We previously developed and validated a nomogram to predict discharge disposition post OLT. The current study analyzes our experience using the discharge calculator.

Patients &Methods:

We retrospectively reviewed data on 430 adults transplanted between 01/2012 and 11/2015. Patients that died <30 days were excluded. The discharge calculator, available online, utilizes the following: Karnofsky score, age, INR, creatinine, diabetes, dialysis, bilirubin, albumin, and BMI. The calculation performed at admission for OLT was shared with the transplant team. Patients with a predictive value of ≥ 60% were considered high risk for discharge to a facility. Early discharge planning and physical therapy (PT) were based on discharge probability. Continuous variables were described by means and standard deviations. Predictions from existing nomogram were validated against the actual observed outcomes from the prospectively collected data using the C-index (ROC AUC).

Results:

132/430 (30%) of patients were discharged to a facility with a predictive score of 60±24% and 298 were discharged home with a predictive score of 27±22% (p<0.0001). The characteristics of the population going to a facility were: older age (p=.004), higher MELD score (p<0.001), higher BMI (p=0.001), number of hospital admissions prior to OLT (p=0.002), higher creatinine (p<0.001), longer ICU length of stay (p<0.001), recipients of cadaveric grafts vs. living donor grafts (p=0.04), hospitalized at the time of OLT (p<0.001), encephalopathy (p<0.001), arrhythmia (p<0.001), dialysis (p<0.001), ascites and hydrothorax (p<0.001).

107 patients had a predictive value >60% for discharge to a facility. 31(28%) of these were discharged home. The only significant factor in the patients that were able to be discharged home was less time in ICU, 2.3 ± 1.9 vs. 9.1±13 days (p<0.001).

The concordance index for discharge disposition prediction is 0.835. For higher predictions >80% (n=44) the model over-predicts discharge to a facility, but even at this level 75% (n=33) of patients are actually discharged to a facility. The intensive discharge planning and PT regimen may impact the discharge disposition in these patients.

Conclusion:

The discharge calculator is a vital component in early discharge planning for the transplant team. The concordance index of 0.835 is reflected in its accuracy in clinical practice. The over-prediction at higher values may reflect the response of the clinical team to the initial predictive value.

CITATION INFORMATION: Kerepesi K, Waby A, Eid K, Plescia J, Zinn K, Vass K, Eghtesad B, Miller C, Kelly D. Experience Using a Discharge Calculator for Discharge Planning Post Liver Transplantation (OLT). Am J Transplant. 2016;16 (suppl 3).

« Back to 2016 American Transplant Congress