With the passage of the Health Information Technology for Economic and Clinical Health (HITECH) Act, widespread adoption of certain health information technologies, such as electronic health records (EHRs) and electronic prescribing (e-prescribing), is imminent. Drug-drug interaction (DDI) screening and medication history information are commonly incorporated into health information exchange systems to improve medical decision making, safety, and quality of care, but the value of these features is unclear. To evaluate the effect of providing access to an early generation electronic medication management program with medication history accessible to prescribers via a wireless handheld personal digital assistant (PDA) device on the incidence of potential DDIs (i.e., DDIs that may or may not cause patient harm). This study employed a retrospective pre-intervention/post-intervention study design with a comparison group to evaluate the effectiveness of a wireless handheld medication management program in preventing serious potential DDIs. Licensed prescribers in a state Medicaid program who wrote prescriptions during the period from August 2003 through June 2006 were included in this study. The intervention (PDA) group consisted of clinicians who requested and were granted access to the wireless handheld device containing prescription drug history between August 1, 2004, and June 30, 2005. Initially the device contained 100-day patient-specific medication history, but other functionalities were added during the study period including the ability to check for drug-drug interactions and e-prescribing. The comparison group consisted of prescribers who sent a request to obtain, but did not receive, the wireless handheld device during the same time period. Baseline prescribing patterns of 25 previously identified clinically important potential DDIs were assessed over two 12-month periods, one period prior to (baseline) and one period after (follow-up) an index date (date of device deployment for PDA group; date of request for comparison group). A random-effects negative binomial model was used to analyze the primary outcome, the number of potential DDIs per prescriber per 12-month time period. A secondary outcome of interest, the likelihood that a prescriber would prescribe at least 1 potentially interacting medication pair during the baseline and follow-up periods, was analyzed using a random-effects logistic model. A total of 1,615 prescribers constituted the PDA group, and 600 prescribers made up the comparison group. Prescribers in the 2 groups were significantly different in their specialty practice areas (P less than 0.001), number of pharmacy claims at baseline (P less than 0.001), and the likelihood of prescribing at least 1 potential DDI combination during the 1-year baseline period (P=0.003). However, the prescriber groups were similar in their average age (P=0.241) and geographic location (P=0.181). The most widely prescribed potential DDIs included those involving warfarin with nonsteroidal anti-inflammatory drugs (NSAIDs) and thyroid hormones. The median number of patient medication history updates requested per PDA group prescriber during follow-up was 24 (range 0 to 1,073). At baseline, 1,104 (68.4%) of the PDA group and 449 (74.8%) of the comparison group had no potential DDIs. During the next year, 1,131 (70.0%) and 462 (77.0%) of the PDA group and comparison group, respectively, had no DDIs. The incidence rate ratio was 1.01 (95% CI=0.87-1.17) for the PDA group relative to the comparison group for change in number of potential DDIs. In the logistic regression model, the odds of prescribing at least 1 potential DDI did not significantly differ by group (odds ratio=1.26, 95% CI=0.96-1.66). These results indicate that there was no significant difference between the intervention and comparison group with regard to the change in the rate of potential DDIs between the baseline and follow-up periods. A stand-alone medication management program in a wireless PDA device was not frequently used by most prescribers to update patient medication histories and was not associated with a reduction in the rate of prescribing potentially clinically important DDIs.
Read full abstract