Quality in medicine is a peculiar thing. It is clearly apparent, and yet, can be very difficult to measure and quantify. Surgery, a performance art of sorts, can be even more challenging to qualify or rate. However, as a means to elevate the quality of care for all patients, hospital systems and care providers have aggressively made attempts to do so. This is a noble objective.
In September 2018, the Committee of Gynecologic Practice of the American College of Obstetricians and Gynecologists released ACOG Committee Opinion Number 750, titled, “Perioperative Pathways: Enhanced Recovery After Surgery.”1
The goals of this committee opinion were to advocate for gynecologic surgeons using the “ERAS” pathways in their perioperative care as part of an evidenced-based approach to quality improvement. ERAS pathways have been previously discussed in this column and feature bundled perioperative pathways that incorporate various concepts such as avoidance of prolonged preoperative fasting, early postoperative feeding, multimodal analgesia (with an avoidance of opiates), and inclusion of antibiotic and antiembolic prophylaxis, among other elements.
What was alarming upon closer review of this ACOG Committee Opinion was its omission of the randomized controlled trial by Dickson et al., the only randomized trial published in gynecologic surgery evaluating ERAS pathways.2 This trial compared the length of stay for patients receiving laparotomy for gynecologic cancer surgery who received perioperative care according the ERAS pathway versus those who received standard perioperative care. They found no difference in length of stay – the primary outcome – between the two groups, an impressive 3 days for both. The secondary outcome of postoperative pain was improved for the ERAS group for some of the time points. It was likely that the excellent outcomes in both groups resulted from a Hawthorne effect in which the behavior of study participants is influenced by the fact that they were being observed, in addition to the fact that the physicians involved in the study already were practicing high quality care as part of their “standard” regimen. It simply may be that the act of trying to improve quality is what improves outcomes, not a specific pathway. As senior author, Dr. Peter A. Argenta, explained to me, many of the ERAS elements are “simply good medicine.”
ERAS pathways are an example of process measures of quality. They include elements of care or processes in the delivery of care that are thought to be associated with improved outcomes. Prescription of antibiotics or venous thromboembolism (VTE) prophylaxis are other examples of process measures thought to be associated with improved surgical quality. Rather than rating surgeons’ outcomes (surgical site infection), surgeons are rated on their compliance with a process (the rate of appropriate perioperative antibiotic prescription). However, high compliance with these processes is not automatically associated with improved observed outcomes. For example, hospitals that meet the definition of high quality by virtue of structural measures (such as procedural volume and use of hospital-level quality initiatives) are associated with worse risk-adjusted VTE rates despite demonstrating higher adherence to VTE prophylaxis.3 This is felt to be a function of surveillance bias and the fact that these same hospitals have better capabilities to capture events as part of a feedback mechanism built into their quality initiatives.
What ERAS has favorably done for surgical care is to shine a glaring light on and challenge the unnecessary, old-fashioned, and non–patient-centric interventions that were considered dogma by many. For example, minimizing preoperative fasting is most certainly a patient-friendly adjustment that should absolutely be embraced, regardless of whether or not it speeds up time to discharge. Multimodal approaches to analgesia consistently have been shown to preserve or improve postoperative pain levels with a focus on minimizing opiate use, once again a noble and patient-centered objective.
However, all too many surgical quality interventions focus on their ability to reduce postoperative length of stay. Length of stay is an important driver of health care cost, and an indirect measure of perioperative complications; however, it is not a patient-centered outcome. So long as patients recover from their surgery quickly with respect to pain and function, the location of that recovery (home versus hospital) is less of a focus for most patients. In addition, in the pursuit of shorter hospital stays and less perioperative morbidity, we may encourage practices with unintentional adverse patient-centered outcomes. For example, to preserve a surgeon’s quality metrics, patients who are at high risk for complications may not be offered surgery at all. Long-term ovarian cancer outcomes, such as survival, can be negatively impacted when surgeons opt for less morbid, less radical surgical approaches which have favorable short-term morbidity such as surgical complications and readmissions.4
Ultimately we are most likely to see improvement in quality with a complex, nuanced approach to metrics, not simplistic interventions or pathways. We should recognize interventions that are consistently associated with better outcomes such as high procedural volume, consolidating less common procedures to fewer surgeons, data ascertainment, and reporting data to surgeons.5 Physicians need to take ownership and involvement in the quality metrics that are created to assess the care we provide. Hospital administrators may not fully understand the confounders, such as comorbidities, that contribute to outcomes, which can lead to mischaracterization, cause unfair comparisons between surgeons, or create unintentional incentives that are not patient-centered.6
We all need to understand the epidemiologic science behind evidence-based medicine and to be sophisticated in our ability to review and appraise data so that we can be sensible in what interventions we promote as supported by good evidence. If we fail to correctly identify and characterize what is truly good quality, if we miss the point of what is driving outcomes, or overstate the value of certain interventions, we miss the opportunity to intervene in ways that actually do make a meaningful difference.
Dr. Rossi is assistant professor in the division of gynecologic oncology at the University of North Carolina at Chapel Hill. She said she had no conflicts of interest. Email Dr. Rossi at.