Ellrodt and colleagues' paper [1] highlights the difficulty physicians have with guidelines. However, "noncompliance" may not be an appropriate term to define physician response. Such findings as the high rate (42%) of misclassification and the lack of agreement among independent reviewers over medical appropriateness in records reviewed suggest differences in guideline interpretation.

Although Ellrodt and colleagues reported a similar distribution of physician mix between those who did and did not discharge appropriately, they did not indicate how these physicians compare with those in their respective specialties (for example, the percentages of compliant and noncompliant cardiologists). Were internists more or less compliant than cardiologists? Did the 13 noncompliant physicians differ from those who appropriately discharged the 151 patients? What types of physicians misclassified high-risk patients as having low risk, and the reverse?

Because comorbid conditions are a factor in delaying discharge, would multiple physician management of multiple illnesses contribute to preexisting inefficiencies of the health care system? Could this explain why more cardiologists were noncompliant? Research suggests that cardiologists view internists and family practitioners as lacking the knowledge and practice to treat acute myocardial infarction [2]. Is this also true for noncardiac conditions? Kassirer [3] has noted that there is little difference between generalists and specialists in terms of quality of care.

If perceptions of what constitutes appropriate care are specialty based, how effective is physician-to-physician feedback in changing physician behavior? Do physicians usually consult with those in their own specialty when determining a course of treatment for the diseases they routinely manage? If so, treatment uniformity may be possible, but would such uniformity cross specialty lines?

Finally, is it feasible to expect diagnosis-specific guidelines from a consensus of different physician specialties? If different specialties develop different guidelines for the management of the same disease, can positive outcomes result from different and equally efficacious treatments? If so, whom are guidelines for? More importantly, how much weight should guidelines have for assessing quality of care?