conditions. Indeed, such a fragmented approach could place patients at increased risk for medication errors, similar to the risk that results from having multiple different health care providers each independently prescribing medications in their own discipline. An example of an alternative is the Program of All-Inclusive Care for the Elderly (PACE), which has shown great promise for providing comprehensive care for the very frail elderly using a multidisciplinary approach (Bodenheimer, 1999).
The committee believes that a team approach to medication use is essential. Different providers will be involved at each step of the process. But all providers who are prescribing, whether a physician, nurse practitioner, or physician’s assistant, or administering, whether a registered nurse or licensed practical nurse, need to have the appropriate competencies, which should be determined by their professional organizations and the care organization for which they work.
Providers should make effective use of well-designed technologies, which will vary by setting.
Judicious use of technology will be important in improving medication safety (Bates and Gawande, 2003). While the evidence supporting this statement is strongest for the inpatient setting (AHRQ, 2005), the use of technology will undoubtedly result in major improvements in all settings, although the specific technologies and relative benefits will likely differ by setting. Much remains to be learned in all settings. Moreover, as noted earlier, any technology can introduce errors as well as prevent them (Ash et al., 2004), and it is essential to monitor any new technology and make appropriate midcourse corrections. And even highly promising technologies may not yield the desired safety benefits if issues of safety culture and efficiency are not adequately addressed (Rothschild, 2004).
In the inpatient setting, strong evidence demonstrates that CPOE reduces rates of serious medication errors in adults (Bates et al., 1998; AHRQ, 2005), although the impact on preventable ADEs is uncertain since a large randomized controlled trial has not been conducted. A key issue regarding CPOE is the depth and breadth of the decision support provided. Moroever, the main impact of CPOE is on ordering and transcription errors; the technique has relatively little impact on administration errors. For reducing the frequency of the latter errors, machine identification techniques such as