Background: Current guidelines recommend preoperative stress testing for patients whose predicted risk of a major adverse cardiac event exceeds 1%, whose functional status is poor or unknown, and when stress testing would change clinical management. However, each of those three decision points likely vary by provider, and real-world use is unclear. In order to understand contemporary use and drivers of preoperative cardiac stress testing, we set out to explain variation in preoperative stress testing using rich clinical data from a large integrated health system.
Methods: The Internal Medicine Preoperative Assessment, Consultation and Treatment (IMPACT) Center assesses patients prior to noncardiac surgery at the Cleveland Clinic. Between 1/1/2008 and 12/31/2018, approximately 118,552 patients were seen in that clinic by 104 providers across 159,795 visits.Using a series of multivariable multilevel logistic regression models, we tested variables plausibly linked to stress test ordering and completion rates, either from theory or from previously-published associations. We tested measures of perioperative risk (RCRI, MICA, and surgical categorizations), measures of functional status (METs, functional class), measures of social and financial support (race, marital status, neighborhood deprivation index), medical comorbidities (age, BMI, numerous diagnoses), measures of physician tendencies and experience (years of previous experience, number of patients seen previously in the IMPACT clinic), and date (continuous and dichotomized at the most recent guideline’s publication date).We used multiple imputation to address missing data and used multilevel models clustering visits by either physician or by patient.
Results: Of 159,795 visits to the IMPACT clinic, 8,300 (5.2%) resulted in a referral for cardiac stress testing, 8,085 (5.1%) of whom completed the test before surgery or within 30 days.Key patient factors associated with preoperative stress testing included male sex, a diagnosis of ischemic heart disease, estimated surgical cardiac risk greater than 1%, functional class, and type of surgery. Type of surgery under consideration had the largest effect: a patient scheduled for aortic surgery had an estimated marginal stress testing rate of approximately 19%, and a patient undergoing peripheral vascular surgery has an estimated marginal rate of approximately 6.5%. A patient with an existing diagnosis of ischemic heart disease was approximately 1.9 times more likely to be referred for preoperative stress testing compared to one without such a diagnosis, a patient whose predicted surgical risk exceeded 1% was 1.4 times more likely than a patient below that threshold, and a male was 1.3 times more likely than a female. Even after fully adjusting for patient factors, the number of patients the physician had seen previously was a significant and meaningful predictor of preoperative stress testing: a visit with a provider who had seen 1,000 patients in the IMPACT clinic was 1.62 times more likely to result in a preoperative stress test compared to a visit with a physician who had seen 5,000 previous patients previously.
Conclusions: In this large cohort of patients seen for preoperative risk assessment at a single health system, key drivers of preoperative stress test usage included male sex, a diagnosis of ischemic heart disease, estimated surgical risk greater than 1%, functional class, type of surgery (particularly vascular), and previous physician experience in a preoperative clinic.