Is continuous monitoring the next revolution for patient safety? 

Severe complications occur in about 30% of patients admitted for severe medical diseases or major surgical procedures (1). What worries patient safety is that up to 50% of these could potentially have been prevented (2) if clinicians had timely and sufficient information about the deterioration. These alarming numbers occur despite long-standing efforts within admission prevention in anaesthesia, perioperative medicine, and intensive care unit (ICU). In absolute numbers, mortality is higher in general wards than in high-dependency units such as the ICU, the operating rooms (OR) or the post-anaesthesia care unit (PACU)(3). The reasons for this are multifactorial, including the high number of patients treated outside the OR/ICU, low nurse-to-patient ratios, limited interventions and clinical training, and patient observation relying on manual intermittent monitoring. Vital signs (pulse, SpO2, respiration rate, blood pressure, etc.) are essential to this patient observation. They are typically incorporated into the widely adopted national Early Warning Score (NEWS), although its implementation did not improve clinical outcomes (4,5). However, advancements in wearable vital sign sensor technology and machine learning show promise and will – if developed and implemented correctly – undoubtedly be a valuable part of the perioperative toolbox and improve patient safety. 

It is well-established that wearable vital sign sensors detect deviations up to 50 times more often in the general ward than manual intermittent monitoring (6,7). But without intelligent alarms, continuous vital sign monitoring (CVSM) comes at a cost of hundreds of alerts per patient per day because standard monitors typically re-issue alerts every 2 minutes. This results in alert fatigue due to alerts being irrelevant in the sense that they either have no relation to complications (alerts are given for brief fluctuations in vital signs that are not dangerous) or that the alerts are related to already recognized complications (that staff have already been notified by previous alerts). Alert fatigue results in ignorance of important alerts, with direct consequences for the patients and a false sense of security. 

The high CVSM-generated alert rates come from the fact that we have forgotten to use our physiological knowledge in system design. Thus, current alert systems are simplistic, such as SpO2 falling below a set threshold of 92%. They do not account for frequent deviations in vital signs that don’t lead to serious complications or are part of normal fluctuations in human physiology, such as activity and time of day. Even more fundamental, we have included vital sign variables in the NEWS just because we could – not because the evidence was built from the beginning. The most central example is the blood pressure assessment, which aims to quantify organ perfusion but often has no obvious correlation between the two. Perhaps it is time to revisit the foundation on which we build our alerts, although the task of removing century-long practices such as blood pressure measurement seems unsurmountable. Regardless, reducing irrelevant alerts is critical if the potential of continuous vital sign monitoring is to be fulfilled, as the staffing resources at the general ward do not allow nurses to spend time pausing alerts, as is the case in the OR/ICU. Furthermore, irrelevant alerts may ultimately result in unwarranted investigations and/or treatments with potential patient harm and increased healthcare expenditure. 

Most studies within CVSM use observational or case-control design, and there is a paucity of large randomized clinical trials, especially for systems that include artificial intelligence (AI) augmentation. Nonetheless, there is mounting evidence that CVSM in the general ward improves outcomes, including severe complications, ICU admissions, and mortality (8). CVSM technology can reduce nurses’ workload by automating vital sign monitoring and documentation in electronic medical records, supporting earlier detection of clinical deterioration, and instilling a sense of security (9). The latter factor is not well-explored but may be an important factor in job retention, which is extremely important given the worldwide nursing shortage. Evidence of CVSM cost-effectiveness is still limited, and experience from full-scale clinical implementation is sparse. Barriers include nurse engagement and integration with other important safety aspects in real-time, such as alertness, urine output, pain, nausea, etc., which still require bedside evaluation.  

Combining the novel wearable sensor technology with clinically developed AI seems to be the sound way forward. Not only does AI reduce alerts several-fold without loss of sensitivity (9,10), but it can also help identify stable patients, reducing unnecessary monitoring. This would potentially pave the way for early safe discharge and optimizing resources. Especially if the CVSM is continued in the patients’ homes with real-time contact with the healthcare system in case of need.  

Integrating AI into CVSM requires attention to implementation barriers, such as human-machine trust, staff training, device design, technological stability, and internet connectivity. Simple obstacles such as device design and user-fit may not be advanced, but experiences show that this is critical to obtain true continuous data, the foundation for the systems. Thus, there must be an interaction between device producers/designers and clinicians to further development. Nurses generally have positive attitudes toward CVSM, citing time savings and earlier interventions as benefits, but technological limitations remain a concern. The success of CVSM relies on incorporating staff feedback during development to ensure clinical uptake and building the evidence for the clinical impact through large trials with relevant clinical outcomes (9). 

It is important to emphasize that AI must be subjected to the same standards for all medical technologies and interventions (the MDR regulation). This will also assess bias from development in one patient population and potential use in another. The recent focus on pulse oximeter bias in people with dark-coloured skin highlights the issues within CVSM, and regulatory approvals and clinical benefits documented in well-conducted trials in relevant populations must be obtained before widespread use. 

New approaches to interpreting continuous vital signs may bridge the current gap in postoperative patient safety, enabling earlier recognition of complications, reducing unplanned ICU admissions, and potentially improving outcomes. Continuous monitoring must be implemented carefully with strategies to reduce alarm fatigue and improve decision-making. Current evidence supports further academic investments in exploring CVSM technology, from sensors and alert design to large-scale clinical tests and implementation. Focusing resources will undoubtedly aid in fulfilling the promise of CVSM as the next revolution in patient safety. 

Authors

• Eske Kvanner Aasvang, Department of Anesthesia, Centre for cancer and organ dysfunction. Copenhagen University Hospital, Rigshospitalet, Copenhagen. Denmark, Department of Clinical Medicine, University of Copenhagen, Denmark  
• Christian S. Meyhoff, Department of Clinical Medicine, University of Copenhagen, Denmark. Department of Anaesthesia and Intensive Care, Copenhagen University Hospital – Bispebjerg and Frederiksberg, Copenhagen, Denmark 

Acknowledgements: None 

Financial support and sponsorship: None 

Conflicts of interest: CSM and EKA have founded WARD247 ApS with the aim of pursuing the regulatory and commercial activities of the WARD project (developing a clinical support system (WARD-CSS) for continuous wireless monitoring of vital signs). WARD247 ApS has obtained a license agreement for any WARD-project software and patents. One patent has been filed: “Wireless Assessment of Respiratory and Circulatory Distress (WARD), EP 21184712.4 and EP 21205557.8”. WARD247-ApS has obtained CE approval of the WARD-clinical support system and filed for FDA approval. 

References 

1. International Surgical Outcomes Study group. Global patient outcomes after elective surgery: prospective cohort study in 27 low-, middle- and high-income countries. Br J Anaesth. 2016;117:601–609. 
2. Panagioti M, Khan K, Keers RN, Abuzour A, Phipps D, Kontopantelis E et al. Prevalence, severity, and nature of preventable patient harm across medical care settings: systematic review and meta-analysis. BMJ. 2019;366:4185.  
3. Vascular Events in Noncardiac Surgery Patients Cohort Evaluation (VISION) Study Investigators; Association between complications and death within 30 days after noncardiac surgery. CMAJ . 2019;191:E830-E837. 
4. McGaughey J, Fergusson DA, Van Bogaert P, et al. Early warning systems and rapid response systems for the prevention of patient deterioration in acute adult hospital wards. Cochrane Database Syst Rev. 2021;11:CD005529. 
5. Gerry S, Bonnici T, Birks J, et al. Early warning scores for detecting deterioration in adult hospital patients: systematic review and critical appraisal of methodology. BMJ. 2020;369:m1501. 
6. Z. Sun et al., Postoperative Hypoxemia Is Common and Persistent: A Prospective Blinded Observational Study. Anesthesia and Analgesia. 2015;121:709–715.   
7. Duus CL, Aasvang EK, Olsen RM, et al. Continuous vital sign monitoring after major abdominal surgery-Quantification of micro events. Acta Anaesthesiol Scand. 2018;62:1200–1208. 
8. Rowland BA, Motamedi V, Michard F, et al. Impact of continuous and wireless monitoring of vital signs on clinical outcomes: a propensity-matched observational study of surgical ward patients. Br J Anaesth. 2024;132:519-527 
9. Aasvang EK, Meyhoff CS. The future of postoperative vital sign monitoring in general wards: improving patient safety through continuous artificial intelligence-enabled alert formation and reduction. Curr Opin Anaesthesiol. 202336:683-690. 
10. Kjærgaard K, Mølgaard J, Rasmussen SM, Meyhoff CS, Aasvang EK. The effect of technical filtering and clinical criteria on alert rates from continuous vital sign monitoring in the general ward. Hosp Pract (1995). 2023;51:295-302. 

Figure Example of an MDR-approved AI-augmented continuous vital sign monitoring system 

Legend: HR: Heart Rate, RR: respiration Rate, BP: Blood Pressure, SAT: Peripheral Oxygen Saturation, ECG: ElectroCardioGram  WARD: Wireless Assessment of respiratory and circulatory Distress.