMICs [15]. Among those who seek treatment for diarrhea or fever/ARI, use of the private sector is positively correlated with household wealth quintile, even in countries with an overall low reliance upon the private sector. Patterns differ by type of private sector facility, with less than 10% of sick children in the poorest quintile visiting a private clinic in Kenya, Haiti and Liberia, compared with 25%–40% of those in the richest quintile in those same countries. Use of private sector pharmacies, by contrast, is much more equitable, especially in the Dominican Republic, Ghana, Kenya and Nigeria. However, many childhood illnesses are not treated in the formal medical sector, revealing an unmet need for quality care.

Across the 12 comparison countries, more than one-third of all children did not receive treatment for diarrhea. Over 50% of children with fever/ARI in Haiti and Mali received no treatment, or treatment from an unqualified source. The proportion of children who received no treatment for fever/ARI was strongly negatively correlated with wealth status in Haiti, Liberia, Mali and Senegal, with a similar finding for no treatment for diarrhea only present in Haiti. It can be argued that not all episodes of diarrhea or fever require formal medical care. In fact, promotion of appropriate home treatment of diarrhea with oral rehydration solution has been part of global recommendations since the 1970s, and has been estimated to reduce 93% of all diarrheal mortality in children under 5 with 100% coverage [28]. However, given that almost 40% of under 5 mortality is caused by diarrheal disease, respiratory infection and malaria, the inequity in the proportion of children who do not receive any treatment necessitates further investigation [29].

The private sector is not a homogeneous entity. It contains different types of providers, ranging from formal medical clinics, to registered pharmacies, drug sellers, general retailers, faith-based and non-governmental organizations, and can be for profit or non-profit. As we have seen in our analyses, there is substantial variation in outpatient care-seeking through the private sector across countries [15, 22, 30]. A deeper examination of the total relevance of the private sector in providing care for specific services, plus the relative usage of the private sector by wealth group and urban/rural residence, can help policymakers direct resources towards improving coverage for needed interventions.

Several systematic reviews have examined the use of the private sector in outpatient health services, including whether private sector providers improve coverage, have quality services or provide services efficiently and equitably [31–34]. Others have examined inequity in health services provision, with notable efforts such as the Countdown to 2015 or meta-analyses of national household survey data for inequity, specifically for child health or FP services [1, 5, 35, 36]. Few have combined both lines of inquiry, and investigated the use of the private sector by wealth status [13, 14, 21]. Of these, Campbell et al., and Grepin aggregate data from 57 and 70 countries, respectively, masking variation, while Hotchkiss focuses only on contraceptives. Our analyses focused on two types of services which are routinely provided in the private sector, and which are services that usually require a payment. Data is disaggregated by country to allow for more nuanced analyses. Improvement in use of FP and diarrhea and fever/ARI services for those in need, who have been shown to be more likely to be poor, will lead to lower mortality and morbidity among women and children [1, 28, 37, 38].

This investigation of type of care sought by wealth, residence and reason for need across 12 countries combines information often found in siloes, such as only looking at FP care seeking, or investigating wealth inequities without addressing type of care sought. Overall, parents were more likely to take their children to public sector services for the treatment of childhood illnesses across all wealth quintiles. However, parents chose the private sector more often for treatment of fever/ARI than for diarrhea. Reasons would have to be explored, but present an opportunity to increase use of the same private sector facilities for diarrhea treatment, as well as among those seeking no treatment across all wealth quintiles. Contrasted to diarrhea and fever/ARI services, use of public and private sector services for FP varied substantially country by country, and use of modern contraception was not correlated with the use of public or private services. Furthermore, there was little correlation between use of the private sector for childhood illness, and use of the private sector for FP in the same household (between a woman and her child). This secondary analysis of DHS data demonstrated differences in where individuals from different wealth quintiles sought FP services and care for children with diarrhea and fever/ARI. Additional research to understand why motivations for seeking private services differ for FP and childhood illness may help to inform strategies to improve coverage of care.

In an era where policymakers, ministry officials and program leaders are continually pressed to do more with less, while simultaneously striving to implement UHC and supporting policies that improve access without causing financial hardship, it is important to understand where individuals of different socioeconomic status seek care for different health services. Being able to leverage a successful program to increase use of family planning through investments in the private sector along with improvements in appropriate care for children can stretch funds for greater results. Similarly, advocating for increased availability of a range of family planning methods at the pharmacies where treatment for diarrhea and fever are sought may improve access for women. Coordination across public and private sectors may help to streamline and maximize the benefits of the services. Examining the role of the private sector furthers our understanding of its utility in expanding access to services across wealth quintiles and providing equitable care.