Does Integration of Laboratory Data Improve Prescribing Decisions and Patient Outcomes?

Abstract

<p>Integrating laboratory information into prescribing tasks may improve medication safety. This thesis addresses several methodological issues in the progress of two studies: a systematic review of randomized trials addressing the impact of drug-lab safety alerts on adverse drug events and changes in prescribing or lab monitoring and a randomized trial using an electronic survey to compare prescribing decisions in complex clinical scenarios including integrated lab data with those in which the lab data were available on request. The systematic review found 32 studies; 10 addressed multiple drug-lab combinations, and 22 addressed single drug-lab combinations, including 14 targeting anticoagulation. We report a benefit of anticoagulation-related alerts (OR of an adverse event (bleeding or thrombosis) 0.88 (95% CI 0.78-1.00) and improved prescribing in multi-drug studies (OR 2.22, 95% CI 1.19-4.17), but substantial study heterogeneity precluded combining studies of other drugs. Methodological issues addressed in the RCT include medication selection, scenario design, recruitment, and assessment of the representativeness of the sample. We selected medications for study scenarios that are commonly prescribed by Canadian primary care physicians, and are associated with clinically important harm that may be preventable through laboratory monitoring. Data sources included IMS Brogan data on prescribing patterns and the Discharge Abstracts Database (DAD) and the National Ambulatory Care Reporting System (NACRS) from 2006-2007 to 2008-2009. Our study had 148 completed surveys. The study sample differed from the population of Ontario family physicians by gender, and use of electronic medical records. We found no difference in prescribing decisions (OR 1.21, 95% CI 0.84-1.75) between the study groups and no predictors of improved prescribing decisions. The lack of demonstrated impact of integrating lab data into clinical decision-making may be related to the study being underpowered, to a true lack of clinical benefit, or to a lack of discriminatory power in the scenarios.</p>