Computer support for determining drug dose: systematic review and meta-analysis

Objective To review the effectiveness of computer support for determining o ptimum drug dose. Design Systematic review of comparative studies where computers gave advice to clinicians on the most appropriate drug dose. Search methods used were standard for the Cochrane Collaboration on Effective Professional Practice. Subjects Comparative studies conducted worldwide and published between 1966 and 1996. Main outcome measures For qualitative review, relative percentage differenc es were calculated to compare effects of computer support in different sett ings. For quantitative data, effect sizes were calculated and combined in m ete-analyses. Results Eighteen studies met the inclusion criteria The drugs studied were theophylline, warfarin, heparin, aminoglycosides, nitroprusside, lignocaine , oxytocin, fentanyl, and midazolam. The computer programs used individuali sed pharmacokinetic models to calculate the most appropriate dose. Meta-ana lysis of data from 671 patients showed higher blood concentrations of drug with computer support (effect size 0.69, 95% confidence interval 0.36 to 1. 02) and reduced time to achieve therapeutic control (0.44, 0.17 to 0.71). T he total dose of drug used was unchanged, and there were fewer unwanted eff ects of treatment Five of six studies measuring outcomes of care showed ben efit from computer assistance. Conclusions This review suggests that using computers to determine the corr ect dose of certain drugs in acute hospital settings is beneficial. Compute rs may give doctors the confidence to use higher doses when necessary, adju sting the drug dose more accurately to individual patients. Further researc h is necessary to evaluate the benefits in general use.