A total of 300 records were reviewed, of which 285 were eligible for analysis; exclusions from analysis were due primarily to inadequate information regarding organ system function. The primary diagnosis at the acute care hospital was available in 87% of patients, and complete data regarding comorbidities were available in 69% of patients. Comorbidities that occurred in at least 5% of patients are shown. On LTAC admission, urinary catheters were present in 75% of patients, feeding tubes in 64%, and central venous catheters in 45%.

The number of patients with each OSF and the associated percent mortality with the OSF. Survival is compared among subjects in different OSF categories. One patient with six OSFs and nine patients with five OSFs are grouped together with patients with four OSFs, resulting in five OSF categories Prednisone in Canada. The relative risks of mortality in those with OSFs, compared to those with no OSFs, were 3.3 with one OSF (confidence interval [CI], 1.4 to 7.8; p < .001), 11.9 with two OSFs (CI, 4.8 to 29.6; p < .001), 31.3 with three OSFs (CI, 9.1 to 107.3; p < .001), and 58.2 with four or more OSFs (CI, 12.0 to 281.6; p < .001). No III score. The area under the receiver operating characteristic curve for the prediction equation was 0.81 (p < 0.001) indicating good model discrimination. The Hosmer-Lemeshow test p value was 0.39, indicating good model calibration.

A Kaplan-Meier survival analysis comparing inpatient survival probabilities among those in different OSF categories (log-rank test, p < 0.0001). Most patients with no OSFs were removed from analysis due to discharge from the LTAC facility, whereas patients with four or more OSFs were removed from analysis primarily due to death. No effort was made to determine the long-term survival of those discharged alive. A Kaplan-Meier survival analysis comparing survival at 28 days after LTAC admission in the same groups yielded similar results, with a log-rank test p value < 0.0001.

Of those patients who remained at the LTAC at day 28 (n = 159), OSF score increased in those who ultimately died (p = 0.05) and decreased by half in those who were subsequently discharged home (p = 0.04). There was a nonsignificant trend toward improvement in those who were discharged to nursing homes (p = 0.07), and no change in those who were eventually transferred to acute care.