Is “Ol’ Reliable” Still Reliable?

“Plus ça change, plus c’est la même chose” (the more it changes, the more it stays the same). Jean-Baptiste Alphonse Karr —“The Wasps” 1849 Imagine the challenge of developing an index, preferably numerical, to enable uniform collection and tabulation of perioperative statistical data. In 1941, a committee of the American Society of Anesthetists, Inc. (the precursor to the American Society of Anesthesiologists, Inc. , founded in 1945) comprising Drs. Emery A. Rovenstine, Ivan B. Taylor, and Meyer Saklad were tasked with such a challenge. 1 The committee took painstaking care to explain that the system was not to be an assessment of operative risk. Rather the patient was to be graded “in relation to his physical state only. ”1 It was the expectation of the authors to correlate “the relationship between result, the operative procedure, and the patient’s preoperative condition. ” This set of principles was ultimately refined into a “physical state” document or more contemporaneously, the ASA physical status classification system. The classification method was never tested in the clinical setting. As incredible as it may seem, not one patient was ever entered into a database to test its validity. Thus, a group of knowledgeable pioneers in anesthesiology devised the most widely used index of preoperative status basically with a No. 2 lead pencil! Although the ASA physical status has undergone periodic revisions, it has stood the test of time. Its core features remain intact nearly 75 years after its introduction. As the classification system was adopted into clinical practice, its focus and application changed. Paradoxically, Visnjevac et al. 2 report in this issue that the scale that was specifically designed to exclude “operative risk”1 has been found to “better predict perioperative risk for both 30-day and long-term mortality after surgery. ” As a matter of fact, based on its utility and simplicity, the use of instruments similar to the ASA physical status classification system has migrated to other specialties and settings. For instance, the risk stratification index based on Medicare Provider Analysis and Review files3 is an index that relies entirely on administrative data such as International Classification of Disease codes. The National Surgical Quality Improvement Program of the American College of Surgeons uses the Healthcare Common Procedure Coding System to compare surgical outcomes among institutions for several surgical subspecialties. 4 Despite the presence of computers, availability of large databases, and the proliferation of large-scale data analysis, no other perioperative index has been developed that rivals the simplicity and uniform acceptance of ASA physical status classification. Regardless of incursions by other assessment indicators, ASA physical status compares very favorably with much more complex methodology. So the question is, given radical changes in health care management, reimbursement, and the functional impairments of an increasingly elderly population, do we need to change the “ol’ reliable” for some new, yet unvalidated modification of the original? Before debating the merits of introducing “an evaluation of patients’ preoperative functional capacity into the ASA classification for surgical patients, ”2 we need to enumerate the limitations and the strengths of the current classification system. First, the glaring limitation is that the ASA physical status classification is subjective: Interrater agreement was found to be only “fair. ”5–7 Of 10 hypothetical patients presented to 97 anesthesiologists, none had unanimous agreement on ASA physical status assessment. 6 In fact, the authors concluded that the variation among the assessments of anesthesiologists was so large, that “the ASA grade alone cannot be considered to satisfactorily describe the physical status of a patient. ”6 However, these assessments were performed on hypothetical case scenarios, so the reliability of the ASA physical status classification had to be evaluated in clinical practice on real patients. In such studies, the interrater agreement was found to be moderate, in both adult8–10 and pediatric11 patients. Over the years, the limitations became accepted, and the ASA physical status classification system started being used as the “backbone” for its application in other settings. For instance, a clinically useful and reliable classification system was used for determining optimal allocation of resources (expenditures), such as the cost of routine preoperative laboratory testing. The institutional costs of unindicated laboratory testing were lower for ASA physical status III patients (661) than for ASA physical status I patients (4953), although the percentage of unindicated tests (45%–96%) was still unacceptably high. 12 For decades, studies also have attempted to correlate the ASA physical status with perioperative risk and morbidity/mortality. 13, 14 Highly predictive risk adjustment models for 30-day mortality and morbidity have been developed to monitor quality of care and help patients (and third-party payers) interpret public reports of hospital outcomes. Such models have included Current Procedural Terminology code, age or hospitalization type (inpatient versus outpatient), and ASA physical status. 13 Finally, anesthetic risk is used by the Centers for Medicare in Canada, anesthesiologists receive financial premiums from the government health insurance plan to care for sicker, ASA physical status III or IV patients. 8 However, the incentive to upcode the patients’ ASA physical status classification to gain increased reimbursement previously has been shown not to be significant, as even practice areas that “have high medical liability premiums did not rate the patients higher. ”7 Almost 40 years later, and after the introduction of the Affordable Care Act, the physicians’ sensitivity to different reimbursement rates based on the ASA physical status classification will need to be revisited, particularly with the increased specificity of the International Classification of Diseases, 10th Revision that is purported to allow physician billing for more complex treatments in high-risk, sicker patients. Over the past decade, we have learned that preoperative physical and psychosocial interventions (“prehabilitation”) can result in improved postoperative functional exercise capacity, enhanced recovery, and reduced postoperative adverse events and complications. 15 As clinicians, we should, therefore, rejoice at the finding by Visnjevac et al. 2 that patients’ preoperative functional capacity is an independent predictor of postoperative mortality. As the authors elegantly describe in their analysis, postoperative mortality was significantly higher in patients who were either partially or fully dependent on others for the activities of daily living (ADLs) compared with patients who were independent. This increase in mortality was present in every ASA physical status class investigated. In addition, the rate of postoperative complications (myocardial infarction, cardiac arrest, postoperative pneumonia, urinary tract infections, wound dehiscence, renal insufficiency, reoperations, and hospital length of stay) was significantly higher in ADL-dependent versus independent patients. In fact, the effect of the patients’ preoperative functional capacity on outcome (mortality) was greater than that of the physical status and associated disease burden. In other words, ADL-dependent patients in any one ASA class had significantly higher mortality than ADL-independent patients in the next-higher ASA physical status class. This important and timely validation study2 that patients’ functional capacity is an independent predictor of mortality in each ASA class adds to and confirms that physiological reserve, whether measured as “functional capacity, ” “frailty, ”16 or simple gait speed (by preoperative gait speed test), 17 can be used to identify those patients who may require additional postoperative support, time, and resources. Were we able to do this consistently, we could better advise patients as to their expectations during the informed consent process and advanced directives. To their credit, the authors2 also describe in detail many of the study limitations. Functional status, much like the ASA physical status classification, is relatively subjective when assessed during the preoperative interview. The authors correctly point out that the functional status is not an all-encompassing predictor, as suggested by the limited increase in the receiver operating characteristic area under the curve. Thus, the authors call for further prospective research “prior to considering formal modification of the ASA physical status system. ”2 It seems that >7 decades after the introduction of the “simple” ASA physical status classification system, we can improve the prediction in the quality of care. This will take a significant amount of time, resources, and proof of improvement in the patients’ clinical outcome. Until then, “ol’ reliable” seems to remain the only game in town. DISCLOSURES Name: Sorin J. Brull, MD, FCARCSI (Hon). Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript. Attestation: Sorin J. Brull approved the final manuscript. Name: Paul G. Barash, MD. Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript. Attestation: Paul G. Barash approved the final manuscript. RECUSE NOTE Dr. Sorin J. Brull is the Section Editor of Patient Safety for the Journal. The manuscript was handled by Dr. Steven Shafer, Editor-in-Chief for the Journal, and Dr. Brull was not involved in any way with the editorial process or decision.