HandheldBased Decision Support in Trauma Medicine Bla Zupan

Handheld-Based Decision Support in Trauma Medicine Blaž Zupan Faculty of Computer Science, University of Ljubljana Baylor College of Medicine, Houston

Examples of Medical Decision Support • Probability of recurrence after radical prostatectomy • Predicting patient’s long term clinical status after hip arthroplasty • Wrist injuries: should we operate?

Why? • Second opinion • Systematic approach • Decision analysis – which factors are important – how do they influence the decision • Prognosis, • Treatment planning, • Diagnosis, . . .

Why handhelds? • Handhelds are practical • Decision support tool should be available wherever decision takes place • They provide enough computing power

Decision Support Models data (EPR, clinical data bases, . . . ) decision support model statistical analysis data mining background (expert) knowledge patient’s data decision (prognosis, diagnosis, treatment plan, . . . )

Sucess Story: Prostogram • MW Kattan, P Fern: From 1999 in regular use at Memorial Sloan-Kettering Hospital and BCM

Decision Support Shell on Palm(TM) statistical analysis data mining decision support models in XML (synchronization) decision support

An example. . . Aoki N. et al: Mathematical Analysis of Data from the Hanshin. Awaji Earthquake.

. . . An example

Current Status • Applications – BCM & Ben Taub Hospital, Houston: traumatology – Smrke (Clinical Center in Ljubljana, Slovenia): hip, wrist injury – Kattan (Sloan-Kettering, NY): prostate cancer • Methodology – Logistic regression, other methods (table lookup, naive Bayes, survival models) coming soon – Integration with data mining tools coming soon