Why Are Patients Still Dying after Surgery? The Problem of Postoperative Deterioration

Mortality related solely to anesthesia has declined precipitously since Beecher and Todd's landmark study in 1954, which found a staggering mortality rate of one death per 1,560 anesthetics (Ann Surg 1954;140:2-35). More recent investigations have demonstrated a mortality rate of one death per 45,000 anesthetics, a 28-fold decrease (Anesth Analg 2018;127:730-5). Yet, despite the improvements in anesthetic-related mortality, patients continue to suffer harm after perioperative care. A recent study found that nearly a quarter of patients admitted for inpatient hospital care suffered from an adverse event, and events related to a surgery or procedure were the second most common type, after medication-related events (N Engl J Med 2023;388:142-53). A follow-up study found that at least 35% of perioperative adverse events were preventable or partially preventable. The most common setting for adverse events was the general care unit, not the ICU or OR (BMJ 2024;387:e080480). Postoperative deaths are now the third leading cause of mortality worldwide (Lancet 2019;393:401). Postoperative deterioration therefore represents the next great frontier to be conquered in anesthesia patient safety. As Daniel Sessler, MD, stated in his 2023 Emery A. Rovenstine Memorial Lecture, anesthesiologists should take the lead in immediate postoperative care, which represents the “fourth branch of anesthesia,” after surgical anesthesia, pain medicine, and critical care (Anesthesiology 2024;140:1068-75). Innovations in both technology and systems are needed to fill this critical patient safety gap.The Anesthesia Patient Safety Foundation (APSF) addressed the issue of postoperative deterioration at the APSF Stoelting Conference 2019 (Anesth Analg 2020;131:e155-9). A review of the ASA Closed Claims Database found that “failure-to-rescue” events occurred in one in five claims for severe permanent injury or death, and communication and teamwork were significant risk factors for injury. A separate review of the Anesthesia Incident Reporting System found the majority of failure-to-rescue cases were preventable, mostly due to failure to recognize deterioration, failure to monitor, delay in escalation, and delay in making the definitive diagnosis. The prevalence of patient harm due to postoperative deterioration is not necessarily surprising. Patients move from the intense, 1:1, minute-by-minute monitoring of the OR to the PACU, with low nursing ratios and anesthesiologist oversight, to the general floor where nursing care may be 1:4 with intermittent monitoring. Continuous monitoring: Panacea or problematic? Hemodynamic derangements often precede ICU admission, which makes continuous monitoring a commonly sought-after intervention to prevent postoperative deterioration (Resuscitation 2004;62:275-82). A survey of anesthesiologists across Western Europe and the United States found that the majority of respondents were in favor of wireless continuous monitoring systems on the surgical floors, and that such systems were thought to decrease both clinical deterioration and mortality (BJA Open 2022;1:100002). However, the evidence on the benefits of continuous monitoring is mixed; while some studies have shown reduced ICU admissions and rapid response events, others have shown no difference (J Clin Monit Comput 2018;32:945-51; Int J Clin Pract 2016;70:806-24). Experts at the APSF Stoelting Conference 2019 noted that more monitoring may not necessarily be the answer; rather, interventions should focus on risk stratification, criteria-based monitoring, and early warning systems (Anesth Analg 2020;131:e155-9). Others have stated that it is not the monitoring itself that changes outcomes, but rather the response to such derangements; systems must be in place to provide early recognition and robust response (BJA Open 2022;2:100009). At this time, the Society of Critical Care Medicine makes no recommendation regarding continuous monitoring for patients outside of the ICU (Crit Care Med 2024;52:314-30). Early warning systems (EWS), especially in conjunction with machine learning models, have shown promise in decreasing clinical deterioration in the general ward (BMJ 2020;369:m1501; eClinicalMedicine 2023;66:102312). In the postoperative setting, a study of an early warning system based on vital sign abnormalities with automated clinician notification found decreases in the rates of cardiac arrest and ICU admission (Ann Surg 2020;271:100-5). However, these systems also come with their own challenges. Such systems cannot replace human clinical assessment and may be difficult to implement from a logistic and economic standpoint. Furthermore, such systems may be plagued with false alarms that might decrease responses to deterioration events; in one retrospective study, false alarms for a perioperative early warning system numbered up to 70 per patient per day (JMIR Perioper Med 2023;6:e44483). Future directions As with many other perioperative safety challenges, clinical deterioration could be improved by system-level interventions focused on safety culture, teamwork, and communication. Factors such as inadequate transfer of information, lack of psychological safety, and hierarchical professional relationships have been found to contribute to adverse events related to clinical deterioration (J Clin Nurs 2023;32:6165-78). Standardized criteria for a rapid response activation, as well as established protocols for how the response should be conducted, can allow providers from multiple disciplines to have a shared mental model of expectations and can help facilitate teamwork. Hospital and departmental leadership should model transparency, honesty, and psychological safety, thereby enabling a robust and resilient safety culture that is constantly learning and iterating toward improvement. Given the current need for more definitive evidence on early warning systems and technological innovations in the setting of postoperative deterioration, future research should be aimed at which ways postoperative monitoring can be optimized to prevent deterioration. For example, perhaps it is not continuous monitoring that will make a difference, but intermittent monitoring intentionally tailored to individual patient and surgical factors. In addition, artificial intelligence has tremendous potential to help identify patients at risk, and further investigations should look into clinical decision support algorithms and warning models that can help prevent deterioration. Anesthesiologists have not traditionally stepped outside of the OR or PACU to care for the deteriorating patient. However, the patient safety gains made in the OR have not necessarily transferred to the general ward, and postoperative patients are still suffering from immense preventable harm. Anesthesiologists have a unique skill set that makes them particularly attuned to physiologic changes that may normally be missed on the floor, and they also possess expertise in the potential cardiovascular and respiratory implications. It is time that anesthesiologists truly embrace the “perioperative medicine” piece that is in the title of departments nationwide; such a title encompasses not just preoperative optimization and intraoperative management, but postoperative care as well.Emily Methangkool, MD, MPH, Member, Anesthesia Patient Safety Foundation Board of Directors, Chair, ASA Committee on Patient Safety and Education, and Chair, Department of Anesthesiology, Olive View-UCLA Medical Center, Sylmar, California.