PATIENT CARE REMINDER MODULES DEMONSTRATION PROJECT

点击翻译按钮获取中文摘要

Computerized patient care reminders will be designed and their impact and cost tested in a large free-standing ambulatory health care center. Reminders will be tested for two tasks concerning follow-up of abnormal laboratory test results, and one task concerning prompts to perform health maintenance checks. Specific aims are to develop computerized reminder modules for public-use COSTAR; to produce a manual of procedures for the development of computerized reminder modules; to measure and describe difficulties in implementation of the modules; to evaluate the effectiveness of computerized reminders on compliance with patient care protocols, service use, and medical yield; and to assess the costs of the development, implementation, and operation of the computerized reminder modules. For the evaluation, a randomized controlled trial is proposed. For one test task (follow-up of positive tests from gonorrhea), patients will be randomly assigned to the reminder or control group (N=96). For the other two tasks (follow-up of low hematocrits, N=230, and cancer screening in women, N=96) three of six medical assistant units will be randomly assigned to reminder status for one of the tasks, while the other three will be in reminder status for the other task. Each medical assistant unit will be in control status for the task for which they are not in reminder status. Patients will be in reminder versus control status according to the status of the medical assistant unit by which they are served. Data on compliance with reminders, and on service use and yield, will be gathered from medical records. Data on the use made of reminders will be gathered from a computer file of all reminder transactions. Detailed analysis of development and operating costs of the modules will be based on personenel time, material supplies and physical resources used, and computer costs. Analysis of changes in practitioner compliance with protocols, service use, and medical yield will be accomplished using logistic regression models.