Mortality during a surgical procedure is so rare that it is hard to quantify risk. However, death within 30 days after surgery comprises approximately 7.7% of all causes of death per year (Lancet 2019;393:401). Stated another way, postoperative mortality is about 140 times higher than intraoperative mortality! Nearly half of all adverse events in hospitalized patients occur on the ward, under our direct care, and are responsible for nearly 85% of all postoperative mortality (Lancet 2012;380:1059-65; Resuscitation 2016;105:123-9; J Am Heart Assoc 2016;5:e003638). Code blue events are often thought of as sudden catastrophic cardiorespiratory deteriorations. However, contrary to this perception, most patients who suffered perioperative mortality had at least one abnormal vital sign within the four hours preceding cardiopulmonary arrest (code blue) (Resuscitation 2016;105:123-9). Further, there is an association of increasing mortality with an increasing number of abnormal or severely abnormal pre-arrest vital signs (Resuscitation 2016;105:123-9). Capturing trending changes in vital signs on the ward is therefore necessary if an attempt at early intervention to reverse a trajectory of harm is to be implemented.

Compact, portable monitoring is not foreign; portable pulse oximetry, adhesive cardiac arrhythmia monitors, continuous glucose monitoring devices, and even wireless breast and insulin infusion pumps are ubiquitous in both the medical setting and ambulatory community. Likewise, a number of wearable monitoring devices designed for the inpatient health care setting have been introduced to the market (BJA Open 2022;1:100002). However, validation and reliability remain a challenge along with alarm fatigue, which blunts any chance of demonstrating clinical effectiveness.

The ideal monitoring system should be continuous, reliable, automated, and noninvasive. It should also communicate with patient monitors, central nursing station platforms, and/or mobile devices with threshold-based alarms and titratable delays at the level of the device. Most of the important data from ward monitoring systems should easily stream into the electronic medical record (EMR) and should be extractable along with accurate time stamps (BJA Open 2022;1:100002). In a survey of nearly 6,000 anesthesiologists from western Europe and the United States, 91% of respondents believed that continuous monitoring of most vital signs, including oxygen saturation, heart rate, and blood pressure, is indicated on surgical wards (BJA Open 2022;1:100002).

While the intensive care unit (ICU) usually has low patient-to-nurse ratios and continuous or, at minimum, hourly vital signs, patients admitted to hospital ward beds are cared for by nurses with a much higher patient-to-nurse ratio and only intermittent monitoring; this usually means that vital signs are only spot-checked every four to 12 hours depending on local culture or hospital policy (BJA Open 2022;1:100002). Further, subjective, intermittent manual measurements of some vital signs and automatically measured oxygen saturations and noninvasive blood pressures are prone to artifact and inaccuracy (BMJ Qual Saf 2017;26:832-36). Respiratory rate is an independent predictor of poor outcomes and thus incorporated into many risk-prediction indices (BMJ Qual Saf 2017;26:832-36). Whereas most other vital signs are reported based on objective measurements, the respiratory rate is usually a clinical assessment with observation and counting the patient’s breathing pattern and is frequently inaccurate (BMJ Qual Saf 2017;26:832-36).

Despite these recognized disparities, no standard of care for ward monitoring has been established in the U.S. or other parts of the world. This means that many hemodynamic and respiratory disturbances are frequently missed, and interventions on the ward are usually reactive rather than preventative, as is often seen in the ICU setting (Anesthesiology 2019;130:550-9; Br J Anaesth 2021;127:760-68; Anesth Analg 2015;121:709-15; J Clin Anesth 2023;89:111159). Importantly, most patients (75%) who experienced in-hospital mortality were not admitted to an ICU at any stage of the postoperative period; thus, the vast majority of patients who die are potentially insufficiently monitored (Lancet 2012;380:1059-65; Resuscitation 2016;105:123-9; J Am Heart Assoc 2016;5:e003638).

The POISE and POISE-2 trials demonstrated a high incidence of prolonged, serious hypotension and tachycardia in ward patients, which has been reconfirmed in some observational studies over a decade later (Lancet 2008;371:1839-47; N Engl J Med 2014;370:1504-13; Anesthesiology 2019;130:550-9; J Clin Anesth 2023;89:111159). Importantly, this persists even when threshold-based alarms are activated in an unblinded fashion to enable recognition and intervention by staff; however, the duration and severity of vital sign abnormalities appeared to be less than others where blinded ward monitoring was tested (Anesthesiology 2019;130:550-9; J Clin Anesth 2023;89:111159). A post-hoc analysis of POISE-2 demonstrated that postoperative hypotension is associated with an approximately three-fold increase in odds of mortality when occurring on the ward rather than intraoperatively or in the postanesthesia care unit (Anesthesiology 2018;128:317-27). Nonetheless, vitals obtained with only intermittent spot checks miss at least half of all episodes of hypotension with a MAP less than 65 mmHg and significant amounts of hypertension and heart rate changes as well (Anesthesiology 2019;130:550-9; J Clin Anesth 2023;89:111159).

Postoperative respiratory depression, particularly opioid-induced, often presents with both hypoxemia and hypoventilation and is frequently underrecognized (J Healthc Qual 2022;44:e7-14). A pivotal study of continuous blinded ward oximetry and capnography monitoring demonstrated a 46% incidence of opioid-induced respiratory depression when assessed with predefined threshold criteria and adjudicated for artifacts (Knee Surg Sports Traumatol Arthrosc 2023;31:2917-26). Acute respiratory events on the wards are associated with an in-hospital mortality of nearly 40% (Circulation 2013;127:1538-63). Malpractice claims related to this have been deemed preventable in closed claims analysis, with insufficient monitoring being a contributing factor (Anesthesiology 2015;122:659-65). In fact, among one observational study of postsurgical patients, approximately 20% of patients experience a pulse oximetry reading less than 90% for more than 10 minutes each hour, with almost all of these missed during routine, intermittent nursing checks (Br J Anaesth 2021;127:760-8). Despite this, continuous pulse oximetry has produced conflicting data, with one randomized trial demonstrating no difference compared to standard of care (Anesth Analg 2006;102:868-75). However, observational literature has shown shorter length of stay, a decreased incidence of rescue events, and fewer ICU transfers when it is deployed; meta-analyses have confirmed the same (Anaesthesia 2006;61:1031-9; Anesthesiology 2010;112:282-7; Am J Med 2014;127:226-32; Anesth Analg 2017;125:2019-29). Considering more awareness around the utility of continuous monitoring in the last five years, an updated systematic review and meta-analysis of 23 studies (17 with continuous pulse oximetry alone and five with capnography) and about 56,000 adult hospital ward patients (only eight studies had a comparator intermittent monitoring group) determined that continuous pulse oximetry was better at recognizing desaturation (SpO2 <90%) (OR: 11.94 (95% CI: 6.85, 20.82; P<0.01) compared to standard monitoring. Importantly, no significant differences were reported for ICU transfer, reintubation, and noninvasive ventilation between the two groups (J Clin Anesth 2024;94:111374).

Continuous ward monitoring, preferably via wearable wireless systems, does reduce the burden of vital signs perturbations and has the potential to improve perioperative patient safety. However, most available evidence is observational or based on retrospective datasets. These analyses demonstrate that ward monitoring detects far more vital sign changes than traditional intermittent spot-check monitoring. Other data also report a significant reduction in ICU transfers, rapid response calls, and, in some cases, mortality; however, these are based on historical controls in comparison to current interventional cohorts. Before-and-after study designs have inherent statistical flaws that are difficult to overcome. Much-needed evidence for continuous ward monitoring would be an appropriately powered, pragmatic, interventional trial with patient-centric clinical outcomes.

When implemented, wireless monitoring systems do improve event detection and patient outcomes. Taenzer et al. performed a before-and-after concurrence study in an orthopedic ward where a continuous pulse oximetry surveillance system was linked to a nurse notification system. The system was associated with a reduction in the rate of ICU transfers and an over 50% reduction in rescue events (Anesthesiology 2010;112:282-7). Similarly, a continuous monitoring system of heart and respiratory rates in a medical-surgical unit was associated with lower rates of cardiac arrest (6.3 before to 0.9 per 1000 patients post-implementation) (Am J Med 2014;127:226-32). In a recent study of 2,303 patients, continuous monitoring with wearable monitoring devices linked wirelessly to hospital systems reduced unplanned ICU admissions and rapid response team calls (Br J Anaesth 2022;128:857-63). Khanna and colleagues performed an analysis on nearly 35,000 patients recovering on surgical wards that compared continuous monitoring with contemporaneous controls with intermittent monitoring. After propensity matching, intermittent monitoring (n=12,345) was associated with increased risk of a composite of mortality or ICU admission (OR 3.42, 95% CI 3.19-3.67; P<0.001) and heart failure (OR 1.48, 95% CI 1.21-1.81; P<0.001), myocardial infarction (OR 3.87, 95% CI 2.71-5.71; P<0.001), and acute kidney injury (OR 1.32, 95% CI 1.09-1.57; P<0.001) compared with continuous wireless monitoring (n=7,955). Interestingly, there was no difference in the odds of a rapid response team activation in the two monitoring groups (Br J Anaesth 2024;132:519-27).

While desirable and with clear benefit, adoption of wireless continuous monitoring systems designed for ward use has been slow and fraught with challenges (Br J Anaesth 2021;127:675-7). Implementation challenges are primarily cost-related and technical, with connectivity issues and artifact generation being of primary concern (BJA Open 2022;1:100002; Crit Care 2019;23:194). Ultimately, there is evidence to support an overall cost-savings with implementation of continuous ward monitoring, with one study suggesting that just a 1.5% reduction in episodes of respiratory depression would achieve a break-even point that would negate startup costs (Knee Surg Sports Traumatol Arthrosc 2023;31:2917-26; Best Pract Res Clin Anaesthesiol 2019;33:229-45). Nonetheless, the initial investment to establish such a monitoring system is likely to deter many institutions from rapid and widespread adoption of the technology. Finally, given the dependence on theoretically vulnerable wireless networking, there are potential concerns regarding patient privacy rights and data security. Ultimately, the startup costs for a continuous monitoring system also must factor in the “efferent arm,” i.e., the cost of effective communication and intervention systems as well providing hospital IT enough resources to handle the data burden.

Further, false alarms are known to lead to alarm fatigue and neglect, and the sheer amount of data and thus potential perturbations to be recognized with universal ward monitoring is staggering. Given this, there is debate as to which patients should require continuous ward monitoring. While it is tempting to monitor all patients, there is certainly a population of patients who are unlikely to suffer clinically significant harm with brief vital sign abnormalities. As such, implementation of risk prediction based on age, cardiovascular risk factors, and the newly described Risk Stratification Index 3.0 may help to select those most at risk and thus poised to benefit (Knee Surg Sports Traumatol Arthrosc 2023;31:2917-26; JAMA 2017;317:1642-51; Anesthesiology 2022;137:673-86). In terms of monitoring parameters, variables like pulse oximetry, electrocardiography, and respiratory rate are easy to acquire with existing continuous ward monitors, portable blood pressure monitoring seems fraught with inaccuracies, and minimally invasive advanced portable hemodynamic monitoring is in its early stages of validation. Most importantly, better monitoring alone is not likely to improve outcomes without effective early warning systems and protocolized bundles of interventions, i.e., the “efferent arm,” so ward nurses and providers know what to do in response to alarms. A key future direction for optimizing ward monitoring options includes introduction of pattern detection technology as well as machine learning and artificial intelligence using waveform data.

In summary, continuous ward monitoring poses a unique opportunity to improve postoperative patient outcomes with observational data supporting beneficial effects on clinical outcomes. However, appropriately powered and preferably pragmatic prospective outcomes trials need to be performed to pave the way for implementation and universal adoption at the bedside.