Objective:

To establish and quantify the determinants of hospital stay in a high complexity university hospital in Medellin between 2013 and 2018, assess their importance, and model the expected length of stay.

Methodology:

Retrospective analytical observational study of aggregate data. While following the method step by step, seven models were used, where mean hospital stay was the dependent variable and the respective independent variables were complexity, timeliness of diagnostic procedures, availability of supplies, cases of prolonged stay and financial capacity. The best model was selected using the Akaike and Bayesian information criterion, along with measures of both overall significance and individual significance of the coefficients. Statistical tests of model validity were performed and standardized coefficients were calculated.

Results:

The mean values of the most relevant variables and their standard deviation (SD) were: mean hospital stay, 8.09 days (SD = 0.40); complexity by resource consumption, 1.28 units (SD = 0.07); diagnostic procedures, 90.74 thousand studies (SD = 10.05); cases of extremely prolonged stay, 4.36% (SD = 0.70), and complexity by casuistry, 1 (SD = 0.03). Overall significance: F = 55.2, p < 0.001. Significance of coefficients: complexity by resource consumption, p < 0.01; diagnostic procedures and cases of extremely prolonged stay, p < 0.001; complexity by casuistry, p < 0.05. Standardized coefficients: complexity by resource consumption, 0.35; diagnostic procedures, 0.35; cases of extremely prolonged stay, 0.26; and complexity by casuistry, 0.24. Adjusted R2 0.82.

Conclusion:

In order of importance, the determinants of hospital stay are complexity by resource consumption, diagnostic procedures, extremely prolonged stay, complexity by casuistry, available inventory and gross profit.

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