Multilevel Zero-One Inflated Beta Regression Model for the Analysis of the Relationship between Exogenous Health Variables and Technical Efficiency in the Spanish National Health System Hospitals.

BACKGROUND: This article proposes a methodological innovation in health economics for the second stage analysis of technical efficiency in hospitals. It investigates the relationship between the installed capacity in regions and hospitals and their ownership structure. METHODS: A multilevel zero-one inflated beta regression model is employed to model pure technical efficiency more adequately than other models frequently used in econometrics. RESULTS: Compared to publicly managed hospitals, the mean efficiency index of hospitals with public-private partnership (PPP) formulas was 4.27-fold. This figure was 1.90-fold for private hospitals. Concerning the efficiency frontier, the odds ratio (OR) of PPP models vs. public hospitals was 42.06. The OR of private hospitals vs. public hospitals was 8.17. A one standard deviation increase in the percentage of beds in intensive care units increases the odds of being situated on the efficiency frontier by 50%. CONCLUSIONS: The proportion of hospital beds in intensive care units relates to a higher chance of being on the efficiency frontier. Hospital ownership structure is related to the mean efficiency index of Spanish National Health Service hospitals, as well as the odds of being situated on the efficiency frontier.

Multilevel Analysis of the Relationship between Ownership Structure and Technical Efficiency Frontier in the Spanish National Health System Hospitals.

Objective: To evaluate the relationship between the ownership structure of hospitals and the possibility of their being positioned on the frontier of technical efficiency in the economic crisis period 2010-2012, adjusting for hospital variables and regional characteristics in the areas where the Spanish National Health System (SNHS) hospitals are located. Methods: 230 National Health System hospitals were studied over the two-year period 2010-2012 according to their ownership structure-public hospitals, private hospitals and public-private partnership (PPP)-data envelopment analysis orientated to inputs was used to measure the overall technical efficiency, pure efficiency and efficiency of scale. A generalised linear mixed model (GLMM) with binomial distribution and logit link function was used to analyse the hospital and regional variables associated with positioning on the frontier. Results: There are substantial differences between the average pure technical efficiency of public, private and PPP hospitals, as well as a greater number of PPP models being positioned on the efficiency frontier (91.67% in 2012). The odds of being positioned on the frontier are 41.7 times higher in PPP models than in public hospitals. The average annual household income per region is related to the greater odds of hospitals being positioned on the frontier of efficiency. Conclusions: During the most acute period of recession in the Spanish economy, PPP formulas favoured hospital efficiency, by increasing the odds of being positioned on the frontier of efficiency when compared to private and public hospitals. The position on the frontier of efficiency of a hospital is related to the wealth of its region.

Análisis multinivel de la eficiencia técnica de los hospitales del Sistema Nacional de Salud español por tipo de propiedad y gestión / Multilevel analysis of the technical efficiency of hospitals in the Spanish National Health System by property and type of management

Objetivo: Analizar la eficiencia técnica por tipo de propiedad y gestión de los hospitales generales del Sistema Nacional de Salud español (2010-2012) y sus variables explicativas hospitalarias y regionales. Método: Se estudian 230 hospitales combinando el análisis envolvente de datos y modelos transversales de regresión lineal multinivel de efectos fijos. Con el análisis envolvente de datos se miden la eficiencia técnica global, pura y de escala, y con los modelos multinivel, las variables explicativas de eficiencia. Resultados: El índice medio de eficiencia técnica global de los hospitales sin personalidad jurídica es inferior al de los hospitales con personalidad jurídica (0,691 y 0,876 en 2012). Existe una importante variabilidad en eficiencia técnica pura (ETP) por formas de gestión directa, indirecta y mixta. Un 29% de la variabilidad en la ETP es atribuible a diferencias entre comunidades autónomas. La dotación de personalidad jurídica del hospital aumenta en 11,14 puntos la ETP. Por otra parte, la mayoría de las formas de gestión alternativas al modelo tradicional aumentan en porcentajes variables la ETP. En el ámbito regional, según el escenario considerado, la insularidad y la renta media por hogar son variables explicativas de la ETP. Discusión: Tener personalidad jurídica favorece la eficiencia técnica. El marco de regulación y gestión de los hospitales, más que la propiedad pública o privada, parecen explicar la eficiencia técnica. Las características regionales explican de forma relevante la variabilidad en la ETP

Objective: To analyze technical efficiency by type of property and management of general hospitals in the Spanish National Health System (2010-2012) and identify hospital and regional explanatory variables. Method: 230 hospitals were analyzed combining data envelopment analysis and fixed effects multilevel linear models. Data envelopment analysis measured overall, technical and scale efficiency, and the analysis of explanatory factors was performed using multilevel models. Results: The average rate of overall technical efficiency of hospitals without legal personality is lower than hospitals with legal personality (0.691 and 0.876 in 2012). There is a significant variability in efficiency under variable returns (TE) by direct, indirect and mixed forms of management. The 29% of the variability in TE es attributable to the Region. Legal personality increased the TE of the hospitals by 11.14 points. On the other hand, most of the forms of management (different to those of the traditional hospitals) increased TE in varying percentages. At regional level, according to the model considered, insularity and average annual income per household are explanatory variables of TE. Discussion: Having legal personality favours technical efficiency. The regulatory and management framework of hospitals, more than public or private ownership, seem to explain technical efficiency. Regional characteristics explain the variability in TE

[Multilevel analysis of the technical efficiency of hospitals in the Spanish National Health System by property and type of management]. / Análisis multinivel de la eficiencia técnica de los hospitales del Sistema Nacional de Salud español por tipo de propiedad y gestión.

OBJECTIVE: To analyze technical efficiency by type of property and management of general hospitals in the Spanish National Health System (2010-2012) and identify hospital and regional explanatory variables. METHOD: 230 hospitals were analyzed combining data envelopment analysis and fixed effects multilevel linear models. Data envelopment analysis measured overall, technical and scale efficiency, and the analysis of explanatory factors was performed using multilevel models. RESULTS: The average rate of overall technical efficiency of hospitals without legal personality is lower than hospitals with legal personality (0.691 and 0.876 in 2012). There is a significant variability in efficiency under variable returns (TE) by direct, indirect and mixed forms of management. The 29% of the variability in TE es attributable to the Region. Legal personality increased the TE of the hospitals by 11.14 points. On the other hand, most of the forms of management (different to those of the traditional hospitals) increased TE in varying percentages. At regional level, according to the model considered, insularity and average annual income per household are explanatory variables of TE. DISCUSSION: Having legal personality favours technical efficiency. The regulatory and management framework of hospitals, more than public or private ownership, seem to explain technical efficiency. Regional characteristics explain the variability in TE.

Análisis de la eficiencia técnica en los hospitales del Sistema Nacional de Salud español / Analysis of the technical efficiency of hospitals in the Spanish National Health Service

Objetivo: Analizar la eficiencia técnica y la productividad de los hospitales generales del Sistema Nacional de Salud (SNS) español (2010-2012) e identificar variables hospitalarias y regionales explicativas. Métodos: Se estudian 230 hospitales del SNS mediante análisis envolvente de datos midiendo la eficiencia técnica global, pura, de escala, y el índice de Malmquist. La robustez del análisis se evalúa con modelos input-output alternativos. Se emplean modelos multinivel lineales transversales de efectos fijos para analizar las variables explicativas de eficiencia. Resultados: El índice medio de eficiencia técnica global (ETG) es de 0,736 en 2012, con una importante variabilidad por comunidades autónomas. El índice de Malmquist (2010-2012) es de 1,013. Un 23% de la variabilidad en ETG es atribuible a la comunidad autónoma. Las variables exógenas estadísticamente significativas (residentes por cada 100 facultativos, índice de envejecimiento, renta media anual por hogar, gasto en servicios públicos fundamentales y gasto público sanitario per cápita) explican el 42% de la variabilidad de ETG entre hospitales y el 64% entre comunidades autónomas. El número de residentes es estadísticamente significativo. En todas las comunidades autónomas existe una relación lineal directa significativa entre la ETG y la renta anual per cápita y el gasto en servicios públicos fundamentales, e indirecta con el índice de envejecimiento y el gasto público sanitario per cápita. Discusión El importante margen de mejora en eficiencia de los hospitales está condicionado por características específicas de cada comunidad autónoma, particularmente el envejecimiento, la riqueza y las políticas de gasto público de cada una (AU)

Objective: To analyse the technical efficiency and productivity of general hospitals in the Spanish National Health Service (NHS) (2010-2012) and identify explanatory hospital and regional variables. Methods: 230 NHS hospitals were analysed by data envelopment analysis for overall, technical and scale efficiency, and Malmquist index. The robustness of the analysis is contrasted with alternative input-output models. A fixed effects multilevel cross-sectional linear model was used to analyse the explanatory efficiency variables. Results: The average rate of overall technical efficiency (OTE) was 0.736 in 2012; there was considerable variability by region. Malmquist index (2010-2012) is 1.013. A 23% variability in OTE is attributable to the region in question. Statistically significant exogenous variables (residents per 100 physicians, aging index, average annual income per household, essential public service expenditure and public health expenditure per capita) explain 42% of the OTE variability between hospitals and 64% between regions. The number of residents showed a statistically significant relationship. As regards regions, there is a statistically significant direct linear association between OTE and annual income per capita and essential public service expenditure, and an indirect association with the aging index and annual public health expenditure per capita. Discussion: The significant room for improvement in the efficiency of hospitals is conditioned by region-specific characteristics, specifically aging, wealth and the public expenditure policies of each one (AU)

[Analysis of the technical efficiency of hospitals in the Spanish National Health Service]. / Análisis de la eficiencia técnica en los hospitales del Sistema Nacional de Salud español.

OBJECTIVE: To analyse the technical efficiency and productivity of general hospitals in the Spanish National Health Service (NHS) (2010-2012) and identify explanatory hospital and regional variables. METHODS: 230 NHS hospitals were analysed by data envelopment analysis for overall, technical and scale efficiency, and Malmquist index. The robustness of the analysis is contrasted with alternative input-output models. A fixed effects multilevel cross-sectional linear model was used to analyse the explanatory efficiency variables. RESULTS: The average rate of overall technical efficiency (OTE) was 0.736 in 2012; there was considerable variability by region. Malmquist index (2010-2012) is 1.013. A 23% variability in OTE is attributable to the region in question. Statistically significant exogenous variables (residents per 100 physicians, aging index, average annual income per household, essential public service expenditure and public health expenditure per capita) explain 42% of the OTE variability between hospitals and 64% between regions. The number of residents showed a statistically significant relationship. As regards regions, there is a statistically significant direct linear association between OTE and annual income per capita and essential public service expenditure, and an indirect association with the aging index and annual public health expenditure per capita. DISCUSSION: The significant room for improvement in the efficiency of hospitals is conditioned by region-specific characteristics, specifically aging, wealth and the public expenditure policies of each one.

Cost-effectiveness of percent free PSA for prostate cancer detection in men with a total PSA of 4-10 ng/ml.

OBJECTIVE: To assess the cost-effectiveness of two diagnostic strategies for prostate cancer in men with prostate-specific antigen (PSA) levels of 4-10 ng/ml and normal digital rectal examination (DRE). DESIGN: Cost-effectiveness analysis was performed using a decision tree model. Data collection and a systematic review of patients at the Urology Department (Carlos Haya Hospital) were made. 101 patients over the age of 40 with PSA levels of 4-10 ng/ml and normal DRE were selected. Transrectal ultrasound-guided prostate biopsy (TRUS-Bx) and percent free PSA testing prior to TRUS-Bx were performed. The outcome measures used were the incremental cost-effectiveness ratio, and costs were calculated through activity-based costing. The effectiveness was measured by means of the number of detected cases, test utility and actual cases (detected cases minus lost cases). RESULTS: Using base-case analysis, the strategy of percent free PSA + TRUS-Bx was found to be the most cost-effective. The incremental cost-effectiveness ratio for free PSA + TRUS-Bx compared with TRUS-Bx was EUR 2,277.40. Strategy 2 (TRUS-Bx) would be more cost-effective if the cost of percent free PSA increased to EUR 21.64 or if prostate cancer prevalence increased to 26%. CONCLUSIONS: The use of percent free PSA prior to TRUS-Bx is the most cost-effective diagnostic strategy. However, this result is very sensitive and strategy 2 (TRUS-Bx) would be more cost-effective if the cost of the percent free PSA increased to EUR 21.64 or if the prevalence of prostate cancer increased to above 26%.