Tropical Medicine and International Health
doi:10.1111/tmi.12414
volume 20 no 2 pp 230–239 february 2015
Series: Who cares for women? Towards a greater understanding of reproductive and maternal healthcare markets
The role of the private sector in the provision of antenatal care: a study of Demographic and Health Surveys from 46 low- and middle-income countries Timothy Powell-Jackson1, David Macleod2, Lenka Benova2, Caroline Lynch2 and Oona M. R. Campbell2 1 Faculty of Public Health and Policy, London School of Hygiene and Tropical Medicine, London, UK 2 Faculty of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, UK
Abstract
objective To examine the role of the private sector in the provision of antenatal care (ANC) across low- and middle-income countries. methods Demographic and Health Survey (DHS) data from 46 countries (representing 2.6 billion people) on components of ANC given to 303 908 women aged 15–49 years for most recent birth were used. We identified 79 unique sources of care which were re-coded into home, public, private (commercial) and private (not-for-profit). Use of ANC and a quality of care index (scaled 0–1) were stratified by type of provider, region and wealth quintile. Linear regressions were used to examine the association between provider type and antenatal quality of care score. results Across all countries, the main source of ANC was public (54%), followed by private commercial (36%) and home (5%), but there were large variations by region. Home-based ANC was associated with worse quality of care (0.2; 95% CI 0.2 to 0.19) relative to the public sector, while the private not-for-profit sector (0.03; 95% CI 0.02 to 0.04) was better. There were no differences in quality of care between public and private commercial providers. conclusions The market for ANC varies considerably between regions. The two largest sectors – public and private commercial – perform similarly in terms of quality of care. Future research should examine the role of the private sector in other health service domains across multiple countries and test what policies and programmes can encourage private providers to contribute to increased coverage, quality and equity of maternal care. keywords antenatal care, maternal health, private providers, developing countries, quality of care, socioeconomic gradient
Introduction The role of the private sector in low- and middle-income countries has been the subject of heated debate in both the academic literature and policy circles (Hanson et al. 2008). Even agreement on the size of the private sector is hard to come by (Wagstaff 2013). Those in favour of an expansion in the role of the private sector argue that market forces improve the efficiency and responsiveness of health services. Advocates of a strong public sector point towards the inequitable nature of private sector provision and the fact that markets often fail in health – for example in the provision of preventive care. Despite these differences, both camps would agree that more evidence on the perfor230
mance of public and private health providers is needed to inform the debate on the relative merits of the two sectors. The diversity of the private sector can make simple comparisons with public providers difficult to interpret. Nevertheless, there have been a number of efforts to review the broad literature. One expansive review found that the private providers had poorer patient outcomes and were more likely to flout standards, but were responsive to patients (Basu et al. 2012). Perverse incentives for unnecessary care in the private sector meant efficiency was lower than in the public sector. Availability of inputs (e.g. trained healthcare workers) was typically worse in the public sector. A second review focused on the quality of ambulatory health care found that many services,
© 2014 John Wiley & Sons Ltd
Tropical Medicine and International Health
volume 20 no 2 pp 230–239 february 2015
T. Powell-Jackson et al. Antenatal care and the private sector
irrespective of whether public or private, scored low on infrastructure, clinical competence and practice (Berendes et al. 2011). The private sector performed better in relation to drug supply, responsiveness and effort. No difference between provider groups was detected for patient satisfaction or competence. Both reviews are notable for the lack of cross-country evidence on the performance of public and private providers. In this article, we examine the role of the private sector in the provision of antenatal care (ANC) across low- and middle-income countries. We focus on ANC for two important reasons. First, ANC and its components are widely recognised as critical for the health of the mother and her baby. Evidence suggests that routine ANC visits matter for maternal and newborn outcomes (Villar et al. 2001; Dowswell et al. 2010), even if the precise number of recommended visits is disputed (WHO 2006; Hofmeyr & Hodnett 2013; Vogel et al. 2013). Indeed, use of ANC is a core indicator for monitoring progress towards the fifth MDG concerning universal access to reproductive health. Second, ANC provides one of few entry points to examine process of care measures at the patient level across a large sample of countries. The data we used contain information on whether women received specific components of care during an ANC consultation, giving us the opportunity to develop an overall quality of care measure that can be compared between different types of health provider.
Methods Data Demographic and Health Surveys are cross-sectional nationally representative household surveys, usually covering 5000–30 000 households. In the DHS, standard model questionnaires are used but can be adapted by each country; optional modules can also be added. Manuals and technical assistance ensure that the survey procedures followed in each country are similar, providing comparable data across countries. The surveys include questions on household and individual characteristics, fertility and family planning, maternal and child health and details on antenatal and delivery care. The sampling design is a multilevel cluster survey, which often oversamples certain areas. For each country, we obtained the most recent DHS survey conducted between 2000 and 2012 from the DHS Measure website. Data sets on 57 countries were available, of which 47 had information about ANC provision. Bangladesh only had one of the eight components used to generate the content of care measure and therefore was excluded. Our analysis was based on 46 countries, listed in
© 2014 John Wiley & Sons Ltd
Appendix 1, with information on 303 908 women aged 15–49 years with a recent birth. We grouped countries according to DHS regions, in line with other studies (Montagu et al. 2011). There were 23 countries in subSaharan Africa, eight countries in the North Africa/West Asia/Europe Region, eight countries in the South/SouthEast Asia Region and seven countries in Latin America & the Caribbean. These represent 68%, 25%, 83% and 20% percentage of the populations of these regions, respectively. The 46 countries represent a total population of 2.6 billion people. Quality of care measure The DHS asked each woman who reported using ANC whether she received specific services during any consultation. Each survey included questions relating to up to eight components of ANC – including whether the woman was weighed, had her blood pressure taken, had a urine sample taken, had a blood sample taken, took intestinal parasite drugs, took iron tablets or syrup, took malaria prophylaxis or received information on pregnancy complications. Not every intervention is needed in a particular setting. Hence, the number of questions varied by country, ranging from five to eight in our sample. Quality of care was considered by looking at the number of antenatal care components received (Hodgins & D’Agostino 2014). For every woman who received at least one ANC visit, we generated a score, calculated as the number of components received divided by the number of components asked about in that country. We use the terms ANC content of care and ANC quality interchangeably when referring to this measure. The score ranges from zero (no components received) to one (all components received). Aggregating items in this way provides a score that places equal weight on each component and is thus easy to interpret. It is worth noting, however, that some components of ANC may be more effective than others in promoting better health outcomes. Source of provision Demographic and Health Surveys collect information on where women sought ANC. Women can specify multiple responses, for example if they had more than one visit. While provider categories are reasonably standardised, some flexibility is allowed to reflect the country context. Accordingly, we identified 79 unique types of ANC provider in the data. Using these labels, we re-coded the responses into the following four categories that form the basis of our analysis (i) home, (ii) public, (iii) private (commercial) and (iv) private (not-for profit), that is, run 231
Tropical Medicine and International Health
volume 20 no 2 pp 230–239 february 2015
T. Powell-Jackson et al. Antenatal care and the private sector
by a faith-based or non-governmental organisation. A total of 6364 women used both public and private providers, accounting for 2.1% of the sample. Because this category has no clear interpretation and is negligible, we dropped these observations from the analysis. Statistical analysis After a basic description of the data, we used OLS to examine the association between the type of provider sought for ANC and the content of care score, adjusted for various potential confounding factors. Covariates included in the models were urban residence, parity, education, age, age squared, asset wealth, country and the number of ANC visits. The latter controls for the fact that women seeking more ANC have more opportunities to receive each component of care. We ran a regression of the ANC quality score on a set of dummies indicating whether the visit was at a private commercial provider, private not-for-profit provider or at home. The public sector was omitted and thus serves as the reference category. We used the entire sample across all countries and then stratified the analysis by region and wealth quintile. Stata (version 13.0) was used to conduct the analysis. Features of the complex survey design were accounted for in the analysis. Specifically, we adjusted standard errors for clustering at the primary sampling unit and applied weights that accounted for both survey design and country population, using UN population estimates for 2005, to ensure that estimates are representative of the entire population residing in the study countries. Results Figure 1 shows the source of ANC by region. In aggregate, the public sector has the largest market share (54.0%), followed by the private commercial sector (36.0%). In sub-Saharan Africa and Latin America & the Caribbean, the public sector was dominant, providing 81.2% and 86.3% of care, respectively. By contrast, it was the private commercial sector that was largest in the North Africa/West Asia/Europe (52.5%) and South/ South-East Asia (46.4%) regions. The private not-forprofit sector represents a small fraction of ANC provision, both overall and within regions, even in sub-Saharan Africa where its market share was only 3.2%. It is worth bearing in mind that the percentage of women without any ANC was 20.1% across the study countries (21.1% in sub-Saharan Africa, 19.2% in North Africa/ West Asia/Europe, 20.5% in South/South-East Asia and 7.8% in Latin America & the Caribbean).
232
Figure 2 shows the distribution of the content of care score in the study countries, stratified by type of provider. As expected, a large proportion of women given ANC at home received few of the individual components, with a mean score of 0.28. While differences between public, private commercial and private not-for-profit exist, they were not substantial. The private not-for-profit sector has the highest ANC score (0.71), followed by private commercial (0.63) and public providers (0.59). Clearly, there is substantial room for improvement across all sectors. There was a large spread in ANC quality, irrespective of provider type, with a long tail suggesting that a large proportion of women received poor quality of care. Figure 3 plots the ANC score for each wealth quintile by type of provider and region. Several points are of note. First, wealthier women consistently received a higher proportion of ANC components during pregnancy. Second, the gap between wealth quintiles in the content of care was greater amongst private commercial providers than those in the public or private not-for-profit sectors, perhaps reflecting the diversity of providers operating in the private commercial sector. Only in one region, South/ South-East Asia, was this not the case. Third, the private not-for-profit sector consistently provided the highest quality of care, with the least variation between wealth groups. For completeness, we show in Appendix 2 the mean coverage and the percentage point gap in coverage between the poorest and richest quintile for each component of ANC by type of provider and region. We next turn to the regression results. Table 1 examines the relationship between ANC quality and the source of care. Across the entire sample, home care was associated with substantially lower quality than public health providers, the reference category. The private not-for-profit sector was significantly correlated with better care relative to public providers, but the magnitude of 0.031 was modest compared to the mean ANC score of 0.59 in the public sector. The difference in ANC quality between the private commercial and public sectors appears negligible. The results stratified by region are notable in several respects. ANC quality was worse in the private commercial sector than the public sector in all regions except South/SouthEast Asia, where it was better largely because of India. Indeed, when we exclude India, there is a negative and significant association between ANC quality and private commercial providers (relative to public), indicating that the private commercial sector was worse. ANC quality was highest in the private not-for-profit sector, although significantly so only in the sub-Saharan Africa region. Finally, Table 2 shows the differences in ANC quality between types of provider within wealth quintiles across all study countries. The private not-for-profit sector con-
© 2014 John Wiley & Sons Ltd
Tropical Medicine and International Health
volume 20 no 2 pp 230–239 february 2015
T. Powell-Jackson et al. Antenatal care and the private sector
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%
All countries
Sub-Saharan North Africa/ South/Southeast Latin America Africa West asia/Europe Asia & Caribbean
Home
Public
Public & Private
Private not-forprofit
Private commercial
0
0.5
Density 1
1.5
2
Figure 1 Sources of antenatal care.
0
Figure 2 Antenatal content of care by type of provider.
sistently provided better quality than the public sector, irrespective of whether the woman was poor or rich. Private commercial providers performed worse than the public sector for the poorest one-fifth of women. This difference becomes either negligible or positive for women in the higher wealth quintiles. This pattern is driven entirely by India. When we exclude India, we find that the
© 2014 John Wiley & Sons Ltd
0.2
0.4 0.6 Antenatal content of care score
Public Private - Not for Profit
0.8
1
Private - Commercial Home
private commercial sector provided a lower level of ANC quality in all five wealth groups. Discussion We examined the content of ANC using nationally representative data in 46 low- and middle-income countries,
233
Tropical Medicine and International Health
volume 20 no 2 pp 230–239 february 2015
T. Powell-Jackson et al. Antenatal care and the private sector
Poorest
Poorer
Middle
Richer
Richest
100% 90% 80% 70% 60% 50% 40% 30% 20%
North Africa/West Asia/Europe
Home Public Private commercial Private not-for-profit
Sub-Saharan Africa
Home Public Private commercial Private not-for-profit
All countries
Home Public Private commercial Private not-for-profit
Home Public Private commercial Private not-for-profit
0%
Home Public Private commercial Private not-for-profit
10%
South/Southeast Asia
Latin America & Caribbean
Figure 3 Antenatal content of care by provider type and wealth group.
Table 1 Association of provider type with antenatal content of care score
n All countries (46) Private commercial 48 559 Private not-for-profit 6133 Home 6882 Sub-Saharan Africa (23) Private commercial 8273 Private not-for-profit 5317 Home 2306 North Africa/West Asia/Europe (8) Private commercial 9487 Private not for-profit 400 Home 314 South/South-East Asia (8) Private commercial 24 793 Private not-for-profit 314 Home 4100 Latin America & Caribbean (7) Private commercial 6006 Private not-for-profit 102 Home 162
Weighted percentage
Adjusted estimate (95% CI)
P value
0.375 0.012 0.055
0.003 ( 0.0004 to –0.007) 0.031 (0.024 to –0.039) 0.197 ( 0.207 to 0.186)
0.078
doi:10.1111/tmi.12414
volume 20 no 2 pp 230–239 february 2015
Series: Who cares for women? Towards a greater understanding of reproductive and maternal healthcare markets
The role of the private sector in the provision of antenatal care: a study of Demographic and Health Surveys from 46 low- and middle-income countries Timothy Powell-Jackson1, David Macleod2, Lenka Benova2, Caroline Lynch2 and Oona M. R. Campbell2 1 Faculty of Public Health and Policy, London School of Hygiene and Tropical Medicine, London, UK 2 Faculty of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, UK
Abstract
objective To examine the role of the private sector in the provision of antenatal care (ANC) across low- and middle-income countries. methods Demographic and Health Survey (DHS) data from 46 countries (representing 2.6 billion people) on components of ANC given to 303 908 women aged 15–49 years for most recent birth were used. We identified 79 unique sources of care which were re-coded into home, public, private (commercial) and private (not-for-profit). Use of ANC and a quality of care index (scaled 0–1) were stratified by type of provider, region and wealth quintile. Linear regressions were used to examine the association between provider type and antenatal quality of care score. results Across all countries, the main source of ANC was public (54%), followed by private commercial (36%) and home (5%), but there were large variations by region. Home-based ANC was associated with worse quality of care (0.2; 95% CI 0.2 to 0.19) relative to the public sector, while the private not-for-profit sector (0.03; 95% CI 0.02 to 0.04) was better. There were no differences in quality of care between public and private commercial providers. conclusions The market for ANC varies considerably between regions. The two largest sectors – public and private commercial – perform similarly in terms of quality of care. Future research should examine the role of the private sector in other health service domains across multiple countries and test what policies and programmes can encourage private providers to contribute to increased coverage, quality and equity of maternal care. keywords antenatal care, maternal health, private providers, developing countries, quality of care, socioeconomic gradient
Introduction The role of the private sector in low- and middle-income countries has been the subject of heated debate in both the academic literature and policy circles (Hanson et al. 2008). Even agreement on the size of the private sector is hard to come by (Wagstaff 2013). Those in favour of an expansion in the role of the private sector argue that market forces improve the efficiency and responsiveness of health services. Advocates of a strong public sector point towards the inequitable nature of private sector provision and the fact that markets often fail in health – for example in the provision of preventive care. Despite these differences, both camps would agree that more evidence on the perfor230
mance of public and private health providers is needed to inform the debate on the relative merits of the two sectors. The diversity of the private sector can make simple comparisons with public providers difficult to interpret. Nevertheless, there have been a number of efforts to review the broad literature. One expansive review found that the private providers had poorer patient outcomes and were more likely to flout standards, but were responsive to patients (Basu et al. 2012). Perverse incentives for unnecessary care in the private sector meant efficiency was lower than in the public sector. Availability of inputs (e.g. trained healthcare workers) was typically worse in the public sector. A second review focused on the quality of ambulatory health care found that many services,
© 2014 John Wiley & Sons Ltd
Tropical Medicine and International Health
volume 20 no 2 pp 230–239 february 2015
T. Powell-Jackson et al. Antenatal care and the private sector
irrespective of whether public or private, scored low on infrastructure, clinical competence and practice (Berendes et al. 2011). The private sector performed better in relation to drug supply, responsiveness and effort. No difference between provider groups was detected for patient satisfaction or competence. Both reviews are notable for the lack of cross-country evidence on the performance of public and private providers. In this article, we examine the role of the private sector in the provision of antenatal care (ANC) across low- and middle-income countries. We focus on ANC for two important reasons. First, ANC and its components are widely recognised as critical for the health of the mother and her baby. Evidence suggests that routine ANC visits matter for maternal and newborn outcomes (Villar et al. 2001; Dowswell et al. 2010), even if the precise number of recommended visits is disputed (WHO 2006; Hofmeyr & Hodnett 2013; Vogel et al. 2013). Indeed, use of ANC is a core indicator for monitoring progress towards the fifth MDG concerning universal access to reproductive health. Second, ANC provides one of few entry points to examine process of care measures at the patient level across a large sample of countries. The data we used contain information on whether women received specific components of care during an ANC consultation, giving us the opportunity to develop an overall quality of care measure that can be compared between different types of health provider.
Methods Data Demographic and Health Surveys are cross-sectional nationally representative household surveys, usually covering 5000–30 000 households. In the DHS, standard model questionnaires are used but can be adapted by each country; optional modules can also be added. Manuals and technical assistance ensure that the survey procedures followed in each country are similar, providing comparable data across countries. The surveys include questions on household and individual characteristics, fertility and family planning, maternal and child health and details on antenatal and delivery care. The sampling design is a multilevel cluster survey, which often oversamples certain areas. For each country, we obtained the most recent DHS survey conducted between 2000 and 2012 from the DHS Measure website. Data sets on 57 countries were available, of which 47 had information about ANC provision. Bangladesh only had one of the eight components used to generate the content of care measure and therefore was excluded. Our analysis was based on 46 countries, listed in
© 2014 John Wiley & Sons Ltd
Appendix 1, with information on 303 908 women aged 15–49 years with a recent birth. We grouped countries according to DHS regions, in line with other studies (Montagu et al. 2011). There were 23 countries in subSaharan Africa, eight countries in the North Africa/West Asia/Europe Region, eight countries in the South/SouthEast Asia Region and seven countries in Latin America & the Caribbean. These represent 68%, 25%, 83% and 20% percentage of the populations of these regions, respectively. The 46 countries represent a total population of 2.6 billion people. Quality of care measure The DHS asked each woman who reported using ANC whether she received specific services during any consultation. Each survey included questions relating to up to eight components of ANC – including whether the woman was weighed, had her blood pressure taken, had a urine sample taken, had a blood sample taken, took intestinal parasite drugs, took iron tablets or syrup, took malaria prophylaxis or received information on pregnancy complications. Not every intervention is needed in a particular setting. Hence, the number of questions varied by country, ranging from five to eight in our sample. Quality of care was considered by looking at the number of antenatal care components received (Hodgins & D’Agostino 2014). For every woman who received at least one ANC visit, we generated a score, calculated as the number of components received divided by the number of components asked about in that country. We use the terms ANC content of care and ANC quality interchangeably when referring to this measure. The score ranges from zero (no components received) to one (all components received). Aggregating items in this way provides a score that places equal weight on each component and is thus easy to interpret. It is worth noting, however, that some components of ANC may be more effective than others in promoting better health outcomes. Source of provision Demographic and Health Surveys collect information on where women sought ANC. Women can specify multiple responses, for example if they had more than one visit. While provider categories are reasonably standardised, some flexibility is allowed to reflect the country context. Accordingly, we identified 79 unique types of ANC provider in the data. Using these labels, we re-coded the responses into the following four categories that form the basis of our analysis (i) home, (ii) public, (iii) private (commercial) and (iv) private (not-for profit), that is, run 231
Tropical Medicine and International Health
volume 20 no 2 pp 230–239 february 2015
T. Powell-Jackson et al. Antenatal care and the private sector
by a faith-based or non-governmental organisation. A total of 6364 women used both public and private providers, accounting for 2.1% of the sample. Because this category has no clear interpretation and is negligible, we dropped these observations from the analysis. Statistical analysis After a basic description of the data, we used OLS to examine the association between the type of provider sought for ANC and the content of care score, adjusted for various potential confounding factors. Covariates included in the models were urban residence, parity, education, age, age squared, asset wealth, country and the number of ANC visits. The latter controls for the fact that women seeking more ANC have more opportunities to receive each component of care. We ran a regression of the ANC quality score on a set of dummies indicating whether the visit was at a private commercial provider, private not-for-profit provider or at home. The public sector was omitted and thus serves as the reference category. We used the entire sample across all countries and then stratified the analysis by region and wealth quintile. Stata (version 13.0) was used to conduct the analysis. Features of the complex survey design were accounted for in the analysis. Specifically, we adjusted standard errors for clustering at the primary sampling unit and applied weights that accounted for both survey design and country population, using UN population estimates for 2005, to ensure that estimates are representative of the entire population residing in the study countries. Results Figure 1 shows the source of ANC by region. In aggregate, the public sector has the largest market share (54.0%), followed by the private commercial sector (36.0%). In sub-Saharan Africa and Latin America & the Caribbean, the public sector was dominant, providing 81.2% and 86.3% of care, respectively. By contrast, it was the private commercial sector that was largest in the North Africa/West Asia/Europe (52.5%) and South/ South-East Asia (46.4%) regions. The private not-forprofit sector represents a small fraction of ANC provision, both overall and within regions, even in sub-Saharan Africa where its market share was only 3.2%. It is worth bearing in mind that the percentage of women without any ANC was 20.1% across the study countries (21.1% in sub-Saharan Africa, 19.2% in North Africa/ West Asia/Europe, 20.5% in South/South-East Asia and 7.8% in Latin America & the Caribbean).
232
Figure 2 shows the distribution of the content of care score in the study countries, stratified by type of provider. As expected, a large proportion of women given ANC at home received few of the individual components, with a mean score of 0.28. While differences between public, private commercial and private not-for-profit exist, they were not substantial. The private not-for-profit sector has the highest ANC score (0.71), followed by private commercial (0.63) and public providers (0.59). Clearly, there is substantial room for improvement across all sectors. There was a large spread in ANC quality, irrespective of provider type, with a long tail suggesting that a large proportion of women received poor quality of care. Figure 3 plots the ANC score for each wealth quintile by type of provider and region. Several points are of note. First, wealthier women consistently received a higher proportion of ANC components during pregnancy. Second, the gap between wealth quintiles in the content of care was greater amongst private commercial providers than those in the public or private not-for-profit sectors, perhaps reflecting the diversity of providers operating in the private commercial sector. Only in one region, South/ South-East Asia, was this not the case. Third, the private not-for-profit sector consistently provided the highest quality of care, with the least variation between wealth groups. For completeness, we show in Appendix 2 the mean coverage and the percentage point gap in coverage between the poorest and richest quintile for each component of ANC by type of provider and region. We next turn to the regression results. Table 1 examines the relationship between ANC quality and the source of care. Across the entire sample, home care was associated with substantially lower quality than public health providers, the reference category. The private not-for-profit sector was significantly correlated with better care relative to public providers, but the magnitude of 0.031 was modest compared to the mean ANC score of 0.59 in the public sector. The difference in ANC quality between the private commercial and public sectors appears negligible. The results stratified by region are notable in several respects. ANC quality was worse in the private commercial sector than the public sector in all regions except South/SouthEast Asia, where it was better largely because of India. Indeed, when we exclude India, there is a negative and significant association between ANC quality and private commercial providers (relative to public), indicating that the private commercial sector was worse. ANC quality was highest in the private not-for-profit sector, although significantly so only in the sub-Saharan Africa region. Finally, Table 2 shows the differences in ANC quality between types of provider within wealth quintiles across all study countries. The private not-for-profit sector con-
© 2014 John Wiley & Sons Ltd
Tropical Medicine and International Health
volume 20 no 2 pp 230–239 february 2015
T. Powell-Jackson et al. Antenatal care and the private sector
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%
All countries
Sub-Saharan North Africa/ South/Southeast Latin America Africa West asia/Europe Asia & Caribbean
Home
Public
Public & Private
Private not-forprofit
Private commercial
0
0.5
Density 1
1.5
2
Figure 1 Sources of antenatal care.
0
Figure 2 Antenatal content of care by type of provider.
sistently provided better quality than the public sector, irrespective of whether the woman was poor or rich. Private commercial providers performed worse than the public sector for the poorest one-fifth of women. This difference becomes either negligible or positive for women in the higher wealth quintiles. This pattern is driven entirely by India. When we exclude India, we find that the
© 2014 John Wiley & Sons Ltd
0.2
0.4 0.6 Antenatal content of care score
Public Private - Not for Profit
0.8
1
Private - Commercial Home
private commercial sector provided a lower level of ANC quality in all five wealth groups. Discussion We examined the content of ANC using nationally representative data in 46 low- and middle-income countries,
233
Tropical Medicine and International Health
volume 20 no 2 pp 230–239 february 2015
T. Powell-Jackson et al. Antenatal care and the private sector
Poorest
Poorer
Middle
Richer
Richest
100% 90% 80% 70% 60% 50% 40% 30% 20%
North Africa/West Asia/Europe
Home Public Private commercial Private not-for-profit
Sub-Saharan Africa
Home Public Private commercial Private not-for-profit
All countries
Home Public Private commercial Private not-for-profit
Home Public Private commercial Private not-for-profit
0%
Home Public Private commercial Private not-for-profit
10%
South/Southeast Asia
Latin America & Caribbean
Figure 3 Antenatal content of care by provider type and wealth group.
Table 1 Association of provider type with antenatal content of care score
n All countries (46) Private commercial 48 559 Private not-for-profit 6133 Home 6882 Sub-Saharan Africa (23) Private commercial 8273 Private not-for-profit 5317 Home 2306 North Africa/West Asia/Europe (8) Private commercial 9487 Private not for-profit 400 Home 314 South/South-East Asia (8) Private commercial 24 793 Private not-for-profit 314 Home 4100 Latin America & Caribbean (7) Private commercial 6006 Private not-for-profit 102 Home 162
Weighted percentage
Adjusted estimate (95% CI)
P value
0.375 0.012 0.055
0.003 ( 0.0004 to –0.007) 0.031 (0.024 to –0.039) 0.197 ( 0.207 to 0.186)
0.078
