Determining the Effects of Patient Casemix on Length of Hospital Stay: A Proportional Hazards Frailty Model Approach

Summary

Objectives: To identify factors associated with hospital length of stay (LOS) and to model variations in LOS within Diagnosis Related Groups (DRGs).

Methods: A proportional hazards frailty modelling approach is proposed that accounts for patient transfers and the inherent correlation of patients clustered within hospitals. The investigation is based on patient discharge data extracted for a group of obstetrical DRGs.

Results: Application of the frailty approach has highlighted several significant factors after adjustment for patient casemix and random hospital effects. In particular, patients admitted for childbirth with private medical insurance coverage have higher risk of prolonged hospitalization compared to public patients.

Conclusions: The determination of pertinent factors provides important information to hospital management and clinicians in assessing the risk of prolonged hospitalization. The analysis also enables the comparison of inter-hospital variations across adjacent DRGs.

Keywords

Diagnosis-Related Groups - Length of Stay - Proportional Hazards Models - Survival Analysis

• References

• 1 Fetter RB, Youngsoo S, Freeman JL. et al. Case-mix: definition by diagnosis related groups. Medical Care 1980; 18 Suppl 2 1-53.
• 2 Hindle D, Degeling P, van der Wel O. Severity variations within DRGs: Measurement of hospital effects by use of data on significant secondary diagnoses and procedures. Aust Health Review 1998; 21: 37-49.
• 3 Beaver C, Zhao Y, McDermind S, Hindle D. Casemix-based funding of Northern Territory public hospitals: adjusting for severity and socio-economic variations. Health Economics 1998; 7: 53-61.
• 4 Leung K-M, Elashoff RM, Rees KS. et al. Hosptial and patient-related characteristics determining maternity length of stay: a hierarchical linear model approach. Amer J Pub Health 1998; 88: 377-81.
• 5 Marazzi A, Paccaud F, Ruffieux C, Beguin C. Fitting the distributions of length of stay by parametric models. Medical Care 1998; 36: 915-27.
• 6 Lee AH, Xiao J, Vemuri SR, Zhao Y. A discordancy test approach to identify outliers of length of hospital stay. Statist Med 1998; 17: 2199-206.
• 7 McGilchrist CA. REML estimation for survival models with frailty. Biometrics 1993; 49: 221-5.
• 8 Yau KKW, McGilchrist CA. ML and REML estimation in survival analysis with time dependent correlated frailty. Statist Med 1998; 17: 1201-13.
• 9 MathSoft.. S-PLUS 2000 Guide to Statistics. MathSoft, Inc; 1999
• 10 Efron B. The efficiency of Cox’s likelihood function for censored data. J Amer Statist Assoc 1977; 72: 557-65.
• 11 Carty EM, Bradley CF. A randomized controlled evaluation of early postpartum hospital discharge. Birth 1990; 18: 199-204.
• 12 Brooten D, Roncoli M, Finkler S. et al. A randomized trial of early hospital discharge and home follow-up of women having cesarean birth. Obstet Gynecol 1994; 84: 832-8.
• 13 Shorten A. Obstetric early discharge versus traditional hospital stay. Aust Health Review 1995; 18: 19-39.
• 14 Solomon GL. Length of the hospital stay for mothers and newborns. N Engl J Med 1996; 334: 1134.
• 15 Braveman P, Egerter S, Pearl M. et al. Problems associated with early discharge of newborn infants. Early discharge of newborns and mothers: a critical review of the literature. Pediatrics 1995; 96: 716-26.
• 16 Xiao J, Douglas D, Lee AH, Vemuri SR. A Delphi evaluation of the factors influencing length of stay in Australian hospitals. Int J Health Plan Manage 1997; 12: 207-18.