SciELO - Scientific Electronic Library Online

vol.38 issue1Adverse events after treatment with the transfer factorAbortion in teenagers within the legal contex author indexsubject indexarticles search
Home Page  

Services on Demand




Related links

  • Have no similar articlesSimilars in SciELO


Revista Cubana de Salud Pública

On-line version ISSN 1561-3127Print version ISSN 0864-3466


TAMARGO BARBEITO, Teddy Osmin; JIMENEZ PANEQUE, Rosa Eugenia; GUTIERREZ ROJAS, Ángela Rosa  and  MORA DIAZ, Isabel. Two risk adjustment methods for length of stay as indicator of hospital performance. Rev. cub. salud pública [online]. 2012, vol.38, n.1, pp.29-44. ISSN 1561-3127.

Introduction: the length of stay at hospital is the indicator of the rendered service efficiency par excellence, so any assessment of the hospital performance based on this indicator should take into account in one way or another the characteristics of the patients considered for the value estimation. Objective: to evaluate two risk adjustment methods for the length of stay as indicator of the hospital performance. Methods: a retrospective study was conducted at the internal medicine service of "Hermanos Ameijeiras" hospital from May to October, 2006. The sample of 606 medical histories was randomly divided into two parts, that is, one group comprised 304 medical histories for the predicted length of stay estimation from the multiple linear regression function and from the patient classification by related diagnosis groups, and the other included 302 medical histories for the validation of these two risk adjustment methods. In the validation phase, the capacity of each procedure to detect care deficiencies in another group of medical histories through the variance analysis and the ROC curve construction was evaluated. Results: age, severity index, main diagnosis on discharge and its interactions, and sex influenced the length of stay. The area under the ROC curve with the multiple linear regression was 0.747 (p< 0,001) (95 % CI, 0.690-0.805) whereas the same are with the related diagnosis groups was 0,738 (p< 0,001) (95 % CI, 0.680-0.796). Conclusions: both risk adjustment methods are equally effective in detecting efficiency problems, but the multiple linear regression model is better than the related diagnosis groups in estimating the predicted length of stay due to economic reasons. This aspect supports its use in poor-resource countries or institutions as is the case of underdeveloped countries.

Keywords : Length of stay at hospital; efficiency; Related Diagnosis Groups; risk adjustment.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )