Findings from the implementation of a validated readmission predictive tool in the discharge workflow of a medical intensive care unit

Uchenna R. Ofoma, Subhash Chandra, Rahul Kashyap, Vitaly Herasevich, Adil Ahmed, Ognjen Gajic, Brian W. Pickering, Christopher J. Farmer

Research output: Contribution to journalArticlepeer-review

6 Scopus citations


Rationale: Provider decisions about patients to be discharged from the intensive care unit (ICU) are often based on subjective intuition, sometimes leading to premature discharge and early readmission. The Stability and Work Load Index for Transfer (SWIFT) score, as a risk stratification tool, has moderate ability to predict patients at risk of ICU readmission. Objectives: To describe findings following the incorporation of the SWIFT score into the discharge workflow of a medical ICU. Methods: The study involved 5,293 consecutive patients discharged alive from the medical ICU of an academic medical center. The SWIFT score and associated percentage risk for readmission were incorporated into daily rounds for purpose of discharge decision-making. We measured readmission rates before and after implementation and observed changes in provider discharge decisions for individual patients after SWIFT discussions. Measurements and Main Results: Baseline (n = 1,906) and implementation (n = 1,938) cohorts differed with respect to APACHE III scores (P = 0.03). In the implementation cohort, 26.2% of subjects had SWIFT scores greater than 15 and thus were predicted to have a higher risk of unplanned readmissions. In this high-risk group, 25% had SWIFT discussed in their discharge planning. There was modification of provider discharge decisions in 108 (30%) of cases in which the SWIFT was discussed. SWIFT score values above a prespecified cutoff of 15 were associated with physician tendency to prolong ICU stay or to discharge to a monitored setting (P < 0.001). There was no difference in 24-hour or 7-day readmission rates between the baseline and implementation cohorts (1.9 vs. 2.4%, P = 0.24; 6.5 vs. 7.4%, P = 0.26, respectively) even after adjustment for severity of illness. Conclusions: Using the SWIFT score as an adjunct to clinical judgment, physicians modified their discharge decisions in onethird of subjects. Introducing such tools into the discharge workflow may present change management challenges that limit the evaluation of their impact on readmission rates and other relevant ICU outcomes.

Original languageEnglish (US)
Pages (from-to)737-743
Number of pages7
JournalAnnals of the American Thoracic Society
Issue number5
StatePublished - Jun 2014


  • Care transitions
  • Quality
  • Readmissions
  • Risk stratification

ASJC Scopus subject areas

  • Pulmonary and Respiratory Medicine


Dive into the research topics of 'Findings from the implementation of a validated readmission predictive tool in the discharge workflow of a medical intensive care unit'. Together they form a unique fingerprint.

Cite this