Paediatric Hepatology Dependency score (PHD score): an audit tool.

Background: The aim was to develop a tool that measures patient dependency and disease severity in children with mild to severe liver disease.
Methods: The initial pilot score was based on known markers of disease severity in acute and chronic liver disease. Between 1997 and 2001, the score was modified 4 times and reduced to comprise 10 key parameters: aspartate transaminase, prothrombin time, albumin, bilirubin, ascites, nutritional support, organ dysfunction, blood product support, sepsis and intravenous access. The score's face validity, internal consistency, interobserver agreement and construct validity were evaluated statistically and by the use of endpoints such as survival and transplant after 6 months.
Results: The final 10 variable score was tested on 71 children admitted over two 3-month periods. The Cronbach's alpha score (a test for internal consistency) for the total score was 0.72. Serial data paralleled clinical deterioration and response to interventions. A Paediatric Hepatology Dependency score >15 had an odds ratio of 7 (P = 0.0125) for death or transplant at 6 months. The score was also found to agree with the paediatric end-stage liver disease score for the 12 patients being listed for liver transplantation (r = 0.660, P < 0.05).
Conclusion: The Paediatric Hepatology Dependency score is valid and internally consistent and is a convenient measure of dependency and disease severity in a heterogeneous group of patients with liver disease. It also allows admissions to be audited for comparison between eras and for monitoring the progress of patients while on the ward.