Software Engineering: A Practitioner's Approach

The Worthing physiological scoring system (PSS) was first validated in 2005 as a tool to predict hospital mortality on admission and was subsequently introduced into clinical practice at Worthing Hospital, UK. Five years on, this study was conducted to determine the effects on mortality and length of stay (LOS) after the introduction of electronic alerting software using the PSS. In addition, we investigated whether the Worthing PSS predictive ability could be improved by addition of further variables.Prospective observational study conducted in the acute medical unit, Worthing Hospital, UK. Patient physiological data on admission and discharge/transfer were collected between February and July 2010 from the electronic alerting software VitalPAC™. Patient characteristics, co-morbidity, outcomes, and biochemistry data were taken from the hospital administration and pathology systems.The observed mortality reduction from 8.3% to 5.2% over 5 yr was not statistically significant after adjustment for admission Worthing PSS score. Median LOS was reduced from 4 to 2 days, but this reflected an increase in short stay admissions. Worthing PSS was not significantly improved with the addition of biochemical variables or patient co-morbidity. A score taken before admission to a medical ward showed an improved predictive ability when compared with the initial admission score, but further analysis found no additional clinical benefit.The introduction of an electronic alerting PSS did not lead to a reduction in mortality when adjusted for severity of illness defined by physiological variables. Predictive performance was not enhanced by the addition of biochemical variables and co-morbidities.