AG Gordon, VA Pittsburgh Health Care System; DS Macpherson, VA Pittsburgh Health Care System

Objectives: Clinical practice guidelines are increasingly promoted and utilized in today’s health care system. The assessment of cardiac risk prior to surgery is critical, as perioperative complications due to cardiac events may be deadly. Two guidelines for preoperative cardiac assessment for non-cardiac surgery (American College of Cardiology/American Heart Association (AHA) and the American College of Physicians (ACP)) use different approaches to assess risk of surgery. We sought to learn if these two guidelines lead to important differences in recommendations for a group of patients who attended a Medical Pre-operative Evaluation Clinic (MPEC).

Methods: We reviewed computerized records of all patients evaluated at MPEC in a Veterans Hospital between January 1 and April 1, 1998. Using data abstracted from these notes, we determined what testing should have been done if each guideline was followed and what testing was actually ordered. Inter-rater reliability of data abstraction regarding key variables was determined by a second author. We simplified guideline recommendations to 1) operation without testing (OR), 2) non-invasive stress testing (NIT), 3) cardiac catheterization (CATH), and 4) OTHER (cancel or delay surgery). To compare the guidelines recommendations to each other and practitioner recommendations, we used Cohen’s Kappa (K), weighted Kappa (WK), and Marginal Homogeneity (MH) tests.

Results: We reviewed 138 MPEC notes. Reliability testing of data abstraction revealed 100% agreement. Patients were older (median age 69), male (97%), and had histories of current angina (26%), bypass surgery (15%), and shortness of breath (32%). Patients were undergoing primarily orthopedic (36%), head and neck (20%), neurological (15%), and abdominal (11%) surgeries.

Recommendations were discordant between guidelines for 17% of patients (K = 0.382, (CI: 0.19, 0.57), WK = 0.54 (CI: 0.33,0.75)). Guide-lines never agreed on the need for NIT. That is, in each patient in whom the AHA guideline recommended a non-invasive test, the ACP guideline did not, and visa-versa. Extreme differences in recommendations (i.e., one recommends OR, the other CATH) occurred in 9 (7%) patients. For patients with a history of angina or infarction, the guidelines were in moderate disagreement (K = 0.44, CI: 0.22, 0.66).

MPEC physicians recommended 102 patients go to the OR, 3 for CATH, and 6 for OTHER. They ordered NIT more often (n = 27) than either guideline. Recommendations of physicians were more concordant to the AHA (K = 0.33) than the ACP (K = 0.21). For patients in whom the physician order a NIT, the guidelines recommendations agreed poorly (K = 0.26). Reversible ischemia was discovered in 31% of these patients.

Conclusions: When applied to real patients being evaluated for surgery, AHA and ACP guidelines significantly differ in recommendations for further testing prior to surgery. In patients where practitioners assessed a high likelihood of cardiac risk, guidelines agreed less often.

Impact: Clinical practice guidelines are useful tools to aid in clinical decision making, but different guidelines constructed by different groups can differ in recommendations for the same clinical problem. Guideline makers should be cognizant of competing guidelines and work together for greater consensus of their recommendations.