See Article, page 951

In this Pro-Con commentary article, we discuss continuous physiologic monitoring used to detect respiratory decompensation in patients after surgery. Acute respiratory events are the focus of the debate because they cause critical illness and contribute to substantial postoperative morbidity and mortality.

For postoperative surgical inpatients, acute respiratory complications are common, estimated to occur at an incidence of 1% to 23%.1,2 In the subgroup of patients with complex comorbid conditions (American Society of Anesthesiologists [ASA] physical status >3) the postoperative respiratory complication rate is 30% and the 7-day mortality is significantly increased.3 In many of these situations, failure to rescue is associated with failure to detect substantial changes in respiratory rate in the 3 days before a patient had to be transferred to the intensive care unit (ICU) for emergency care.4 Recent literature indicates that vital sign abnormalities often occur while patients are on inpatient units, 1 to 4 hours before a cardiopulmonary event occurs.5

The physiology of postoperative respiratory dysfunction is well known and can be caused by residual anesthetic and neuromuscular blocking agents, analgesic regimens, combinations of sedative agents, and patient positioning.6 Unsurprisingly, postoperative hypoxemia is common. A prospective study found that 21% of patients had oxygen saturation (Spo2) levels less than 90% for at least 10 minutes after surgery.7 Furthermore, 3% of patients had an Spo2 less than 80% for at least 30 minutes. Contemporary postoperative analgesia regimens are often multimodal, and coadministration of sedative agents and opioids has an additive effect on risk of opioid-induced respiratory depression (OIRD).8 A systematic review and meta-analysis found the incidence of OIRD to be 5 cases per 1000 anesthetics administered.9 A closed-claim analysis by the ASA assessed 92 critical events related to postoperative OIRD and showed the majority occurred within the first 24 postoperative hours, most were preventable with appropriate monitoring, and 42% had a nursing check within the 2 preceding hours.10

Obstructive sleep apnea (OSA) is another important risk factor for postoperative respiratory complications and is associated with a 2-fold increase in risk.11,12 OSA has spawned increasing numbers of malpractice suits, which are generally related to respiratory complications, with use of opioids having a role in 38%.13

Although not all patients who have cardiopulmonary arrests can be saved, virtually all opioid-induced respiratory arrests are treatable and should not be fatal for hospitalized patients after surgery. The Anesthesia Patient Safety Foundation (APSF) has stated, “No patient shall be harmed by OIRD in the postoperative period.”13 To prevent clinically significant OIRD, the guidelines from the APSF and the American Society for Pain Management Nursing recommend that all patients given postoperative opioids should be continuously monitored.14,15 A 10-year retrospective study showed that continuous pulse oximetry used to monitor patients for severe events could virtually eliminate all OIRD-related deaths.16 Unfortunately, the standards recommended by the APSF are not adhered to in many organizations.

This pro-con commentary will address the application of universal monitoring for all patients for their entire postoperative hospitalization versus selective monitoring of a subset of patients for a portion of their hospitalization. The authors’ pro-con positions focus on key characteristics of physiologic monitoring system design and configuration:

Who should be monitored When monitoring should be in place Where monitoring is appropriate How monitoring should be configured

Various constraints and considerations such as clinical resources, alarm fatigue, patient and staff acceptance, risk-assessment algorithms, and respiratory event detection are discussed.

PRO POSITION: UNIVERSAL CONTINUOUS RESPIRATORY MONITORING OF PATIENTS AFTER SURGERY

All patients should receive continuous respiratory monitoring postoperatively for their entire hospitalization. Herein, we (G.B. and S.P.M.) provide justification that

Every hospitalized postoperative patient is at risk for acute severe respiratory events that may progress to cardiopulmonary arrest and death without immediate treatment; and Unmonitored cardiopulmonary arrests are more lethal because of delays in detection and treatment.

Therefore, logic dictates that every patient hospitalized after surgery should be continuously monitored for possible respiratory depression events. We refute the most common criticisms of universal continuous monitoring, that is, effectiveness and feasibility. We provide evidence that continuous monitoring is proven and not limited by unmanageable false alarms. We also provide evidence that such systems are needed, affordable, and acceptable to patients and staff.

Every Hospitalized Postoperative Patient is at Risk for Acute Severe Respiratory Events That May Progress to Cardiopulmonary Arrest and Death Without Immediate Treatment

Surgery and anesthesia are associated with multiple risk factors for respiratory events. The need for postoperative hospitalization is itself predictive of risk for respiratory events because patients at low risk for respiratory complications are usually outpatients. Approximately 60% of operations in the United States are now ambulatory procedures with unmonitored recovery for patients at home.17,18 To approve patients for outpatient surgery, most organizations use formal criteria that include consideration of comorbid conditions, need for potent postoperative analgesia, and type of postoperative care.19,20 Of note, these systems are subject to strict regulatory oversight and quality assurance metrics (eg, outpatient surgery readmission rates).

The 14.4 million patients in the United States requiring postoperative hospitalization are at risk for acute severe respiratory events, regardless of inpatient care setting. In acute care hospitals, the overall acuity level has increased, and more than 50% of postoperative patients receive potent analgesics.21,22 The calculous to determine the level of postoperative care is complex and reliant on patient, perioperative, and postanesthesia care unit (PACU) factors. Patient care–level assessments are performed in hospitals daily to determine the lowest required acuity setting needed (including discharge) to contain limited hospital resources. Despite these assessments, patients considered lower acuity still have adverse events and require rescue and emergent escalation of care. Errors in assessing risk with undertriage to general care units have been associated with increased risk of morbidity and mortality.23 A study of rescue events (rapid response team [RRT] and code blue activations) for 67,000 hospitalized adults showed intermediate care units had the highest rate of RRT activations at 2.8 per 1000 patient days, while general surgical care units had a rate of 1.05 per 1000 patient days.24 Notably, the general care rescue event rate did not decrease from postoperative day 1 to day 28. Similarly, cardiopulmonary arrests in the postoperative group occurred during the entire hospitalization and were associated with a 21.0% mortality rate. Another multicenter study reviewed 402,023 RRT activations from 347,401 unique health care encounters.25 Most indications were respiratory (38.0%) and cardiac (37.4%), and the most common interventions were application of pulse oximetry (66.5%) or other monitoring (59.6%) and supplemental oxygen (62.0%) systems.25

These data support that (1) surgery and (2) the need for continued hospitalization are predictors of risk for serious but treatable deteriorating conditions, even for patients deemed acceptable for admission to a general care unit. Although respiratory depression can progress slowly, a substantial proportion of cases are acute. Between 3.6% and 10.0% of RRT activations reported in the literature are associated with intubation.25–27

Unmonitored Cardiopulmonary Arrests Are More Lethal

National registry data have shown that 8.0% of all cardiopulmonary arrests are unwitnessed/unmonitored and are associated with a more than 2-fold higher in-hospital mortality rate.28 Because of this association, a goal of the National Registry of Cardiopulmonary Resuscitation’s Quality Assurance program has been to eliminate unmonitored cardiac arrests in hospitalized patients.29 A recent meta-analysis has also verified that the mortality rate decreases when cardiac arrests are witnessed.30 Why? During a cardiopulmonary arrest, every minute of delay in initiating resuscitation decreases the survival rate for patients with rescuable conditions.31 Failure to rescue (defined as mortality per 100 serious but treatable complications)32 correlates with nursing ratios, presumably because lower ratios improve the chance for early detection and intervention, which suggests that assessing vital signs every 4 to 8 hours is inadequate.3 These data have compelled notable surgical outcome leaders to call for providing continuous surveillance monitoring in general care units for all patients after surgery, although the leaders do not explicitly advocate for monitoring during the entire hospitalization.33

Continuous pulse oximetry is one of several options that has been used for postoperative monitoring because it measures both oxygenation and pulse rate, allowing for early detection of both severe acute respiratory and cardiac events. Our institution was one of the first to establish this standard of care for all postoperative general care beds over 15 years ago, with the initial goal to eliminate unwitnessed or unmonitored acute cardiopulmonary events. Resuscitation outcomes improved after this change was made.34

Therefore, Every Postoperative Patient Should Be Continuously Monitored While Hospitalized

Early detection of cardiopulmonary arrest and often the precursor conditions create the opportunity to rescue a subset of hospitalized patients with serious but treatable conditions, especially respiratory events—our core argument. The premise that undetected and untreated physiologic derangements are more likely to progress to respiratory decompensation and cause patient harm has strong validity. This situation provided the rationale for the anesthesiology community to adopt minimum monitoring standards for patients in the operating room.35,36 A survey of 5744 anesthesiologists showed that most (91%) supported continuous monitoring of vital signs for all patients being treated postoperatively in general care units.37

Next, we address the most common categories of criticism of universal continuous monitoring: effectiveness and feasibility.

Continuous Monitoring is Effective in Allowing Early Detection of Serious But Treatable Respiratory Events

A seminal before and after cohort-controlled study showed that continuous pulse oximetry monitoring was associated with a 65% reduction in RRT activations and a 48% decrease in ICU transfers.38 Multiple subsequent studies investigating a variety of continuous monitoring modalities have also found decreases in RRTs and ICU transfers, and in 1 study decreases in cardiac arrests and in-hospital mortality were also noted.39–44 Meta-analyses have shown similar results, including reductions in ICU transfers but not in mortality.45,46 However, most of these studies lacked detailed information about the larger rescue system infrastructure and about how monitoring is operationalized and built into clinical workflows. Deploying continuous monitoring technology alone is insufficient for establishing a reliable organizational rescue capability. A mixed-methods engineering assessment of the complex adaptive sociotechnical system needed to mitigate failure to rescue events provides insight into the full range of factors (human, team, system, organizational) that impact rescue system performance.47

Early warning scores are often suggested as an alternative method for identifying higher-risk patients needing more intensive monitoring. None of these scores are perfect predictors of risk and thus leave many patients whose conditions will deteriorate unmonitored.48–50 Early warning scores assess the patient’s present state but are currently calculated only intermittently during a patient’s stay and therefore cannot provide the same detection capability for acute events that occur over minutes or hours.51 Although the evidence supporting the broad use of continuous monitoring is substantial, opportunities remain to build on this body of evidence.

False Alarms With Continuous Monitoring Are Manageable and Minimal

The alarm problem associated with continuous monitoring systems can render good monitors ineffective and distracting as well as contribute to staff errors and burnout.52–56 Alarms have been identified as a top-10 health technology hazard by the Emergency Care Research Institute. Multiple strategies and guidelines for managing excess alarms have been published.57–59 A recent analysis of continuous Spo2 data from patients in general care units was undertaken to understand alarm behavior under different configurations of alarm activation thresholds, annunciation delays, and notification modes.60 Common default condition-specific settings are used to detect any abnormality (Spo2 <90%, pulse rate <60 or >120 beats per minute [bpm]). Such settings with a 5:1 patient-to-nurse ratio would expose a nurse to 1971 audible alarms per 12-hour shift with traditional broadcast notification. In contrast, surveillance-specific settings configure alarms to identify only severe abnormalities (Spo2 <80% for 15 seconds, pulse rate <50 or >140 bpm; settings currently used at Dartmouth) and would reduce alarms to 10 audible alarms per shift when directed notification is used.

A 10-year retrospective study investigated the impact of continuous Spo2 monitoring (with surveillance settings) on mortality and severe morbidity associated with the administration of sedative or analgesic medications in a general care setting.16 Of 111,488 patients in units with surveillance monitoring available, none died or were harmed by OIRD. Only 1 patient died from OIRD; however, that patient was unmonitored at the time of the event. In unmonitored units (15,209 patients), 3 patients died from OIRD (19.73 deaths per 100,000 at-risk patients). These data support the contention that continuous monitoring is effective and alarms are manageable when configured for surveillance.

Selective Monitoring Will Not Prevent Hospitals From Needing to Expand the Availability of Continuous Monitoring in the General Care Setting

Given current patient acuity levels, all acute general care beds should be equipped with continuous monitoring. Even if a selective monitoring approach is used, most hospitalized patients will meet the criteria for continuous monitoring. If all patients receiving opioids are monitored per APSF guidelines, 50% to 88% of general care beds would need continuous monitors.21,22 Even for patients receiving opioid-sparing analgesic agents during postoperative care, the use of gabapentinoids, the prevalence of OSA, chronic obstructive pulmonary disease, atelectasis, and other postoperative respiratory conditions suggest a substantial majority of patients will meet indications for continuous monitoring. However, most US hospitals have continuous monitoring available only in the emergency department, intermediate care units, PACU, and ICU and not in most general care units. A recent hospital trend has been to increase intermediate care capacity and build general care units that are fully acuity adaptable. Hospitals need to invest in increasing capacity for continuous monitoring in the general care setting regardless of whether selective or universal monitoring is used.

Universal Continuous Monitoring Is Both Affordable and Associated With a Positive Return on Investment

The estimates of cost for basic surveillance continuous monitoring vary depending on individual hospital infrastructure and capabilities. Acuity adaptable beds typically require costly full-function physiologic monitors. However, pulse oximetry can be networked to central stations to provide direct telecommunication notification and is currently available at low cost. Depending on the business model used, estimates to install wireless pulse oximetry hardware range from a $0 out-of-pocket capital cost to $7000 per bed.61 Sensor costs per patient also vary depending on the type, disposability, and patient length of stay. A recent study modeling the costs of continuous monitoring reported a combined pulse oximeter capnography unit to cost $4800 and a disposable pulse oximeter probe and capnography line to cost $8.50 and $14.50, respectively.61 Telemetry to a central station with around-the-clock human monitoring has substantial costs. Furthermore, using condition-specific monitoring for a surveillance application is expensive and less effective for detecting general condition deterioration early (eg, cardiotelemetry in patients who do not meet American Heart Association criteria).62 Alternatively, our institution instituted a system where alarms were directly sent to nurse-carried pagers, which avoided the cost associated with a monitoring technician.34

Surveillance monitoring may also result in cost savings.63,64 Our initial study found that implementation of surveillance monitoring in a 36-bed unit with an estimated fixed and annualized cost of $225,000 avoided 135 ICU bed days, which carried an estimated savings of $1.5 million.63 Another study showed that a single unit equipped with continuous monitoring had a sustained decrease in ICU transfer rate (1.73% vs 2.25%, P =.03), corresponding to 360 ICU days avoided, and a savings of $1.1 million.64

Patient and Staff Acceptance Requires Proper Configuration, Training, and Attention to Human Factors

Selection of biosensors and monitoring parameters affect patient and staff acceptance. Acceptance decreases as complexity, weight, wires, and nuisance alarms increase.65 In addition, sustaining complex systems is difficult with staff turnover and unfamiliarity of new staff with equipment, a situation recently exacerbated by post–coronavirus disease-2019 (COVID-19) workstaff shortages. Although patients are more likely to be monitored postoperatively if they are given opioids, failure to configure alarms creates a major alarm-fatigue burden.60 Using default alarm settings with the alarm set on high and leaving the room door open neither builds confidence nor improves safety. Thus, human factors and usability are important considerations in minimizing technological burden.66 For example, capnography via nasal cannulae is less tolerated and understood than pulse oximetry via finger probe.66,67 Newer ergonomically designed bedside systems allow for wireless sensing of full vital signs with automated health record integration even during patient ambulation and allow for surveillance alarm configuration.68

Summary Arguments in Favor of Universal Continuous Respiratory Monitoring of Postoperative Patients

In closing, we argue that all patients are at risk for acute postoperative respiratory events and that early detection improves cardiopulmonary arrest outcomes. We fully addressed common but false narratives that these monitoring systems are not effective, are impractical because of false alarms, are costly, and are an unreasonable burden on staff. We also argue that current standards require hospitals to make monitoring more available. Quite the contrary, continuous monitoring is effective, feasible, and affordable. Continuous surveillance systems provide stressed, understaffed health care practitioners on general care units a backup safety net to allow for early detection and rapid treatment of serious but treatable adverse events during a patient’s entire hospital stay.

THE CON POSITION: SELECTIVE CONTINUOUS MONITORING

We (B.G. and M.A.O.K.) agree that all postoperative patients receiving opioids should have continuous basic monitoring, such as continuous pulse oximetry. However, we believe that centralized continuous and higher levels of monitoring should be reserved for patients deemed at high risk for OIRD. A nuanced approach for advanced monitoring should be aimed at patients at high risk, utilizing focused screening tools.

Patient and Procedural Factors

Known patient factors increase the risk of postoperative OIRD. Older age, often over 60 years, has been reported as a risk factor for postoperative OIRD.6,14,15 A narrative review showed that patients aged 61 to 70 years had 2.8 times the risk of OIRD than patients aged 16 to 45 years.69 Beyond increasing age, studies and meta-analyses have also shown higher disease burden (higher ASA classification; cardiac, neurologic, renal, and respiratory diseases; diabetes), OSA, and obesity all increase the risk for OIRD.14,15,69,70

From a procedural standpoint, orthopedic operations have been associated with a high risk for OIRD.69–71 However, a single-center study showed that patients undergoing general operations, including transplantation and hepatobiliary procedures, had the highest risk for OIRD.71 Other studies have shown that most respiratory complications were reported during the first 24 postoperative hours.9,15,69,70,72

Specific aspects of postoperative analgesic regimens have also been associated with increased risk for postoperative OIRD. Use of patient-controlled analgesic devices, which can increase opioid doses, boosts the risk for respiratory complications.69 Coadministration of multiple sedative medications increases OIRD risk, especially use of gabapentinoids as part of multimodal analgesia.8,72–74

Implications of Postoperative Respiratory Monitoring

The risks of postoperative OIRD are clearly a concern as the APSF noted years ago, and continuous pulse oximetry should be used for patients hospitalized postoperatively.13 However, using advanced monitoring modalities beyond routine nursing checks for all patients has implications for personnel and cost. Evidence suggests that respiratory deterioration typically occurs over a distinct period (4 to 8 hours of alterations in vital signs) rather than in a sudden, unanticipated event.75–78 Sedation or somnolence has also been identified as commonly preceding respiratory decompensation.79 Thus, our aim should be targeting advanced monitoring for patients at high risk and developing institutional RRTs to intervene before decompensation becomes catastrophic when there is evidence that a patient’s condition is declining.

Alarm fatigue, sensory overload, and desensitization to frequent alarms can result from use of advanced monitors and is a major concern in hospitals.80 Thus, universal use of advanced monitoring for all patients hospitalized postoperatively may not have the intended result of timely recognition of a patient’s deteriorating condition. Rather, universal monitoring could result in delayed or even no response to alarms because of staff overload and apathy.81 Studies suggest that 80% to 99% of alarms are false activations,82 and false alarms erode nurses’ trust in monitors, paradoxically decreasing patient safety.83 In a recent review of alarm fatigue in the ICU, an area with the highest staffing levels in most hospitals, nurses reported increased burden and stress from alarms,84 and alarm fatigue also led to compassion fatigue and burnout.85

The burden of universal monitoring on hospital resources is another issue. Demand for ICU resources increased, even before the COVID-19 pandemic,86,87 which further strained monitoring resources.88 Several large hospital systems have estimated an inability to provide ICU beds when needed;32 therefore, demands to monitor all patients with postoperative OIRD could surpass advanced monitoring capacities. This could lead to premature discharge of other ICU patients to lower-acuity care areas such as general units.89–91

The costs associated with unnecessary monitoring of patients also burden the health care system. In 2022, many health care systems in the United States had negative median financial margins, as federal support during the COVID-19 pandemic ended.92 Hospital expenses are increasing, primarily attributable to surging wage costs in response to labor shortages. Medicare rates are also currently projected to fall short of hospital costs. These financial pressures on hospitals are unlikely to resolve soon, and administrators will need to cut costs, not purchase additional monitoring technology.93

Inappropriate use of telemetry increased during the COVID-19 pandemic with the estimated cost reported at $1400 per day.94 This increased use and cost augment the burden of alarm fatigue and resource utilization as well as potentially affect patient care in other units. All of this adds to the cost of wasteful medical practices, estimated at $760 to $935 billion or approximately 25% of total US health care spending.95 As the focus of health care utilization has moved to providing quality and value-based care, hospital systems have responded by optimizing use of existing resources to provide safe care for all patients.96 The ability to rescue patients before a harm is dependent both on the ability to identify patients and on the afferent or response limb to a change in a patient’s clinical status. Creation of high-functioning rescue response teams can allow systems to do this.

What Targets Can We Use?

Preoperative identification of patient groups at high risk of postoperative OIRD can help optimize resource utilization and avoid the need for continuous monitoring. Patients with OSA and those at high risk for undiagnosed OSA are known to be at increased risk for postoperative respiratory complications, especially OIRD.97–101 The STOP-Bang questionnaire, the most used preoperative OSA screening tool for identifying patients at high risk of OIRD, assesses for snoring, tiredness, observed apnea, high blood pressure, body mass index (kg/m2) over 35, age over 50 years, neck circumference over 40 cm, and male sex.98 An analysis showed the sensitivity of a STOP-Bang score of 3 or more to detect moderate to severe OSA at 93% and severe OSA at 100%.98

PRODIGY (Prediction of Opioid-induced Respiratory Depression in Patients Monitored by Capnography) was a large prospective multicenter observational trial of hospitalized patients receiving parenteral opioids. The PRODIGY score uses 5 independent variables (age ≥60 years by decade, sex, opioid naivety, sleep breathing disorders, and chronic heart failure) to calculate a score that assigns patients to low (<8 points), moderate (8–14 points), or high risk (≥15) for OIRD.48 The trial used blinded and alarm-silenced bedside capnography and oximetry to identify episodes of OIRD to develop a risk-prediction model.48 Compared with patients at low risk, patients at high risk had a 6 times greater risk of OIRD, and patients at intermediate risk had 2 times greater risk.48

Risk of pulmonary complications can be predicted by the Assess Respiratory Risk in Surgical Patients in Catalonia (ARISCAT) score.102 The ARISCAT score is based on age, arterial Spo2, recent respiratory infection, preoperative anemia, and surgery-related factors (intrathoracic or upper abdominal location, prolonged duration, and emergency surgery). A score greater than 26 is predictive of an increased risk of postoperative pulmonary complications and can potentially help identify patients who might benefit from more advanced postoperative respiratory monitoring.

A patient’s clinical performance in the PACU is also an important assessment tool for identifying patients at high risk for respiratory complications. In a large study of patients admitted to the PACU, those requiring an unplanned ICU admission from the PACU with subsequent transfer to a general unit had shorter ICU and hospital stays than patients who had been discharged to a general unit and then required transfer to the ICU.103 Therefore, identifying patients in the PACU who require advanced respiratory monitoring and ICU care can potentially minimize unplanned ICU transfers from a general unit as well as minimize ICU and hospital lengths of stay. This strategy will help target patients who would benefit from advanced monitoring technology and ICU care rather than require universal advanced respiratory monitoring.

Capnography has been identified as a possible tool to detect OIRD in the PACU; when capnography was used along with pulse oximetry, adverse respiratory events were reported to be detected earlier (75% of the time) than with standard monitoring.101 However, other studies have noted the limitations of capnography, which include patient compliance,104 high rates of false-positive alarms,105 alarm fatigue, patient sleep disruptions, unfamiliarity of nursing staff, and the required investment and technology for system integration.12

The integrated pulmonary index (IPI) is an algorithm that calculates a score on a 10-point scale from 4 components: end-tidal carbon dioxide, respiratory rate, Spo2, and pulse rate.106 A prospective study of patients who were at least 75 years or obese showed that low initial IPI and fluctuating IPI in the PACU predicted the occurrence of respiratory complications postoperatively and prolonged PACU stay.106 In a prospective observational study, IPI was able to detect adverse respiratory events more rapidly than standard monitoring for 88% of patients.101

Thoracic bioimpedance monitors can continuously monitor minute ventilation (MV), tidal volume, and respiratory rate.107 These monitors have been evaluated for a wide range of patients with various breathing patterns and have an average error of less than 10%.108 In a prospective study that compared patients with low MV (defined as MV <40% of predicted MV for ≥2 minutes) versus normal MV, those with low MV within 30 minutes of PACU discharge were more likely to have episodes of OIRD on the unit than patients who did not have low MV.107 For patients classified as high risk (low MV), the positive predictive value was 61.5%, whereas for patients classified as low risk, the negative predictive value was 90.6% for postoperative respiratory depression.107

At our institution, a study identifying patients at high risk of OSA preoperatively was combined with PACU-specific respiratory assessments (episodes of bradypnea, apnea, oxygen desaturation, and pain-sedation mismatch [sedated patient with report of moderate to severe pain]).109,110 A combination of an OSA preoperative sleep apnea clinical score111 and PACU respiratory assessments helped identify patients at high risk for oxygen desaturation by using continuous recorded pulse oximetry; a higher desaturation index was associated with an increased risk of postoperative respiratory complications.109,110 On the basis of these findings, the standard process for determining which patients receive a higher level of care involves preoperative screening for OSA and PACU recurrent respiratory events. Any patient who screens as high risk in either or both categories receives remotely monitored pulse oximetry on the unit. There is a central area where the continuous pulse oximetry tracing is monitored, with acute events or trends reported to the nurse caring for the patient.

Postoperative Monitoring Options

Based on recommendations, standard monitoring for patients receiving opioids postoperatively, including those considered at low risk of OIRD, should incorporate continuous pulse oximetry or other forms of continuous monitoring.13 Values for Spo2 entered manually into health records have been reported to be 6.5% higher than results from electronically recorded saturations, thus making continuous oximetry more accurate and therefore beneficial.112 However, results of continuous pulse oximetry used to identify patients who were not originally considered at high risk of respiratory depression could be improved if patients do not receive supplemental oxygen postoperatively, which can impair hypoxic respiratory drive.113,114

An enhanced monitoring pathway has been described for bariatric surgery that can monitor continuous pulse oximetry, heart rate, and respiratory rate and that can detect respiratory failure from reduced oxygen status.44 The wireless unit communicated with the surveillance system (in this case Masimo Patient SafetyNet) and notified bedside clinicians when changes occurred, which subsequently led to a lower incidence of cardiorespiratory events and shorter length of stay. This level of monitoring was associated with fewer adverse postoperative events in this high-risk group of patients undergoing bariatric surgery. This system is similar to our institution’s use of remotely monitored oximetry.

Stepdown units can provide an intermediate level of care greater than a general care unit but less than that of an ICU. This type of unit may be an alternative for patients at high risk of respiratory complications who do not warrant ICU-level care.115 This type of unit would decrease the use of ICU resources; however, only limited evidence exists that describes the benefit of this level of care for postoperative patients.

Admission to the ICU is warranted for patients who have active ongoing cardiopulmonary issues in the PACU. As previously discussed, use of this resource limits its availability for other patient populations and increases the need for additional health care providers for these patients. Clearly, some patients require this level of care after surgery. With appropriate triage for patient risk factors, ICU use can be limited without affecting patient care.

How Do We Choose Postoperative Monitoring Strategy?

Multiple factors are associated with a high risk of postoperative OIRD, which should be considered in determining level of postoperative monitoring: continuous bedside pulse oximetry versus higher-level monitoring such as centrally monitored pulse oximetry, capnography, a stepdown unit, or the ICU. The risk factors of age, OSA (history or positive screen), and other preexisting conditions should be considered in choosing the type of unit required for postoperative care. The PRODIGY score can easily be calculated preoperatively, which can help determine which patients are at high risk of respiratory complications requiring a higher level of care.

The PACU can be used in conjunction with preoperative screening to identify the best postoperative location for patients. Respiratory tools including IPI, capnography, thoracic bioimpedance monitors, or PACU respiratory assessments can aid in determining who may need higher levels of monitoring. This approach limits the burden of resource utilization, patient disruptions, and health care provider stressors.

Summary

Postoperative respiratory depression is associated with substantial morbidity. The risk of complications can be lessened by identifying and appropriately managing respiratory compromise. However, the best way to utilize physiologic monitoring in the postoperative period remains unclear.

The pro argument advocates for continuous postoperative monitoring of all patients until hospital discharge by use of a surveillance-specific system at the bedside (Table). This approach recognizes that patients requiring postoperative hospitalization are at sufficient risk of acute respiratory decompensation during their entire hospitalization and should be continuously monitored, which allows for early detection and prevention of severe respiratory depression events. A safety net surveillance system using continuous pulse oximetry can provide the needed monitoring to detect serious but treatable complications and to mobilize timely rescue by RRTs. In addition, we argue that the cost of such a system is budget neutral and in fact has a positive return on investment, and that alarm fatigue can be readily managed if the system is configured properly. Note that applying risk prediction and selective monitoring with a centralized service is associated with institutional cost and will still be unreliable, allowing for preventable harm in unmonitored individuals at risk.

Table. - Summary of Pro and Con Positions for Continuous Physiologic Monitoring Configuration Pro position Con position Who should be continuously monitored? All inpatients All patients receiving opioids after surgery When should patients be monitored? Duration of hospitalization While patients are receiving opioids. Use a nuanced approach for patients at higher risk, as indicated by physician order Where should continuous monitoring be used? PACU
All medical/surgical inpatient units PACU
Surgical inpatient units when patients are receiving opioids How shoul