Predictors and determinants of hospital length of stay in congestive heart failure in ten community hospitals

• Published in: The Journal of heart and lung transplantation Vol. 16; no. 5; p. 548
• Main Authors: Philbin, E F, Rocco, Jr, T A, Lynch, L J, Rogers, V A, Jenkins, P
• Format: Journal Article
• Language: English
• Published: United States 01-05-1997
• Subjects: Age Factors; Aged; Female; Health Services Research; Heart Failure - etiology; Heart Failure - therapy; Hospitals, Community; Humans; Length of Stay; Linear Models; Male; Predictive Value of Tests; Retrospective Studies; Risk Factors; Severity of Illness Index; Survival Analysis
• ISSN: 1053-2498

Little is known about the actual determinants of hospital length of stay (LOS) among patients admitted with congestive heart failure (CHF), in spite of its economic impact. To increase understanding of these factors, we examined the demographic, clinical, laboratory, and treatment characteristics of patients hospitalized with decompensated CHF.
The charts of consecutive patients admitted to 10 acute care community hospitals during 1995 were reviewed. The relationship between LOS and more than 140 patient-specific variables were examined. First, patient characteristics identifiable within the first 24 hours of hospitalization were examined for their relationship with LOS. Then, variables indicative of the processes of care and response to treatment were studied. Finally, administrative data were added to yield the final model for LOS.
During the study period 1402 patients were admitted to the participating centers. The patients were predominantly elderly with moderately severe or severe CHF. With stepwise multiple linear regression, 5% of the variation in LOS could be explained by baseline characteristics alone (r = 0.22, p < 0.0001). When treatment and response variables were added to this model, 15% of the variation in LOS could be explained (r = 0.39, p < 0.0001). When administrative data were added, the final model explained 31% of the variation in LOS (r = 0.56, p < 0.0001).
We conclude that LOS among patients hospitalized with decompensated CHF is partially related to patient demographics, severity of illness, management modalities, response to treatment, and administrative data. However, significant residual variation in LOS exists, which cannot be explained by these factors. These observations may be of value in the design and implementation of initiatives aimed at reducing resource utilization and improving quality of care in CHF.