COMPUTER-BASED PROSPECTIVE DRUG UTILIZATION REVIEW

药物利用审查（DUR）已迅速成为质量的主要焦点 在医疗界的护理，并已成为一个主要的金融 联邦政府内部的利益。 随着1990年OBRA的颁布， 各州必须在1993年1月之前执行DUR方案。 DUR是一个通过比较实际的药物治疗来改善药物治疗的过程。 药物使用，以建立预定的标准，并提供结果 与开处方的医生相比。 DUR有两个基本的 方法：回顾性和前瞻性。 每一种都有其固有的 局限性，但乐观的是，前瞻性DUR提供有效的 优势 目前缺乏对照试验来支持这一点 乐观. 以前在印第安纳州对医生决策的研究 大学和Regenstrief研究所建议， 在做出决定时向卫生保健提供者发送提醒 可能会提高对治疗建议的依从性。 这些提醒使用Regenstrief病历中的患者数据 系统（RMRS）纵向，合并，住院和门诊 电子病历。 遵守情况的改善是 证明了预防保健提醒和实验室测试订购。 这项拟议的研究将在一项随机对照试验中检查 由基于规则的专家系统生成的提醒的效果， 在处方时反馈给处方医生 书面或药剂师在处方配药时。 的 专家系统将结合联合收割机条件和病人的具体治疗 建议与先前记录的存储在 RMRS创建高度针对患者的提醒。 治疗标准 将根据科学文献和纲要制定， 以及当地专家共识小组。 此外，具体条件 将开发预测模型， 例如作为治疗的一部分提供给医生的住院治疗， 反馈 三种目标条件（充血性心力衰竭，反应性 气道疾病和冠状动脉疾病）被选为DUR研究 因为它们代表了常见的、致残的和病态的疾病。 的160 医生在学术，内城，全科医学实践将是 随机分配至（1）常规护理（仅限于医院正在进行的 回顾性DUR）;（2）医生靶向前瞻性DUR;（3） 药剂师针对性的前瞻性DUR;或（4）联合医生和 药剂师针对预期DUR。 五百个病人， 每隔6个月，将确定并随访两个 年 主要结局指标包括功能状态（SF-36 评分）、具体病情问题、患者健康满意度 护理（PSQ评分），以及随后的药物依从性， 住院 次要结局指标将由医生和 药剂师对指南的依从性、患者死亡率和直接 医疗保健费用。 这个项目将有相当大的力量来探测 主要结果的差异。 如果显示出积极影响 患者的功能状态、对医疗保健的满意度或健康状况 护理利用率，基于计算机的DUR将提供一个具有成本效益的工具 来改善医疗服务。

Drug utilization review (DUR) has rapidly become a major focus of quality of care within the medical community and has become a major financial interest within the federal government. With the enactment of OBRA 1990 individual states are mandated to implement DUR programs by January 1993. DUR is a process for improving pharmaceutical therapy by comparing actual drug use to established predetermined criteria and providing the results of this comparison to prescribing physicians. DUR has two basic approaches: retrospective and prospective. Each of these have inherent limitations, but optimism exists that prospective DUR offers efficacious advantages. Controlled trials are currently lacking to support this optimism. Previous research into physician decision making at Indiana University and the Regenstrief Institute suggests that patient specific reminders delivered to the health care provider at the time a decision is being made may improve compliance with treatment recommendations. These reminders utilize patient data from the Regenstrief Medical Record System (RMRS) a longitudinal, combined, in-patient and out-patient electronic medical record. Improvement in compliance has been demonstrated for preventive care reminders and laboratory test ordering. The proposed study will examine in a randomized controlled trial the effect of reminders generated by a rule-based, expert system which will be fed back to the prescribing physician at the time of prescription writing or to pharmacists at the time of prescription dispensing. The expert system will combine condition and patient specific treatment recommendations with previously recorded patient data which is stored in the RMRS to create highly patient specific reminders. Treatment criteria will be developed based ont he scientific literature and compendia as well as a local expert consensus panel. In addition, condition specific predictive models will be developed and probabilities of patient outcomes such as hospitalization provided to the physician as part of the feedback. Three target conditions (congestive heart failure, reactive airways disease, and coronary artery disease) were selected for DUR study because they represent common, disabling, and morbid diseases. The 160 physicians in an academic, inner city, general medicine practice will be randomized to (1) usual care (limited to the hospital's ongoing retrospective DUR); (2) physician targeted prospective DUR; (3) pharmacist targeted prospective DUR; or (4) combined physician and pharmacist targeted prospective DUR. Five hundred patients with each condition will be identified and followed at 6 month intervals for two years. Primary outcome measures will include functional status (SF-36 scores), condition specific questions, patient satisfaction with health care (PSQ scores), and subsequent medicine compliance and hospitalization. Secondary outcome measures will be physician and pharmacist compliance with guidelines, patient mortality, and direct health care costs. This project will have considerable power to detect differences in the primary outcomes. If shown to positively affect patients' functional status, satisfaction with health care, or health care utilization, a computer-based DUR will offer a cost-effective tool for improving health care delivery.