Meeting need vs. sharing the market: a systematic review of methods to measure the use of private sector family planning and childbirth services in sub-Saharan Africa

Forty-seven of the 53 studies used household survey data to estimate use of private sector family planning and/or childbirth services; the majority of these studies used the Demographic and Health Surveys (n = 27), while others used data from demographic surveillance sites [34, 35], national maternal health surveys [36], or other smaller, sub-national surveys [23, 24, 37‐50]. Three studies conducted surveys that sampled women at a health facility [51], market [52], or through respondent-driven sampling [53]. The remaining three studies used routine health service statistics to estimate the proportion of facility births that occurred within the private sector [54‐56].

The studies included in this review contained over 40 unique terms to describe private sector sources of family planning and childbirth services (Additional file 5). Throughout the literature, the “private sector” referred to a range of for-profit, not-for-profit, faith-based, medical, and informal providers that were managed by non-government actors. While some studies reported on service use from any private sector source, others distinguished private sector providers by two key characteristics: (a) their commercialization or profit orientation and/or (b) their position within or outside of the medical sector (Fig. 3). Among the included studies, the private for-profit sector included both medical and non-medical providers, while the private non-profit sector seemed to refer exclusively to medical providers. Additional file 6 displays the frequency with which each unique private sector term appeared in the included studies, categorized by profit orientation.

The studies included in this review examined private sector use within two general population groups: (1) women in need or “at risk” of needing family planning or childbirth services and (2) users of those services. We used the term private sector coverage to indicate the proportion of women in need who were using family planning or childbirth services from a private sector source, or the proportion of health service need met by the private sector. Private sector market share, on the other hand, refers to the proportion of family planning or childbirth service users who received care from a private sector source. Of the 31 studies that examined use of private sector family planning services, five reported coverage estimates [20, 21, 57‐59], 30 reported market share estimates, and four reported both market share and coverage estimates [20, 21, 57, 59] (Additional file 6). For childbirth services, 22 of 26 papers presented private sector coverage estimates, seven reported on private sector market share [35, 52, 54‐57, 60], and three reported both market share and coverage estimates [35, 57, 61] (Additional file 6). Although we grouped these outcomes into two broad categories, there was substantial variation within each category in how these populations were defined. For instance, three studies considered the population in need of family planning to be all women married or in union (regardless of fertility preferences or desires) [21, 58, 59], while two studies reported use among all women with d need for family planning (regardless of marital status) [20, 57], following the most recent consensus definition of need for contraception [62]. We summarized the different populations used to examine private sector coverage and market share in Fig. 4.

There were differences between studies in the unit of analysis for examining source of family planning and childbirth services. Of the 31 studies that reported on use of private sector family planning services, 13 described source of care during women’s most recent supply of contraceptives [20, 22, 24‐26, 29, 31, 41, 50, 53, 58, 59, 63] and one described where women obtained care when they first started using their current method [40]. The majority of studies (n = 17), however, did not state their unit of analysis [19, 21, 27, 28, 38, 39, 42, 44, 45, 51, 57, 64‐68].

When describing use of childbirth services, the included studies generally adopted either a birth-based (n = 6) [23, 54‐56, 69, 70] or a woman-based approach (n = 13) [24, 30, 36, 37, 43, 46‐50, 57, 60, 68]. The unit of analysis for birth-based approaches was all births that occurred over the study recall or review period. The woman-based approach, on the other hand, included only one birth per woman, and all of the included studies adopting this approach used a woman’s most recent birth as the unit of analysis. One study used a hybrid approach, taking information about all of a woman’s births and categorized her according to where she sought care across births [32]. Six of the 26 studies that reported on use of private sector childbirth services did not state their unit of analysis [34, 35, 52, 71‐73].

Although missing information on source of care and family planning and childbirth service use and need has the potential to bias estimates, relatively few studies described how they treated such missing information.

Of the studies that examined private sector market share for family planning and childbirth services, only 9 out of 30 [20, 24, 25, 27, 40, 42, 57, 63, 67] and two out of seven [57, 60], respectively, indicated how missing data on source of care was treated. Among the studies that did provide this information, the convention was to either include women with missing information as part of the market, but report their source of care as missing or unknown, or to exclude them from the market entirely.

Of the five studies that reported on private sector coverage of family planning services, only two described how they treated women with missing data on family planning need [20, 57] and three described how they treated women with missing data on source of care [20, 57, 58]. In those studies, women with incomplete information on family planning need were considered to be not in need of contraception, while those with missing information on source of care were considered to not have received care in the private sector. For childbirth services, all reported deliveries were considered in need of care. Campbell et al. [8] is the only study that discussed missing information on delivery care need, and the authors found no missing data for that variable [57]. Six of the 22 studies that examined private sector coverage of childbirth services reported that women with missing information on source of care were considered to have not received care in the private sector [24, 35, 36, 43, 57, 60].

Ugaz et al. [21] estimated that private sector coverage among women married or in union was relatively low in the sub-Saharan Africa region, ranging between 3 and 6% from 1992 to 2012. Looking exclusively at the population of women in need of contraception, regardless of marital status, Campbell et al. [20, 57] estimated higher private sector coverage, with 14% private sector coverage of modern contraceptive need. Similarly, comparing Fig. 5a and b, we observed that estimates of private sector family planning coverage tended to be higher among women in need compared to all women married or in union [20, 21, 58, 59]. As expected, the proportions of women using modern methods from private non-profit or for-profit providers was smaller than the proportion of women using modern methods from any private sector source.

Benova et al. (2015) estimated that 10% of women in sub-Saharan Africa delivered in the private sector for their most recent birth, either at a private facility or in a non-facility location with a private medical provider, while Wodon et al. (2012) generated a lower estimate, at 6.8% of women [61, 68]. Yoong et al. [70] estimated that an average of 7.7% of all births in the region received childbirth care from a private medical provider, specifically in a private medical facility. In Fig. 5c and d, estimates of private sector coverage appeared to be quite similar and increasing between the 1985–1999 and 2000–2014 periods, with median values increasing from 5 to 14 and 4 to 11% among all births and most recent births, respectively [23, 24, 30, 36, 46, 48, 50, 68, 69]. Private non-profit provider coverage of most recent births was estimated around 17% in both periods, and private for profit provider coverage appeared negligible; however, the studies reporting on this outcome were conducted exclusively in rural areas in Kenya and Tanzania with access to a mission hospital [46, 49].

We observed much greater heterogeneity in estimates of private sector market share estimates for sub-Saharan Africa compared to those of private sector coverage. Campbell et al. [8, 20, 57] estimated that 35% of all modern family planning users and 38% of modern family planning users who received their method from a source with a classifiable sector obtained care from a private sector provider. Figure 6g shows that the family planning market share among women who obtained care from a source with a classifiable sector ranged from 6% in Rwanda (2010) to 80% in Gabon (2012) [20].

Both Ugaz et al. [21] and Wodon et al. [68] estimated that approximately 28% of all modern family planning users in sub-Saharan Africa received their method from a private medical provider. This ranged from countries with less than 2% private medical market share (Burundi, 1987; Sao Tome and Principe, 2008/9) to countries with greater than 60% private medical market share (Democratic Republic of the Congo, 2007; Nigeria, 2008) (Fig. 6e) [50, 68, 74].

Among all users of modern contraception and those married or in union, private for-profit providers appeared to have a greater market share compared to private non-profit providers (Figs. 6e & 6f) [22, 29, 33, 41, 59, 66‐68]. There also seemed to be greater use of private medical providers compared to non-medical providers among current users of modern contraception; however, this may be because Wodon et al. (2012) classified pharmacies as medical providers while others distinguished facilities from pharmacies or specialized drug sellers (Fig. 6e) [50, 67, 68].

Only two studies comprehensively examined childbirth service market share across a large number of countries [57, 61]. Benova and colleagues estimated that 20% of women who gave birth under appropriate care conditions (in a facility or with a skilled-birth attendant) in sub-Saharan Africa received care in the private sector (Fig. 7h) [61]. Looking at source of care among women who received appropriate care from a provider with a classifiable sector increased this estimate slightly to 22% (Fig. 7j) [57, 61]. Three smaller studies from Kenya, South Africa, and Uganda examined private sector market share among facility births, and estimates ranged from 15% (South Africa, 1990) to 36% (Uganda, 2007) [35, 54, 55]. One study from Kenya found that private for- profit providers had a greater market share among facility births (10%) compared to private non-profit providers (3%) [54]. In contrast to private sector family planning market share, none of the estimates of private sector market share for childbirth services exceeded 45%.