Since the prospective payment system began in 1962, hospitals have been incentivized to shorten inpatient length of stay (LOS). Though average LOS has since become shorter, few studies have examined the impact of shorter LOS on patient outcomes. Such studies are particularly difficult because of the need to eliminate confounding. We used a unique study design to examine associations between LOS and outcomes of care.
We examined all admissions to the medical service of an urban medical center from July 1, 2002, through June 30, 2004. At admission, patients were assigned to a physician without bias. We included only physicians to whom admissions had been assigned in both years of the study, and whose mean LOS was not significantly different between years, and then we divided the physicians into quartiles based on their mean LOS during the study period. Next, we created 4 admission groups, which we defined according to the physician to whom the admission had been assigned. Admissions were thus assigned to a physician group that shared an LOS tendency. We then compared admissions assigned to physicians in the highest LOS quartile group (the longest LOS physicians) to admissions assigned to physicians in the lower 3 quartiles (the shorter LOS physicians), with respect to baseline characteristics (demographics, Charlson comorbidity score, prior admissions, and admission albumin) and 30‐day readmission and mortality rates. Finally, we constructed mixed‐effects logistic regression models using physician random intercepts to assess independent associations between physician LOS tendency and readmission and mortality rates.
Examined were 3221 admissions and 23 physicians. Of these, 2763 admissions were assigned to shorter LOS physicians, and 458 admissions were assigned to the longest LOS physicians. There were no significant differences between admission groups with respect to age, sex, race/ethnicity, insurance, Charlson comorbidity score, number of prior admissions, or admission albumin. In univariale analysis, admissions assigned to the longest LOS physicians had a lower 30‐day mortality rate than admissions assigned to shorter LOS physicians, but this difference was not significant (7.4% vs. 96%, P = 0.13). After adjustment forcovariates (age, Charlson score, admission albumin), admissions assigned to the longest LOS physicians had lower 30‐day mortality than did patients assigned to shorter LOS physicians (OR 0.74, 95% Cl 0.47–1.12), but this difference was still not significant. There was also no difference in 30‐day readmissions (16.4% vs. 16.5%, P = 0.96).
We used a unique study design to examine associations between physician LOS tendency and mortality but had limited power to detect differences between patient groups. Compared with admissions assigned to physicians with shorter length of stay, admissions assigned to physicians with the longest LOS had lower 30‐day mortality, but the difference was not significant.
W. Southern, none; J. Arnsten, none.
To cite this abstract:Southern W, Arnsten J. Patients Cared for by Physicians with a Tendency toward Shorter Lengths of Stay May Have Increased Mortality. Abstract published at Hospital Medicine 2010, April 8-11, Washington, D.C. Abstract 137. Journal of Hospital Medicine. 2010; 5 (suppl 1). https://www.shmabstracts.com/abstract/patients-cared-for-by-physicians-with-a-tendency-toward-shorter-lengths-of-stay-may-have-increased-mortality/. Accessed January 26, 2020.