HEALTH APPLICATIONS FOR THE NATIONAL INFORMATION INFRA

点击翻译按钮获取中文摘要

We will reimplement two successful pharmacy-related expert systems, DoseChecker and Adverse Drug Event (ADE) Monitor, to exploit the existence of an extensive telemedicine network. DoseChecker provides drug dosing warnings based on pharmacy orders and patient specific features such as age, sex, weight, and estimated creatinine clearance. ADE Monitor examines patient drug orders and lab and lab test results, alerting hospital pharmacists when it detects signs of adverse drug events such as drug interactions. Once redeployed, these expert systems will be accessible to all clinical settings from which the clinical data repository collects the required data. During 1996, six hospital in Missouri and Illinois--Barnes, Jewish, Christian NE, Christian NW, Alton Memorial, and Barnes West County--will have the demographic, laboratory and pharmacy data required for DoseChecker and ADE Monitor entering the repository. This set of hospitals contains two academic facilities (Barnes and Jewish), two large community facilities (Christian NE and Alton Memorial) a small community facility (Barnes West County), and a predominantly ambulatory care facility (Christian NW). The reimplemented systems will serve as a telemedicine test bed to examine the impact of large-scale, wide-are deployment of clinical decision support in diverse clinical environments. Specifically, we will evaluate the expert system performance, physician acceptance, and clinical settings. By exploiting the existence of a standard clinical vocabulary and extensive networking interconnectivity, we can provide identical CDS functionality at clinical practice settings which preciously would pose significant technological barriers and require substantial rework. By keeping the decision support system constant, our evaluation studies will allow us to study the impact of clinical-setting variables on the effectiveness of these tow systems. In addition, we will implement and evaluate multiple decision support alert notification methods including GUI-based applications, E-mail, fax/printer reports, automated paging, and interactive voice massaging using commercial telephony products to examine the impact of alternative notification methods on the timeliness and clinical impact of decision support alerts. Using this telecommunications-based notification infrastructure, we also will investigate the patient confidentiality and privacy compromises which may occur due to alternative notification methods. By permitting multiple clinicians, possibly in different locations, to be notified of alerting conditions, we can also assess the impact or value of collaborative use of this technology to treat BJC patients. With our study decision, we can examine explicit trade-offs between alerting methods, decision support effectiveness, and patient confidentiality.