Ownership and technical efficiency of hospitals: evidence from Ghana using data envelopment analysis

Background

In order to measure and analyse the technical efficiency of district hospitals in Ghana, the specific objectives of this study were to (a) estimate the relative technical and scale efficiency of government, mission, private and quasi-government district hospitals in Ghana in 2005; (b) estimate the magnitudes of output increases and/or input reductions that would have been required to make relatively inefficient hospitals more efficient; and (c) use Tobit regression analysis to estimate the impact of ownership on hospital efficiency.

Methods

In the first stage, we used data envelopment analysis (DEA) to estimate the efficiency of 128 hospitals comprising of 73 government hospitals, 42 mission hospitals, 7 quasi-government hospitals and 6 private hospitals. In the second stage, the estimated DEA efficiency scores are regressed against hospital ownership variable using a Tobit model. This was a retrospective study.

Results

In our DEA analysis, using the variable returns to scale model, out of 128 district hospitals, 31 (24.0%) were 100% efficient, 25 (19.5%) were very close to being efficient with efficiency scores ranging from 70% to 99.9% and 71 (56.2%) had efficiency scores below 50%. The lowest-performing hospitals had efficiency scores ranging from 21% to 30%.Quasi-government hospitals had the highest mean efficiency score (83.9%) followed by public hospitals (70.4%), mission hospitals (68.6%) and private hospitals (55.8%). However, public hospitals also got the lowest mean technical efficiency scores (27.4%), implying they have some of the most inefficient hospitals.Regarding regional performance, Northern region hospitals had the highest mean efficiency score (83.0%) and Volta Region hospitals had the lowest mean score (43.0%).From our Tobit regression, we found out that while quasi-government ownership is positively associated with hospital technical efficiency, private ownership negatively affects hospital efficiency.

Conclusions

It would be prudent for policy-makers to examine the least efficient hospitals to correct widespread inefficiency. This would include reconsidering the number of hospitals and their distribution, improving efficiency and reducing duplication by closing or scaling down hospitals with efficiency scores below a certain threshold. For private hospitals with inefficiency related to large size, there is a need to break down such hospitals into manageable sizes.

     

Technical efficiency of healthcare systems in African countries: An application based on data envelopment analysis

AimThis study aimed to measure the healthcare system efficiency of 36 African countries and to compare efficiency levels between countries.MethodData envelopment analysis (DEA) was used to evaluate efficiency. The input variables employed within the scope of DEA consisted of the proportion of total health expenditures in the gross domestic product (HE); the number of physicians (PHY), nurses (NUR) and hospital beds (BN) per 1000 people; the unemployment rate (UN); and the Gini coefficient (Gini). The study's output variables were life expectancy at birth and 1/(infant mortality rate). After DEA, the variables affecting the performance of national healthcare systems were identified using a Tobit regression model.ResultsAccording to DEA results, 21 (58.33%) of 36 African healthcare systems were found to be efficient. Among the efficient countries, Senegal was the country most referenced for inefficient countries. According to the Tobit regression analysis results, the number of nurses per 1000 people and Gini coefficient variables statistically significantly affected the inefficiency of national healthcare systems.ConclusionThis study's results suggest that in order to achieve the same health outcomes, national healthcare systems need to use public and private health resources more effectively and efficiently. By assessing the efficiency of countries’ healthcare systems and health services through international comparison, effectiveness and efficiency can be ensured within these systems.     

Assessment of efficiency levels of training and research hospitals in Turkey and the factors affecting their efficiencies

ObjectivesThis study's aim was to compare the efficiency levels of training and research hospitals in Turkey during 2014–2017 and find the factors affecting their efficiency scores.MethodsTo achieve the above objective, input oriented and variable returns to scale (VRS), data envelopment analysis (DEA), and Tobit regression model were used. The number of beds (BD), the number of intensive care beds (IBD), and the number of specialist doctors (SD) were used as the input variables of DEA while the number of polyclinic admissions (PA), number of inpatients (InP), and number of A, B, and C type surgeries as the output variables. The Tobit regression model was created by using some control variables as independent variables and transformed DEA scores as dependent variables. Average length of stay (ALS), bed turnover rate (BTR), bed occupancy rate (BOR), surgery number per specialist doctor (SNSD), polyclinic admission number per specialist doctor (PANSD), and metropolitan city status (MCS) were used as control variables.ResultsStatistically significant control variables of BTR (p < 0.001), SNSD (p = 0.001), and MCS (p = 0.015) affect the transformed DEA score while ALS, BOR, and PANSD do not (p > 0.05).ConclusionsThe research revealed that efficiency increased as BTR and SNSD increased, and that MCS also had a positive effect on efficiency. Policy makers should therefore consider the province's MCS status while allocating resources to relevant educational research hospitals for improving their efficiency. Likewise, policy makers should pay more attention to increasing BTR and SNSD compared to other variables.     

Financial incentives and productive efficiency in Finnish health centres

This paper examines the productive efficiency of Finnish health centres by applying data envelopment analysis (DEA) and econometric methods. The Tobit model was used in an attempt to find out how various economic, structural and demographic factors affect efficiency. The dependent variable of the model, the coefficient of inefficiency, was obtained by deducting the DEA efficiency score from one. According to the results, a high percentage of funding coming from central government matching grants and high taxable income per inhabitant are significant predictors of inefficiency. The results suggest that more generous resources tend to increase inefficiency since they may lessen incentives for tight cost and performance control. A high share of doctors and a low share of administrative, maintenance and support personnel promote efficiency. A low population share of the elderly and a long distance to the nearest hospital are positively associated with inefficiency.     

基于效率的医院规模经济实证分析

目的了解医院是否存在规模经济及适宜床位规模大小。方法利用浙江省500张床及以上综合性医院数据,用主、客观结合的方式和效率回归调整模型筛选评价指标,用数据包络分析方法和Tobit回归模型进行实证分析。结果所有决策单元的总体效率和纯技术效率、规模效率平均得分均大于0.93;36%的单元为总体有效,65%的单元为纯技术有效,43%的单元为规模有效;43%的单元处于规模报酬不变区间,42%的单元处于报酬递增区间,15%的单元处于报酬递减区间;全要素生产率平均呈下降;效率的影响因素主要有床位、药品收入比重、设备数量、医院等级、诊次床位比、业务收支比、每床卫技数、每床固定资产、所在地区千人医生数和经济水平等。结论多数医院存在规模经济。医院床位数在800～1 300张较为合适。提高医院经营效率可采取以下措施：合理控制床位;控制药品收入比重,注重医疗技术的提高;提高管理水平,控制固定资产过度投入;提高地区卫生资源水平和经济发展水平。     

Efficiency and quality of care in nursing homes: an Italian case study

This study investigates efficiency and quality of care in nursing homes. By means of Data Envelopment Analysis (DEA), the efficiency of 40 nursing homes that deliver their services in the north-western area of the Lombardy Region was assessed over a 3-year period (2005–2007). Lombardy is a very peculiar setting, since it is the only Region in Italy where the healthcare industry is organised as a quasi-market, in which the public authority buys health and nursing services from independent providers—establishing a reimbursement system for this purpose. The analysis is conducted by generating bootstrapped DEA efficiency scores for each nursing home (stage one), then regressing those scores on explanatory variables (stage two). Our DEA model employed two input (i.e. costs for health and nursing services and costs for residential services) and three output variables (case mix, extra nursing hours and residential charges). In the second-stage analysis, Tobit regressions and the Kruskall–Wallis tests of hypothesis to the efficiency scores were applied to define what are the factors that affect efficiency: (a) the ownership (private nursing houses outperform their public counterparts); and (b) the capability to implement strategies for labour cost and nursing costs containment, since the efficiency heavily depends upon the alignment of the costs to the public reimbursement system. Lastly, even though the public institutions are less efficient than the private ones, the results suggest that public nursing homes are moving towards their private counterparts, and thus competition is benefiting efficiency.     

Carcinogenic potential of formaldehyde in occupational settings: a critical assessment and possible impact on occupational exposure levels

Objectives To review epidemiological studies which led to a change in the classification of formaldehyde by the International Agency for Research on Cancer (IARC) in 2004 as well as studies published thereafter, with the objective to examine whether occupational exposure levels for formaldehyde should be adapted. Method Cohort and case–control studies investigating the association between occupational exposure to formaldehyde and nasopharyngeal cancer (NPC) and reporting estimates of formaldehyde exposure as well as the most recent meta-analyses, published after 1994, were reviewed. Results Evidence of an association between occupational formaldehyde exposure and NPC appears debatable. Results of the cohort studied by Hauptmann et al. (Am J Epidemiol 159(12):1117–1130, 2004) were key findings in the IARC evaluation. In this study, mortality from NPC was elevated compared with that of the US general population. However, internal comparison analysis using alternative categorization revealed that none of the relative risk for NPC was statistically significantly increased in any category of exposure (Marsh and Youk in Regul Toxicol Pharmacol 42(3):275–283, 2005) and re-analyses of the data highlighted the inappropriateness of the exposure assessment used by Hauptmann et al. (Am J Epidemiol 159(12):1117–1130, 2004) and Marsh et al. (Regul Toxicol Pharmacol 47(1):59–67, 2007). Two other cohorts (Coggon et al. in J Natl Cancer Inst 95(21):1608–1615, 2003; Pinkerton et al. in Occup Environ Med 61(3)193–200, 2004) reported no increase in NPC. Two case–control studies brought some evidence of an increased risk of NPC but the assessment of exposure levels was uncertain. Discussion Human studies fail to raise a convincing conclusion concerning the carcinogenicity of formaldehyde and are not helpful to delineate a possible dose–response relationship. Experimental data indicate that in rats, the carcinogenic activity of formaldehyde is associated with cytotoxic/proliferative mechanisms. Therefore protecting from these effects associated with formaldehyde exposure should be sufficient to protect from its potential carcinogenic effects, if any in humans. Conclusion Current occupational exposure levels to formaldehyde, set to protect against local irritation, should not be adapted.     

The impact of healthcare reform on the efficiency of public county hospitals in China

Background

The new round of Healthcare Reform in China has implemented over 3 years since 2009, and promoted greatly the development of public county hospitals. The purpose of this study is to evaluate county hospitals efficiency before and after the healthcare reform, and further assess the reform effectiveness through the comparative analysis of the efficiency.

Methods

Data envelopment analysis (DEA) was employed to calculate the efficiency of 1105 sample hospitals which were selected from 31 provinces of China, also, Tobit regression was used to regress against those main external environmental factors.

Results

Our results show that the scales and amounts of service of hospitals had increased sharply, however, the efficiency was relatively low and decreased slightly from 2008 to 2012. Thirteen (1.18%) in 2008 and six (0.54%) hospitals in 2012 were defined as technically efficient, and the average scores were 0.2916 and 0.2503. The technical efficiency average score of the post-reform was significantly less than that of the pre-reform (p?<?0.001), and the score of eastern region was highest and the western was lowest among three regions of China.

Conclusions

It suggests the reform had not well improved county hospital efficiency although hospitals have reached a fair developing scale, and the corresponding policies and measures should be put into effect for improving efficiency, especially in the level and structure of health investment, operation and supervision mechanism of county hospitals.

     

Impacts of comprehensive reform on the efficiency of Guangdong's County public hospitals in 2014–2019, China

ObjectivesWe conducted a comparative analysis on the performance of 57 county public medical systems in Guangdong Province in 2014–2019 and tried to find out how vastly the performance differed across counties and what might cause these differences. Answers to these questions could serve as the basis for policymaking in the current reform of China's rural medical systems, helping policymakers to identify the inefficient counties and giving suggestions on possible further reformation of the rural medical system.MethodsOutput-oriented SBM DEA models were employed to measure efficiency performance and a Tobit regression model was employed to explore determinants of efficiency.ResultsThe Pearl River Delta region had the most medical resources while the East region had the least. Average efficiency scores in 2014–2019 show a slight increase of around 1%, i.e., 0.798, 0.766, 0.781, 0.781, 0.779 and 0.802, respectively. 96.5% of counties experienced inefficiency problems and 36.8% of counties were worsened with declined score trends. Average efficiency scores in the Pearl River Delta region decreased mildly. Economic, government, and population factors were related to efficiency scores positively and significantly.ConclusionsMost county public hospitals experienced inefficiency problems and substantial regional disparities existed in Guangdong's county-level medical efficiency, which should be noticed by policymakers. Issues in the reform like insufficient revenues and government subsidies, higher surgical expenses, inadequate technology, and unbalanced regional policies may have hindered medical efficiency, which should be addressed in future reforms.     

Investigating technical efficiency and its determinants: The Case of New Zealand District Health Boards

This study investigates the technical efficiency of New Zealand's District Health Boards (DHBs) in providing hospital services, as well as the effect of certain environmental factors on efficiency. This study is the first to use quarterly data on New Zealand DHBs from 2011 to 2017 and apply the two-stage double-bootstrap methodology of Simar and Wilson. The bias-corrected technical efficiency estimates show that on average, DHBs in the areas with high socioeconomic deprivation operate with low technical efficiency. Furthermore, DHB providing secondary hospital services are less efficient than tertiary DHBs. The result from truncated regression indicates that a higher proportion of surgical, elderly, and acute inpatients is associated with increasing levels of technical efficiency. In contrast, the high average length of hospital stay negatively impacts technical efficiency levels. The findings of this study urge policymakers to adopt policies to address the shortages of healthcare staff, barriers to primary healthcare, lack of investment in hospital capacity, and technology to enhance healthcare sector's long-run technical efficiency. In addition, the existing DHB funding formula needs to be revisited as this tends to include perverse incentives for secondary DHBs where patients are kept longer in hospitals, leading to a higher average length of stays in hospitals and is associated with increasing levels of inefficiency.     

Handling Missing Data in Randomized Experiments with Noncompliance

Treatment noncompliance and missing outcomes at posttreatment assessments are common problems in field experiments in naturalistic settings. Although the two complications often occur simultaneously, statistical methods that address both complications have not been routinely considered in data analysis practice in the prevention research field. This paper shows that identification and estimation of causal treatment effects considering both noncompliance and missing outcomes can be relatively easily conducted under various missing data assumptions. We review a few assumptions on missing data in the presence of noncompliance, including the latent ignorability proposed by Frangakis and Rubin (Biometrika 86:365–379, 1999), and show how these assumptions can be used in the parametric complier average causal effect (CACE) estimation framework. As an easy way of sensitivity analysis, we propose the use of alternative missing data assumptions, which will provide a range of causal effect estimates. In this way, we are less likely to settle with a possibly biased causal effect estimate based on a single assumption. We demonstrate how alternative missing data assumptions affect identification of causal effects, focusing on the CACE. The data from the Johns Hopkins School Intervention Study (Ialongo et al., Am J Community Psychol 27:599–642, 1999) will be used as an example.     

Monitoring Metals near a Hazardous Waste Incinerator. Temporal Trend in Soils and Herbage

In recent years, incineration has been demonstrated to be a commercially available technology for hazardous waste (HW) disposal (Richter and Johnke, 2004). However, because of the potential adverse effects of toxic emissions, waste incinerators are still an important cause for concern for the public. In spite of that, compliance with current EU emissions has vastly reduced the probability of adverse health effects (Glorennec et al., 2005). With respect to metals, a number of studies have shown that these elements are emitted by industrial, medical and municipal waste incinerators (Schumacher et al., 1997; Rimmer et al., 2006). Filter ash is an especially problematic residue because it contains high metal concentrations (Lisk et al., 1989). After combustion in modern HW incinerators (HWIs), metals contained in HW are mainly collected in bottom and fly ash, with only small quantity of metals being discharged from the stack as particulate matter or vapor (Jung et al., 2004). However, the atmospheric emission of these elements is a matter of concern.     

Living in the hands of God. English Sunni e-fatwas on (non-)voluntary euthanasia and assisted suicide

Ever since the start of the twentieth century, a growing interest and importance of studying fatwas can be noted, with a focus on Arabic printed fatwas (Wokoeck 2009). The scholarly study of end-of-life ethics in these fatwas is a very recent feature, taking a first start in the 1980s (Anees 1984; Rispler-Chaim 1993). Since the past two decades, we have witnessed the emergence of a multitude of English fatwas that can easily be consulted through the Internet (‘e-fatwas’), providing Muslims worldwide with a form of Islamic normative guidance on a huge variety of topics. Although English online fatwas do provide guidance for Muslims and Muslim minorities worldwide on a myriad of topics including end-of-life issues, they have hardly been studied. This study analyses Islamic views on (non-)voluntary euthanasia and assisted suicide as expressed in English Sunni fatwas published on independent—i.e. not created by established organisations—Islamic websites. We use Tyan’s definition of a fatwa to distinguish between fatwas and other types of texts offering Islamic guidance through the Internet. The study of e-fatwas is framed in the context of Bunt’s typology of Cyber Islamic Environments (Bunt 2009) and in the framework of Roy’s view on the virtual umma (Roy 2002). ‘(Non-)voluntary euthanasia and assisted suicide’ are defined using Broeckaert’s conceptual framework on treatment decisions at the end of life (Broeckaert 2008). We analysed 32 English Sunni e-fatwas. All of the e-fatwas discussed here firmly speak out against every form of active termination of life. They often bear the same structure, basing themselves solely on Quranic verses and prophetic traditions, leaving aside classical jurisprudential discussions on the subject. In this respect they share the characteristics central in Roy’s typology of the fatwa in the virtual umma. On the level of content, they are in line with the international literature on Islamic end-of-life ethics. English Sunni e-fatwas make up an influential and therefore important developing body of Islamic orthodox normative authority on end-of-life ethics that is still open for further research.     

Answer to comments by A. Lerchl on “Radiofrequency electromagnetic fields (UMTS, 1,950 MHz) induce genotoxic effects in vitro in human fibroblasts but not in lymphocytes” published by C. Schwarz et al. 2008

Genotoxic effects induced in vitro by the third generation mobile communication standard UMTS have recently been described by Schwarz et al. (Int Arch Occup Environ Health 81:755–767, 2008). These findings which may have considerable significance for environmental health have been commented upon by Lerchl (Int Arch Occup Environ Health in press, 2008) (this issue). These comments which are invalid in part have to be set right. Although some of his minor points are correct the objected inconsistencies are largely based on the author′s incomplete and superficial consideration of published data in the field. Moreover, the statistical points being made cannot cast doubts on the validity of the experimental data reported by Schwarz et al. and may not change the principal conclusion of in vitro genotoxic action of UMTS signals.     

Male Sexual Orientation in Independent Samoa: Evidence for Fraternal Birth Order and Maternal Fecundity Effects

In Western cultures, male androphiles tend to have greater numbers of older brothers than male gynephiles (i.e., the fraternal birth order effect). In the non-Western nation of Independent Samoa, androphilic males (known locally as fa’afafine) have been shown to have greater numbers of older brothers, older sisters, and younger brothers (Vasey & VanderLaan, 2007). It is unclear, however, whether the observed older brother effect, in the context of the additional sibling category effects, represented a genuine fraternal birth order effect or was simply associated with elevated maternal fecundity. To differentiate between these two possibilities, this study employed a larger, independent replication sample of fa’afafine and gynephilic males from Independent Samoa. Fa’afafine had greater numbers of older brothers and sisters. The replication sample and the sample from Vasey and VanderLaan were then combined, facilitating a comparison that showed the older brother effect was significantly greater in magnitude than the older sister effect. These results suggest that fraternal birth order and maternal fecundity effects both exist in Samoa. The existence of these effects cross-culturally is discussed in the context of biological theories for the development of male androphilia.     

Social and health determinants of the efficiency of cotton farmers in Northern Côte d'Ivoire

This article assesses the role of malaria and some social determinants on the agricultural development and more precisely on efficiency in the context of cotton crop in the Korhogo region in the North of C?te d'Ivoire. Data envelopment analyses (DEA) was first applied for the purpose of calculating relative efficiencies in production. A Tobit regression model was then used to explain the variation in the DEA scores and check the hypotheses that the efficiency deviations between farmers can be explained by the disparity of malaria morbidity rate among the farmers and their family, by social cohesiveness and cultural behaviour. Field data were collected by the authors between March 1997 and February 1998 on 700 rural households living in three rice production systems differently exposed to the malaria risk. Two malaria indicators were used for the active (11-55 years old) family members of the farm: Plasmodium falciparum infection rate and high parasite density infection rate. The DEA model was applied on the sub-sample of cotton growers (about one third of the households of the full sample). Results of the different DEA and Tobit models (depending of the production process hypothesis) show that high parasite density infection has a direct and indirect negative effect on efficiency in the cotton crop. They also show that more cotton growers in the village improve efficiency, although villages where cotton is growing more widespread have weaker social cohesion.     

Heavy Metals in Wet Method Coffee Processing Wastewater in Soconusco,Chiapas, Mexico

One of the driving forces of the economy in southeast Mexico is agriculture. In Soconusco, Chiapas, coffee is one of the main agricultural products and is traded on the international market. Coffee grown in this region is processed using the wet method in order to be commercialized as green coffee. In the beneficio (coffee processing plant) water is an essential resource which is required in great quantities (Matuk et al., 1997; Sokolov, 2002) as it is used to separate good coffee berries from defective ones, as a method of transporting the coffee berries to the processing machinery, in the elimination of the berry husk from the coffee grains (pulping) and finally in the post-fermentation washing process. This process gives rise to one of the smoothest, high-quality coffees available (Zuluaga, 1989; Herrera, 2002). Currently, many producers in Soconusco are opting for ecological coffee production, which has, among its many criteria, human health and environmental protection (Pohlan, 2005). Furthermore, increasing concern during the past few years regarding the production of food that is free from contaminants such as heavy metals, and recent environmental policies in relation to aquatic ecosystem protection, have given rise to questions concerning the quality of water used in coffee processing, as well as pollutants produced by this agroindustry. Water used in the coffee processing plants originates from the main regional rivers whose hydrological basins stretch from the Sierra Madre mountain range down to the coastal plain. As well as providing water, these rivers also receive the wastewater produced during coffee processing (Sokolov, 2002).     

某院临床科室经营绩效评价分析

目的 对某公立医院各临床科室的运营效率进行评价,为医院和科室管理提供决策参考。方法 根据14个临床科室2015年—2016年的面板数据,运用数据包络分析（DEA）中的CCR和BCC模型进行生产效率绩效评价,并运用Tobit回归分析,探索影响经营绩效的关键因素。结果 14个临床科室综合效率值2016年比2015年提高0.074,表明医院对临床科室的投入与产出资源使用效率2016年比2015年有所提高。运用Tobit模型探索分析影响科室经营绩效的关键因素,发现病床使用率、纯医疗收入占比、医保患者占比与综合效率     

Discriminatory Mass De-housing and Low-Weight Births: Scales of Geography,Time, and Level

Struening et al.1 demonstrated a widening disparity of low birthweight (LOB) rates among New York City health areas from 1980–1986, clearly a dynamic process. In contrast, the New York City Department of Health reported static citywide LOB rate in 1988–2008.2 Struening et al.1 is extended here at the health district level with mapping and regression analyses. Additionally, birthweight data are reported for babies born in 1998–2001 to a group of African-American and Dominican women in Upper Manhattan. The data reported in this paper indicate that both fetal programming of the mother herself (life course model) and stress during or shortly before pregnancy may play a role in LOB. Current stress may arise from past events. Intergenerational effects, thus, could arise from stresses on the grandmother and their residual impacts on the mother as well as new stresses on the mother as an adult. The average weight of babies born to the Upper Manhattan mothers who were born in 1970–1974 was 3,466 g, with 1.6% below 2,500 g; that of babies of mothers born in 1975–1979, 3,320 g, with 6% below 2,500 g. The latter group was born during the 1975–1979 housing destruction. Intergenerational impacts of that event may be reflected in this elevated rate of LOB. Health district maps of LOB incidence ranges show improvement from 1990–2000 and then deterioration in 2005 and 2008. Bivariate regressions of socioeconomic (SE) factors and LOB incidence showed many strong associations in 1990; but by 2000, the number and strength of these associations declined. In 1990, 2000, and 2008, black segregation was the SE factor most strongly associated with LOB. Black segregation and murder rate explained about 85% of the pattern of 1990 LOB. Regressing the 1970–1980 percent population change against the SE factors showed effects even in 2000. The 1990 murder rate and 1989 percentage of public assistance explained over half the 2008 LOB incidence pattern. The housing destruction of the 1970s continued to influence LOB incidence indirectly in 2008. The ability of community and individual to cope with current stressors may hinge on resilience status, which is shaped by past events and circumstances. The present interacts with the past in many ways. Serial displacement exemplifies this interaction of immense importance to public health.     

Assessing the efficiency of rural health centres in Burkina Faso: an application of Data Envelopment Analysis

Background  Effective health care provision benefits from the support of measurement techniques. Contrary to the situation in industrialised countries efficiency analyses in the health care sector in Africa are a very recent phenomenon. Hardly any of the existing studies was conducted at the level of primary care. Aim  The purpose of this study was twofold: (1) to evaluate the relative efficiency of health centres in rural Burkina Faso and (2) to investigate reasons for inefficient performance. Methods  Data Envelopment Analysis (DEA) was applied. To account for the situation in that country, the output-oriented approach was used in connection with different returns to scale assumptions. To identify the spatial effect of the catchment area on efficiency the Tobit model was applied. Results  According to constant returns to scale, 14 health centres were relatively efficient. The DEA projections suggest that the inefficient units were too big to be efficient. Tobit regression showed that the relatively efficient health centres are located close to villages in their catchment area. Conclusions  For ethical reasons it is not appropriate to try to improve the efficiency of health centres by closing some of them. Their efficiency can be improved and lives can be saved if access to health centres is enhanced. Funding: This study was supported by a research grant of the German Research Foundation (Deutsche Forschungsgemeinschaft).