The committee is recommending a strategy for quality enhancement that relies on measurement and reporting of standardized performance measures across the government health care programs. Valid clinical performance measurement depends on the availability of clinical data (McGlynn and Brook, 2001).
Access to data remains problematic in a health care system that still depends largely on claims data, abstraction of data from paper records, and surveys to determine whether patients are receiving identified elements of care. The dependence on abstraction generally limits performance measurement to evaluation of entities with sufficient administrative infrastructure to develop the necessary data, such as hospitals, health plans, and large group practices, thereby excluding many small ambulatory care settings where a large proportion of care is delivered. Record abstraction is a labor-intensive process that usually occurs retrospectively rather than as an integral part of the clinical process, imposing a burden that prohibits more than intermittent review. While less costly than record abstraction, reliance on claims data may not provide the level of clinical detail required to track processes of care accurately (McIntyre et al., 2001; Schneider and Lieberman, 2001). For example, current claims data in many cases do not indicate whether complications in the course of hospitalization arose from preexisting comorbidities or adverse consequences of care. Moreover, claims data are available only for insured populations and are limited to billable services, thus constraining the aspects of care that can be evaluated.
Today’s data sources simply cannot support the strategy for quality enhancement proposed in this report. Indeed, there is broad consensus that the nation must develop a functional health care information technology infrastructure (Becher and Chassin, 2001; Eddy, 1998; Institute of Medicine, 2001; McGlynn and Brook, 2001; National Committee on Vital and Health Statistics, 2001; Schneider et al., 1999). Growing evidence supports the conclusion that automated clinical information and decision-support systems are critical to addressing the nation’s health care quality gap (Institute of Medicine, 2001). Computerized order entry and electronic medical records have been found to result in measurably improved care and better outcomes for patients (Bates et al., 1999; Birkmeyer et al., 2002; Webster, 2001). These results are particularly notable when electronic ordering triggers clinical decision-support information, for example, on antibiotic use (Christakis et al., 2001; Demakis et al., 2000; Rollman et al., 2001; Safran, 2001). Similar evidence suggests that these systems have the potential to reduce costs as well (Birkmeyer et al., 2002; Webster, 2001). In one study in which electronic order entry was accompanied by decision-support tools such as allergy and drug-interaction warnings, serious medication errors were demonstrated to decline by 86 percent (Bates et al.,