Objective To document the patterns of health service utilization and health payments at public and private facilities across countries.
Method We used data from the World Health Surveys from 39 low- and low-middle income countries to examine differences between public and private sectors. Utilization of outpatient and inpatient services, out-of-pocket payments (OOP) at public and private facilities, and transportation costs were compared.
Results Utilization and payments to public and private sectors differ widely. Public facilities dominated in most countries for both outpatient and inpatient services. But, whereas use of private facilities is more common among the rich, poor people also use them, to a considerable extent and in almost all the countries in the study. The majority of OOP were incurred at public providers for inpatient services. On average, this was not the case for outpatient services. Medicines accounted for the largest share of OOP for all services except inpatient services at private facilities, where consultation fees did. Transportation costs were considerable. Price competition is certainly not the only factor that guides choice of provider.
Conclusions The results support continued efforts by the governments to engage strategically with the private sector. However, they also highlight the importance of not generalizing conditions across countries. Governments may need to reconsider simplistic user-fee abolition strategies at public providers if they simply focus on consultation fees. Policies to make health services more accessible need to consider a comprehensive benefit package that includes a wider scope of costs related to care such as expenditures on medicines and transportation.
Objectif: Documenter les schémas d’utilisation des services de santé et les paiements pour la santé dans les établissements publics et privés des différents pays.
Méthode: Nous avons utilisé les données de la Surveillance Mondiale de la Santé de 39 pays à revenus faibles et intermédiaires pour examiner les différences entre les secteurs publics et privés. L’utilisation des services ambulatoires et d’hospitalisation, les paiements directs de sa propre poche dans les établissements publics et privés et les coûts de transport ont été comparés.
Résultats: L’utilisation et les paiements dans les secteurs publics et privés diffèrent largement. Les établissements publics dominaient dans la plupart des pays à la fois pour les services ambulatoires et d’hospitalisation. Mais, alors que l’utilisation des services privés est plus fréquente chez les riches, les pauvres les utilisent aussi dans une mesure considérable et dans presque tous les pays de l’étude. La majorité des paiements directs de sa propre poche ont été engagés pour des prestataires de services publics d’hospitalisation. En moyenne, cela n’était pas le cas pour les services ambulatoires. Les médicaments représentaient la plus grande part de ces paiements pour tous les services sauf pour les services aux patients hospitalisés dans les établissements privés où ces paiements constituaient les frais de consultation. Les coûts de transport étaient considérables. La concurrence des prix n’est certainement pas le seul facteur orientant vers le choix du prestataire.
Conclusions: Les résultats soutiennent la poursuite des efforts par les gouvernements à s’engager de façon stratégique avec le secteur privé. Cependant, ils soulignent aussi l’importance de ne pas généraliser les conditions dans différents pays. Les gouvernements devraient avoir besoin de reconsidérer les stratégies simplistes de suppression des frais d’utilisation pour les prestataires publics s’ils se concentrent simplement sur les frais de consultation. Les politiques visant à rendre les services de santé plus accessibles nécessitent d’envisager un ensemble complet de prestations comprenant un éventail plus large des coûts liés aux soins telles que les dépenses pour les médicaments et le transport.
Objetivo: Documentar los patrones de uso de los centros sanitarios y los pagos sanitarios en centros públicos y privados de diferentes países.
Método: Hemos utilizado datos de las Encuestas Mundiales de Salud de 39 países con renta baja y media, para examinar las diferencias entre los sectores público y privado. Se comparó la utilización de los servicios de consulta externa y hospitalización, los pagos de bolsillo en centros públicos y privados y los costes de transporte.
Resultados: El uso y los pagos en los sectores público y privado diferían ampliamente. Los centros sanitarios públicos dominaban en la mayoría de los países tanto en servicios externos como en ingresos hospitalarios. Sin embargo, mientras que el uso de centros privados era más común entre los ricos, los pobres también los utilizaban bastante y en la mayoría de países en los que se realizó el estudio. La mayoría de los pagos de bolsillo se realizaron a los proveedores públicos y para servicios de hospitalización. En promedio, este no fue el caso para los servicios de consultas externas. Los medicamentos eran los responsables de la mayor parte de los gastos de bolsillo en todos los servicios, excepto en los de hospitalización en centros privados. Los costes de transporte eran considerables. La competencia de precios no es, sin duda, el único factor que se tiene en cuenta a la hora de escoger proveedor.
Conclusiones: Los resultados apoyan los esfuerzos continuos de los gobiernos para crear alianzas estratégicas con el sector privado. Sin embargo, también muestran la importancia de no generalizar las condiciones existentes en diferentes países. Los gobiernos podrían necesitar reconsiderar las estrategias simplistas de abolición de las tarifas para usuarios en proveedores públicos si se centran solamente en los tarifas de consulta. Las políticas para hacer que los servicios sanitarios sean más asequibles deberían considerar un paquete integral de beneficios que incluya un gama de costes más amplia y relacionada con los cuidados recibidos, tales como el gasto en medicamentos y el transporte.
The role of private health providers has sparked controversial debates in low- and middle-income countries. For some, increasing private provision could lead to gains in efficiency, responsiveness, quality and consumer choice (Preker et al. 2000; Bhattacharyya et al. 2010). Indeed, the private sector has complemented or taken on health service delivery functions with positive outcomes in some contexts (Loevinsohn & Harding 2005; Patouillard et al. 2007; Liu et al. 2008). Others have argued that relying on public provision for health care services is the best guarantee for equitable access and for better health outcomes for the whole population (Oxfam 2009; Rannan-Eliya & Sikurajapathy 2009). Globally the evidence on the relative advantages of the private sector in health service provision is largely inconclusive (Mills 1997; Patouillard et al. 2007; Hollingsworth 2008). However, there is one common message from almost all previous research - governments cannot afford to ignore non-state actors (Bustreo et al. 2003; Preker 2007; Hanson et al. 2008; Meessen et al. 2011). This recommendation makes sense but is hard to put in practice as the information on private sector utilization in low-income settings is often incomplete or lacking, making evidence-based policy choices difficult. Indeed, what can we actually say about who is using private facilities, for what reasons and for what prices in low-income settings from an international perspective?
Naturally, an underlying challenge is understanding and defining what is meant by the private sector. As an umbrella group, the private or non-state sector covers non-state non-profit organizations and private for-profit facilities. These in turn can range, for example, from an informal provider operating in a slum area of a large city or a high-end clinic providing sophisticated care for the elite in the rich neighbourhoods of the same city, to a church-run non-profit health centre in a rural village where public services may not even exist. Facilities may in fact be more different than similar. Conceptually, common frameworks to examine all facilities – state and non-state – based on criteria such as their objectives, principles and operation styles to name a few may be superior. However, in practice even within countries, data does not generally allow for examining these subtleties.
Looking beyond the classification issues, one starting point to gather information on the importance of the private sector may be to examine public health expenditure data from National Health Accounts, which employ a consistent method across countries. According to this data general government expenditure on health (GGHE) as a share of total health expenditure (THE) was <50% in half of low-income countries in 2007. Without any doubt, current levels of public funding are simply not sufficient to provide adequate health services in many countries (World Health Organization 2010). However, expenditure data are far from sufficient for understanding the importance of private facilities as it is not possible to dichotomize between public funding and service provision and private funding and service provision. Public funding can be channelled through private facilities and vice versa. Therefore, expenditure data can only provide a very rough picture of public sector capacities.
National health services utilization and expenditure surveys in different countries could also be the useful sources of information on the private sector. The Living Standards and Measurement Study surveys provide one such instrument and are available for a number of countries. However, even surveys such as these do not use a consistent method and as a result are not an ideal instrument for thorough international analysis (Heijink et al. 2011). The issue of comparability is even more accentuated for small-scale surveys. Finally, region- or illness-specific studies exist in some contexts, but these are not reflective of health services as a whole (Amin et al. 2003; Sudha et al. 2003; Rutebemberwa et al. 2009).
Our study strives to provide more information on the private sector in health services provision from an international perspective. We used a unique dataset, the World Health Survey, to explore the patterns of health service utilization and related payments in public and private facilities in 39 low- and low-middle income countries. We do not go into a normative debate on the appropriate role and size of the private sector provision of health services; how the private sector should be integrated to serve national health goals is undoubtedly context-specific. Our objective is simply to provide further evidence on the utilization of public and private sectors as they exist in different countries.
Methodology and data
The World Health Survey of 2003 from 39 low- and low-middle income countries was used to analyse the utilization and spending at different types of facilities. The uniform nature of the survey questions and method provide a useful backdrop for this analysis. We used a descriptive approach in this analysis, which focuses on cross-country patterns and the differences between rich and poor. The survey sample sizes, numbers of observations for relevant sections as well as the country abbreviations used in the figures are provided in Table 1.
Table 1. Country abbreviations used in figures
Survey section sample size
Outpatient visits in the past 4 weeks in the sample
Inpatient stays in the past year in the sample
Bosnia and Herzegovina
Lao People’s Democratic Republic
Service utilization, out-of-pocket payment and transportation costs are from the individual section of the survey. We included outpatient visits that were reported to have occurred in the 30 days preceding the survey and inpatient stays within the 1 year preceding the survey.
The original survey classifies health care providers as: operated by the government; privately operated; NGO; and other. An appendix presents the utilization data using the survey’s original classification. With the exception of a handful of countries, facilities classified as ‘Privately operated’ dominated utilization outside of facilities that are ‘Operated by the government’. Given the potential misclassification of non-state facilities into the three possible groups in the survey, the comparability of these three subgroups across countries as well the skewed nature of utilization in favour of ‘Privately operated’ facilities as opposed to ‘NGO’ or ‘Other’ facilities, we chose to aggregate all the different non-public providers under the single denomination of ‘private providers’. In essence, these providers together can be thought of as different from facilities operated by the government. In addition, we substitute the term ‘public’ providers for facilities ‘Operated by the government’ for ease of reading.
We present utilization rates, out-of-pocket payments (OOP) and per visit charges for outpatient consultations and inpatient stays for public and private providers. Out-of-pocket payments were further separated into consultation/doctors fees, medicines, tests and others costs. Additionally, under-the-table payments and informal fees should be captured, although they cannot be quantified from the survey questions. We also analyse the transportation costs to better understand the distance and travel cost factors in choosing facilities.
For outpatient services, more than half of the utilization was at public facilities in 27 of 39 countries in the study (Figure 1). For some countries, more than 80% of services were provided at public facilities. For inpatient services, public facilities are even more dominant; their share exceeded the share of private facilities in all countries except Pakistan and India. In Brazil, Nepal, the Philippines and the Dominican Republic between 50% and 60% of hospitalizations were at public facilities. These results can be validated with similar country-specific findings in previous literature (Dilip 2010; Sengupta & Nundy 2005; Barnes et al. 2010).
In general, we observed a relationship between the dominance of public facilities in outpatient and inpatient services. Countries with high utilization of public facilities for outpatient services show similar patterns for inpatient services. However, in some countries, private facilities account for a significant share of outpatient visits but not inpatient stays.
A simple hypothesis would be that a larger share of government expenditure in THE would be associated with a higher use of public facilities. Figure 2 shows the utilization of public facilities for outpatient services as compared to GGHE as a share of THE in each country. Similarly, Figure 3 shows utilization of public facilities for inpatient services as compared to GGHE as a share of THE.
There is a general trend of higher utilization of public facilities in countries with larger shares of government expenditure in THE. However, there is much variation across countries. This could be explained by many factors, such as the difference in services available and the efficiency of public facilities. It may also reflect the different financial arrangements between fund holders and service providers. Governments can pay for services directly provided through public facilities, but they can also purchase services from private providers through social health insurance reimbursements that cover also the private sector or through contracting. Indeed, previous research shows that there is not always a clear linear relationship between the size of public funding and the share of public service provision (Hanson & Berman 1998; Gauri et al. 2004).
Figure 4 shows the utilization patterns for outpatient services for the poorest quintile, quintile 1, in a country as compared to the richest quintile, quintile 5. In most countries, among those who reported utilization of outpatient services, the poorest quintile was more likely to use public providers as compared to the richest quintile. For example in Brazil, government operated outpatient services were much more likely to be used by those in quintile 1 relative to those in quintile 5. This could be expected because the Brazilian health care model relies on a dual system where the richer part of the population uses mainly privately provided services, while the Sistema Único de Saúde (SUS) provides public services through pro-poor targeted public funding (Elias & Cohn 2003). However, in some countries such as Swaziland, the pattern is strikingly the opposite – with the richest quintile using public facilities for outpatient services more than the poorest quintile. In other countries, such as Ghana or Republic of Congo, reliance on public facilities seems to be similar across the quintiles.
Figure 5 presents the utilization patterns for inpatient stays for quintile 1 and quintile 5. Similar patterns as those for outpatient services emerge. However, the richest quintile is rarely more likely to use public facilities than the poorest quintile. This figure also confirms the expectation that poorest groups tend to rely almost exclusively on government facilities for inpatient care – with their use exceeding 80% in 25 of the 39 countries presented. However, primarily due to the small sample of people reporting hospitalization, the difference is only statistically significant in 12 countries.
Per visit charges in public and private facilities
Figure 6 presents the ratio of average per visit charges at public facilities to private facilities. The charges include all components of OOP – consultation fees, medicines, tests as well as others costs. The y-axis is the logarithmic ratio of charges. Positive values on the axis reflect charges at private facilities being greater than charges at public facilities and negative values reflect the opposite. The x-axis here (which is the same as the y-axis of Figure 2) plots the share of utilization of outpatient services in the public sector. As would be expected, in general, charges at private facilities are higher than at public facilities. This holds for outpatient services in 27 of the 39 countries presented. Exceptions include countries such as India, Pakistan and Bangladesh, where private provision of services is large as well as Georgia, Lao People’s Democratic Republic, China and Mauritania, where public provision is dominant. However, there are no clear patterns of increasing utilization at public facilities as the price differential increases because there is much variation in these results.
Similarly to Figures 6 and 7 plots the relative charges per visit for public and private providers against the share of utilization in the former for inpatient services. Once again, we observe that the charges per visit are higher at private facilities in 30 of 39 countries in this study. But there are variations in the share of outpatient and inpatient use in public facilities given the same level of difference in per visit charges.
Figure 8 compares the ratios of per visit charges for private and public facilities for outpatient and inpatient services. In most countries with considerable price differentials between private and public providers for outpatient services, there are also higher prices differentials for inpatient services. However, there are some exceptions, such as Bangladesh, where outpatient services are cheaper in the private sector, but inpatient visits are cheaper in the public sector. The opposite trend seems to hold for other countries such as Ethiopia and Viet Nam.
Total out-of-pocket payments
Figure 9 shows the distribution of total outpatient OOP reported in the survey, grouped by component as well as by type of facility. On average, 45% of total OOP (which include consultation/doctor’s fees, medicines, tests and other expenses) for outpatient services are paid to public providers. However, the range across countries is quite wide, with outpatient OOP at public facilities representing less than 10% in countries such as Guatemala and Malawi, and over 95% in China.
Expenditure on medicine accounts for the largest component of OOP in both public and private facilities. On average, medicines represented over 57% of outpatient OOP at public facilities and over 45% of outpatient OOP at private facilities. Exceptions were countries such as Swaziland and South Africa, where consultation fees comprised the largest component of outpatient OOP. Overall, consultation fees were the second largest component, on average, 22% of OOP at public facilities and 40% of OOP at private facilities. Consultation fees in public facilities accounted for less than 10% of total outpatient OOP on average. China and Georgia are the only countries where consultation fees in public facilities exceeded 20% of total outpatient OOP.
With regard to the differences in the distribution of outpatient OOP across different quintiles, Figure 10 shows that quintile 5 spends more at the private facilities than quintile 1. Indeed, on average, only 37% of OOP payments by quintile 5 are to public facilities, whereas 57% of payments by quintile 1 are. However, the range is once again very wide with no OOP reported at public facilities by quintile 1 in countries like Swaziland and Mali, but as much as 100% of OOP in others. Similarly, the range for quintile 5 is from under 2.5% in Ecuador to over 95% in other countries. With respect to the components of OOP, quintile 5 seems to spend more on consultation fees than quintile 1, although the patterns are generally similar to the overall level.
On average, almost 60% of inpatient OOP was at public facilities (Figure 11). It exceeded 80% in 12 countries. Overall, it ranged from just over 10% in Guatemala to over 98% in China. Consultation fees represented the biggest share of inpatient OOP at private facilities, where they accounted for 43%, while medicines represented 31% of inpatient OOP on average. However, in some countries such as Nepal, medicines accounted for a bigger share of inpatient OOP at private facilities than consultation fees. Among public facilities, expenditure on medicines was still the most dominant component of inpatient OOP, on average of 48%. The share of OOP for consultation fees at public facilities was slightly higher for inpatient services at 26% than for outpatient services. Notably in countries such as South Africa and Swaziland, consultation fees were the biggest component of inpatient OOP at public facilities.
Differences in patterns of inpatient OOP across different quintiles also emerge (Figure 12). As with outpatient OOP, quintile 5 spends less at public facilities than quintile 1 for inpatient OOP. On average, quintile 5 spends 54% of inpatient OOP at public facilities compared to 72% for quintile 1. However, once again a fairly wide range is observed – for quintile 1 from under 10% in Mali to 100% in Bosnia and Herzegovina, and for quintile 5 from 6% in Brazil to almost a 100% in China. With respect components of inpatient OOP by type of services, a similar picture to that of outpatient OOP emerges and once again, quintile 5 seems to spend more on treatment than quintile 1.
Transportation costs were not included in per visit treatment charges or OOP. However, the average ratio of transportation costs to per visit treatment charges shows that the transportation costs can be considerable. Our results show that on average, transportation costs were 12% of per visit treatment charges for outpatient services. In 22 of the 39 countries, they are more than 10% of outpatient treatment charges. Transportation costs for hospitalization were 17% of inpatient treatment charges on average. They were less than 15% for 28 of the countries presented.
Figure 13 shows the ratio of transportation costs to per visit treatment charges for outpatient and inpatient services, respectively. The ratios of transportation costs to treatment charges are smaller for outpatient services at private facilities than at public facilities in most countries but there are exceptions such as Swaziland. On average for outpatient services, transportation costs to private facilities were 8% of treatment charges, and the ratio exceeded 10% in only eight countries. For inpatient services, they were 12% on average but as low as around 1% in the eastern European countries shown here. In seven countries, they exceeded 15% of inpatient treatment charges.
Before interpreting the results, some limitations of the study need to be discussed. People’s understanding of whether a facility is a government, private, NGO or other provider may differ. We think grouping the subcategories of non-state providers into one single category eliminates some potential classification errors. However, even with just having ‘public’ and ‘private’ providers, the possibility of misclassification arising from personal interpretation as well as differences in health systems across different countries should be kept in mind. Furthermore, whereas the objective of the study is to compare government and non-government providers, there is diversity within this classification of providers, particularly for private providers with respect to costs, quality and target clienteles. Indeed, we recognize that the denomination of ‘public’vs.‘private’ may mask many real-life complexities, and this should be kept in mind when interpreting the findings.
We should also be cautious in comparing the results for per visit charges between public and private facilities without information on the quality and the intensity of the services. Additionally, types of OOP may also be misclassified to a certain degree. This issue may be particularly pertinent with regard to under-the-table payments, which are likely to occur more frequently in public facilities. No explicit information on these types of payments was elicited and as such, it is not clear to what degree they have been captured in this data. The general data quality may also vary from country to country. Certain countries may appear to be outliers or have questionable data in some of the results presented. However, we consciously shy away from making country-specific conclusions in this last section not only because this is intended as a cross-country analysis, but also because of a lack of thorough validation data for all countries. Some countries such as Swaziland also have few observations for some of the results and thus we should be particularly cautious in drawing conclusions based on these data.
Some cross-country findings are worth highlighting here. Firstly, the study found that the individuals in the richest quintile are more likely to use private facilities than the lowest income group. However, the use of the private sector is not limited to the elite. Even in the poorest quintile, private facilities are used for more than 20% of outpatient visits in the majority of countries. This result is in line with previous studies that have also noted a considerable use of private health services among the poor (Bhatia & Cleland 2001; Prata et al. 2005; Harding 2009; World Bank 2011).
Secondly, we find that the price competition may not be the main criterion for patients in many countries. Indeed, in many countries where the price at private facilities is well above that of public facilities, use of public sector providers is comparatively low. Patients may value services provided by private facilities more in these settings. However, the opposite pattern, where public facilities are preferred despite higher prices, is also seen in some countries. Other factors such as perceived quality, responsiveness and geographical access may play an important role in determining what kind of facility is used. Indeed, our study also found that in most countries included in this study, transportation costs represented more than 10% of what the individuals were paying for the treatment.
The study also found that the consultation fees in public facilities do not necessarily represent the largest component of OOP. Consultation fees for inpatient and outpatient visits in public facilities account for, on average, only 10% of total OOP. In most countries, the level is well below 15%. Importantly, the majority of OOP at public facilities is for the purchase of medicines.
These findings suggest that blanket policies of abolishing user fees, which often only include consultation fees in public facilities, do not solve the whole problem. This will also apply to contexts where in theory user-fee abolition covers costs of consultation and medicines at public facilities, but where in reality provision in the public sector faces challenges, forcing individuals to buy their medicines from private providers. As this study shows, these approaches may ignore the bulk of payments that are likely to pose immense barriers for equitable access. Overall, government policies should focus on reducing the broader set of OOP rather than exclusively concentrating on any particular component. Countries should also consider lowering transportation costs through long-term strategies, such as increasing the number of facilities and allocating resources based on population need, as well as other policies such as vouchers or other reimbursement schemes to cover these costs (Noirhomme et al. 2007; National Health Service England 2010). In addition to addressing OOP, policies to improve the perceived value of services may be needed. Our data showed that people used private facilities despite higher prices in many settings. Increasing the attractiveness of services requires considering factors such as quality and responsiveness to fully translate financial accessibility to better health.
More importantly perhaps, national health financing systems need to engage with non-state actors in a way that enables greater and better access for the population. Governments, as health sector stewards, should ensure that the quality standards are met and routinely gather information from both public and non-state facilities. In addition, governments may need to implement regulatory measures or establish strategic contractual agreements with the non-state sector that are guided by the national health and health systems objectives. These policies also need to implicitly address patient needs with respect to responsiveness, quality of care as well as financial risk protection. Indeed, there are many pioneering programs engaging with the private sector in various settings (Balique et al. 2001; Mahmud et al. 2002; Danel & La Forgia 2005; Mavalankar et al. 2009; Meessen et al. 2011). Building on existing experience and the lessons learnt, policymakers in every country could take steps to better engage with the private sector to achieve national goals. But we should also be conscious that facilities, both private and public, may provide drastically different care in different settings and for different people (Needham et al. 2001; Harding & Preker 2003; Rathore et al. 2003; World Health Organization 2008; Victora et al. 2010). Indeed, our study found much variation, in both utilization and the cost of services. As such, countries should not adopt dogmatic views on the role of the non-state sector in health care. This is a complex reality, and policies should reflect rather than ignore context-specific subtleties.