graphic

While the intraoperative period is safe, the postoperative period is not. About 4,000 deaths per year occur in the first 30 days after non-cardiac surgery (Lancet 2019;393:401). Most cardio-respiratory events happen while patients are recovering from surgery during the transition from postanesthesia care unit to the general care floor and beyond. Monitoring technology that was once limited to a tethered non-invasive blood pressure cuff – and some basic estimates of hemodynamics and respiration – has now evolved to be innovative, smarter, and more portable.

When considering novel postoperative technology, we need to be mindful of the perioperative continuum that begins preoperatively but eventually includes the postoperative phase. The immense and fascinating progress in recent years in information and communication technology, including mobile health (m-Health) apps and wearable wireless remote monitoring equipment, may have outpaced our current perioperative concepts, infrastructure, and environment. The latter will increasingly include pre- and post-hospital patients’ homes to improve care and outcomes. Common examples of current wireless and possibly remote monitoring are already embedded in our smartphones as m-Health and other apps that monitor our daily physical activity, sleep patterns, and other parameters of daily life. Individual private use of this information is customary, and connected utilization for extended health care use is still infrequent. Medical applications in connection with wearable detethered monitors that allow continuous transcutaneous glucose and possibly heart rate and EKG monitoring are becoming more common. Often these approaches are employed for personal use but they enable daily self-monitoring and affect self-directed treatment of chronic conditions (J Diabetes Sci Technol 2019;13:664-73). Integration of novel mobile and wireless hardware and software into the patient’s health care continuum is an attractive proposition and will eventually have a profound impact on perioperative care and enhanced recovery pathways and patient outcomes. Use of a mobile m-Health app-based individual prehabilitation system at home before total knee replacement has been shown to improve hospital length of stay and postoperative disposition while reducing cost (Ann Transl Med 2019;7:68). This approach includes daily exercise instructions, nutritional advice, mindfulness programs to reduce anxiety, education regarding home safety and medical risk reduction, and pain management skills. It can track patients’ preoperative physical activity and other parameters for remote or self-monitoring and lead to improved preoperative active engagement of patients in their own health and may even extend to include their home caregivers (Korean J Anesthesiol 2017;70:493-9). An estimated 25% of same-day case cancellations result from inadequate patient preparation, of which more than 70% are thought to be preventable. m-Health apps and monitors hold the promise to improve these outcomes.

For the postoperative phase following inpatient or ambulatory surgery, exciting remote wireless monitoring equipment has been developed and continues to evolve at a rapid pace. The COVID-19 pandemic may have spurred the exploration of integrated home-based telemonitoring for successful triage and acute care as a connected health care delivery option (J Am Med Inform Assoc 2020;27:1326-30). Comfortable wirelessly communicating biosensors other than for glucose and hemoglobin monitoring now also permit assessment of additional critical postoperative physiological parameters. A chest patch can assess heartrate and variability (HR, HRV) and respiratory rate (RR). A novel bio-impedance necklace includes these three parameters and can add thoracic fluid content assessment. Wrist sensors may include watches and piezoelectric-based membranes that assess HR, HRV, one-lead EKG, oxygen saturation, continuous blood pressure and physical activity (J Clin Monit Comput 2017;31:253-9). Soft, thin, stretchable, and comfortable chest patches, also known as e-tattoos or digital tattoos, can easily, continuously, and in real-time assess and transmit EKG data and blood pressure estimates, the latter being based on seismocardiography, a measure of chest vibration from heart beats (Adv Sci (Weinh) 2019;6:1900290). Sonar-based, bed-side smartphone apps that track respiration and can send alerts to first responders or other health care professionals have been described and are likely to contribute to our growing remote monitoring armamentarium (Sci Transl Med 2019;11:eaau8914). Expanding postoperative care to the patient’s home, taking advantage of novel remote continuous physiologic monitoring, requires prediction of clinically relevant events, selection of the relevant parameters to be monitored, and a systematic alert and protocolized response system to serve the right patient with the right data at the right time with the right responses (Anesth Analg 2019;129:726-34). Likewise the innovative technology needs to be robust, validated, approved and user-friendly. Provided the appropriate perioperative infrastructure and architecture in health care facilities and patients’ private environments, currently available and evolving wireless wearable continuous monitoring equipment with the ability for predictive analytics, connectivity to m-Health apps, health care systems, and providers equipped with response protocols could easily transform our current concepts of perioperative care. However, some recent data suggest the presence of digital health inequity in parallel to the existing health disparities in our society (J Am Med Inform Assoc 2021;28:33-41). Therefore, this digital health care transformation must ensure adoption across racial, ethnic, socioeconomic, and geographic lines, conquer the digital divide, include vulnerable populations, and address data sharing and privacy concerns. Lastly, in concert with answers to these interests, a sustainable telecommunication health care reimbursement program needs to evolve, which may happen rapidly as demonstrated again by the COVID-19 example where the value of telehealth visits was quickly recognized and helped facilitate the adaptation of existing compensation models (Anesth Analg 2020;131:335-9).

A more immediate question for contemporary postoperative care is who, how, and what needs to be monitored with innovative and portable wearable technology? The potential for dynamic physiologic perturbations immediately and for several days following anesthesia and surgery ideally requires reliable prediction of the patient who is at a higher risk for cardio-respiratory impairment – an open question that is now the subject of much investigation. While both hypotension and respiratory compromise in the postoperative period are common, the prediction of hypotension is not entirely possible outside the OR. On the other hand, some progress has been made regarding the prediction of respiratory depression in this period of perioperative care (Anesth Analg 2016;123:1471-9). Using available information from prospective population-based cohorts from hospital systems, predictive factors for postoperative pulmonary complications have been described and include age, preoperative SpO2, recent respiratory infection, preoperative anemia, surgical incision site, duration, and emergent nature of surgery (Anesthesiology 2010;113:1338-50; Ann Intern Med 2001;135:847-57). These factors, however, are descriptive and clinically documented, such as respiratory pathology like pneumonia, atelectasis, effusion, bronchospasm, and others. Such descriptors do not include monitored physiological signals indicating acute deviations of respiratory physiology from normal, including prolonged hypoxemia and hypoventilation alone or in combination. Clearly, improved continuous postoperative physiological monitoring currently is an important missing link in the accurate, timely, and improved detection of vulnerabilities. Acute cardiorespiratory events during postoperative patient recovery are typically preceded by gradual deterioration in clinical vital signs (Crit Care Resusc 2011;13:162-6). Continuous monitoring may detect these subtle patterns earlier and accelerate rescue prior to cardiopulmonary arrest. A variable severity and often prolonged cumulative time of postoperative hypoxemia is common and persistent after non-cardiac surgery (Anesth Analg 2015;121:709-15). Continuous SpO2 monitoring is associated with a significant improvement in the detection of hypoxemia and a trend toward fewer ICU transfers versus intermittent monitoring (Anesth Analg 2017;125:2019-29). Hypoventilation, especially in cases of opioid-induced respiratory depression (RD), may be more common and prevalent than hypoxemia (PLoS One 2016;11:e0150214; Anesth Analg 2007;105:412-8; Anesth Analg 2020;131:1012-24). The PRODIGY trial investigated continuous oximetry and capnography data, along with expert verification of waveform records and clinical relevance of patients after postoperative recovery room discharge on the hospital general care floors. The PRODIGY risk prediction score was derived from a high incidence of 46% monitor detected RD episodes and included age ≥60 years, sex, opioid naivety, sleep disorders, and chronic heart failure. This is an easy-to-implement risk prediction tool with the potential for easy integration into hospital electronic medical records and real-time use with clinical care and opioid prescribing perioperatively. Age is an important feature of several RD risk prediction models (Anesthesiology 2010;113:1338-50; Ann Intern Med 2001;135:847-57; Curr Opin Anaesthesiol 2018;31:110-9), and age-related changes in respiratory physiology and altered opioid pharmacokinetics could contribute to this observation (Clin Geriatr Med 2016;32:725-35). This may be a combination of age and related frailty and where the use of novel frailty indices in the preoperative period may help predict the same. The clinical implications of improved detection of respiratory impairment beyond better patient outcomes include health economics and resource utilization, such as a reduction of rapid response activation when using unblinded continuous capnography (J Patient Saf July 2017). The high-risk patient could therefore likely benefit from continuous monitoring that facilitates judicious use of opioids and sedatives and early clinical interventions when validated thresholds in continuously monitored physiological data are reached.

The vision for the future inpatient care journey is a wireless continuous monitoring portable device for every individual patient that links physiological data in real time to integrated patient management algorithms. However, we must understand that tools to filter noise and artifact from true signals are critical for this type of high-intensity monitoring to allow effective interventions and benefit patient safety. The average hospital nurse carries six to eight patients/shift/day on a standard general care unit. Continuous wearable technology-based monitoring would first have to solve the potential alarm burden that its use could translate into. There would also need to be a continuous assessment of useful alarm thresholds and delays on alarms (For example, a drop in oxygen saturation to less than 90% for 10 seconds is clearly a false alarm, whereas a drop for >180 seconds is likely not). Alarm burden and fatigue prompted some institutions to implement smarter monitoring, and there is an estimated ideal alarm target of up to five alarms per patient/shift/day. Moving away from reactive alarms and into the future, artificial intelligence and pattern recognition promise to allow for early risk identification from streaming vital signs and waveform analysis to anticipate and prevent predicted patient harm. This type of artificial intelligence engine would also need to digest patient data, process free text notes, imaging, labs, and other vital components of clinical information to be truly effective. Therefore, the surgery center of the future would need to be linked to such a platform and a central remote monitoring “data lake” where perioperative electronic validated patient management algorithms employ prediction models, with a combination of patient examination data, and automatically generate alerts to provider and patient in the medical center or at home with the hope to avert adverse outcomes.