Tamblyn Robyn, Huang Allen, Perreault Robert, Jacques André, Roy Denis, Hanley James, McLeod Peter, Laprise Réjean
Department of Medicine, McGill University, Montreal, Que.
CMAJ. 2003 Sep 16;169(6):549-56.
Adverse drug-related events are common in the elderly, and inappropriate prescribing is a preventable risk factor. Our objective was to determine whether inappropriate prescribing could be reduced when primary care physicians had computer-based access to information on all prescriptions dispensed and automated alerts for potential prescribing problems.
We randomly assigned 107 primary care physicians with at least 100 patients aged 66 years and older (total 12 560) to a group receiving computerized decision-making support (CDS) or a control group. Physicians in the CDS group had access to information on current and past prescriptions through a dedicated computer link to the provincial seniors' drug-insurance program. When any of 159 clinically relevant prescribing problems were identified by the CDS software, the physician received an alert that identified the nature of the problem, possible consequences and alternative therapy. The rate of initiation and discontinuation of potentially inappropriate prescriptions was assessed over a 13-month period.
In the 2 months before the study, 31.8% of the patients in the CDS group and 33.3% of those in the control group had at least 1 potentially inappropriate prescription. During the study the number of new potentially inappropriate prescriptions per 1000 visits was significantly lower (18%) in the CDS group than in the control group (relative rate [RR] 0.82, 95% confidence interval [CI] 0.69-0.98), but differences between the groups in the rate of discontinuation of potentially inappropriate prescriptions were significant only for therapeutic duplication by the study physician and another physician (RR 1.66, 95% CI 0.99-2.79) and drug interactions caused by prescriptions written by the study physician (RR 2.15, 95% CI 0.98-4.70).
Computer-based access to complete drug profiles and alerts about potential prescribing problems reduces the rate of initiation of potentially inappropriate prescriptions but has a more selective effect on the discontinuation of such prescriptions.
药物相关不良事件在老年人中很常见，不适当的处方是一个可预防的风险因素。我们的目的是确定当基层医疗医生能够通过计算机获取所有已配发处方的信息以及针对潜在处方问题的自动警报时，不适当处方是否能够减少。
我们将107名拥有至少100名66岁及以上患者（共12560名患者）的基层医疗医生随机分为接受计算机化决策支持（CDS）的组或对照组。CDS组的医生可通过与省级老年人药物保险计划的专用计算机链接获取当前和过去处方的信息。当CDS软件识别出159个临床相关处方问题中的任何一个时，医生会收到警报，该警报会识别问题的性质、可能的后果和替代疗法。在13个月的时间里评估了潜在不适当处方的起始和停用率。
在研究前的2个月，CDS组31.8%的患者和对照组33.3%的患者至少有1张潜在不适当处方。在研究期间，CDS组每1000次就诊中新的潜在不适当处方数量显著低于对照组（18%）（相对率[RR]0.82，95%置信区间[CI]0.69 - 0.98），但两组在潜在不适当处方停用率方面的差异仅在研究医生与另一位医生的治疗重复（RR 1.66，95% CI 0.99 - 2.79）以及研究医生开具的处方引起的药物相互作用方面有显著性（RR 2.15，95% CI 0.98 - 4.70）。
通过计算机获取完整的药物信息档案以及关于潜在处方问题的警报可降低潜在不适当处方的起始率，但对这类处方的停用有更具选择性的影响。
