by some hospitals, medical centers, and group practices) (Turban et al., 1996). In the future, however, the Internet will likely be the platform of choice for many if not most health applications because of the ready access it provides to both consumers and clinicians, as well as other financial and technical considerations.
It must be acknowledged that although the potential benefits of IT are compelling, the evidence in support of these benefits varies greatly by type of application. As discussed in Chapter 6, there is strong evidence to support the effectiveness of computerized reminder systems in improving compliance with practice guidelines. For computerized medication order entry systems, recent studies substantiate reductions in errors and unnecessary services, but such studies are few in number (Bates et al., 1998a). A recent review of 80 controlled trials carried out between 1966 and 1996 concluded that telephone-based distance medicine or telemedicine technologies are beneficial in the areas of preventive care and the management of osteoarthritis, cardiac rehabilitation, and diabetes care (Balas et al., 1997). In a review of 15 controlled trials in which diabetic patients received computer-generated information, it was found that 12 of the 15 trials documented positive clinical outcomes, such as improved hemoglobin and blood glucose levels (Balas et al., 1998).
In summary, the strength of the evidence on the effects of various IT applications is highly varied. Many applications, such as simulation of surgical procedures for educational purposes and remote and virtual surgery, are in the early developmental stages. Others may be highly promising, but their adoption and testing are hampered by the lack of computerized patient information (e.g., computer-aided diagnosis), regulatory or legal impediments (e.g., e-mail communications across state lines), and payment issues (e.g., for e-visits). Still other applications, such as telemedicine, have not been rigorously evaluated (Grigsby and Sanders, 1998; Institute of Medicine, 1996).
Much of the potential of IT to improve quality is predicated on the automation of at least some types of clinical data. Automated clinical data are required by many of the most promising IT applications, including computer-aided decision support systems that couple medical evidence with patient-specific clinical data to assist clinicians and patients in making diagnoses and evaluating treatment options (see Chapter 6) (Burger, 1997; Weed and Weed, 1999). Automated clinical data also open up the potential to glean medical knowledge from patient care (Institute of Medicine, 2000). An example is the extraordinary gains in cancer survival for children as compared with adults, attributable in part to the participation of virtually all pediatric cancer patients in clinical trials that systematically collect, pool, and analyze data and disseminate results to all participants (Simone and Lyons, 2000). Automated clinical and administrative data also enable many types of health service research applications, such as assessment of