• English 
  • Spanish 


Healthcare performance indicators for the Spanish Health System (length of stay, mortality, etc.)

MinisterioSanidad.png Project type: National project
Funding institution: Spanish Health Ministry
Program: Research/Trasfer Contracts
Code: UC 20.640.01.2346
Period: March 2007 - March 2008
Status: Finished
Web: web

The assessment of Health Systems (HS) performance is an important and difficult problem which has been exposed to continuous debate over the last decades. Traditionally, this problem was approached using a combination of empiricism and expert knowledge, with high doses of complexity and personal judgment. However, modern policies aimed at HS development and improvement have favored the development of indicators (quantitative performance measures), to allow the rating and benchmarking of healthcare delivers (e.g., hospitals). These measures are mainly derived from indicators of hospital outcomes and each of them provides a particular picture of some aspects of the healthcare process (length of stay, mortality rate, etc.).

Spain is a 44.4 million inhabitants’ country (51% females) with a decentralized management National Health System (NHS) covering 94.4% of the population. There are a total of 798 hospitals with different clinical profiles attending an average of 4.5 million admissions per year. A basic dataset for the NHS was adopted in 1987 (MBDS: Minimum Basic Data Set of hospitalization of Spanish public hospitals), including clinical and administrative data describing different aspects of the patients and the hospital activity.

In this project we analyze the variability observed in hospital outcomes, as measured by indicators related to hospital performance and quality of care (length of stay, mortality rate, etc.) obtained using the MBDS (the 3.5 million records for 2005):

  • Average stay (in days)
  • Mortality rate (%)
  • Reentry rate, after 30 days (%)
  • Nosocomial infection rate (%)
  • Surgery average wait time (in days)
  • Cesarean rate (%)

In order to use these indicators for benchmarking purposes, the hospital indicator values must be adjusted according to the influence of external factors (e.g., different clinical and patient profiles) not attributable to the hospital performance. To this aim, we introduce a novel multivariate statistical model for discrete data (a probabilistic Bayesian network) to find relevant potential factors of influence for the indicators among the above set of variables. The model allows obtaining the most significant variables influencing the hospital outcome indicators, thus explaining part of the variability observed among different hospitals in terms of clinical and diagnostic profiles. Thus, the outcome indicators can be adjusted for an unbiased estimation of hospital care quality.