Secondary Logo

Using Continuous Vital Sign Monitoring to Detect Early Deterioration in Adult Postoperative Inpatients

Verrillo, Sue Carol, DNP, RN, CRRN; Cvach, Maria, DNP, RN, FAAN; Hudson, Krysia Warren, DNP, RN, BC; Winters, Bradford D., PhD, MD, FCCM

doi: 10.1097/NCQ.0000000000000350

Background: Episodic vital sign collection (eVSC), as single data points, gives an incomplete picture of adult patients' postoperative physiologic status.

Local Problem: Late detection of patient deterioration resulted in poor patient outcomes on a postsurgical unit.

Methods: Baseline demographic and outcome data were collected through retrospective chart review of all patients admitted to the surgical unit for 12 weeks prior to this quality improvement project. Data on the same outcomes were collected during the 12-week project.

Intervention: This project compared outcomes between the current standard of eVSC and the proposed standard of continuous vital sign monitoring (cVSM).

Results: Using cVSM demonstrated a statistically significant 27% decrease in the complication rate, and a clinically significant decrease in transfers to an intensive care unit and failure-to-rescue (FTR) events rate.

Conclusions: cVSM demonstrated detection of early signs of patient deterioration to prevent FTR.

Johns Hopkins Hospital, Baltimore, Maryland (Drs Verrillo and Winters); Office of Integrated Healthcare Delivery, Johns Hopkins Health System, Baltimore, Maryland (Dr Cvach); and The Johns Hopkins University School of Nursing, Baltimore, Maryland (Dr Hudson).

Correspondence: Sue Carol Verrillo, DNP, RN, CRRN, Johns Hopkins Hospital, 1800 Orleans St, Baltimore, MD 21287 (

The authors declare no conflicts of interest.

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Website (

Accepted for publication: May 21, 2018

Published ahead of print: August 7, 2018

Failure to rescue (FTR) is failure to prevent a clinically important deterioration of a patient from a complication of an underlying condition or of medical care.1 Currently, most adult postoperative inpatient units rely on episodic, manual vital sign collection (eVSC) on which to base clinical decisions. Intermittent vital signs, taken every 4 to 12 hours, provide a single data point and may present an incomplete picture of the patient's physiologic status. A HealthGrades Report found that 1 of every 10 postoperative Medicare patients currently dies after developing pulmonary embolism/deep vein thrombosis, pneumonia/sepsis, shock/cardiac arrest, or gastrointestinal bleeding, which are categorized as hospital-acquired conditions (HACs) that are trended in many states.2 Multiparameter continuous vital sign monitoring (cVSM) allows trending of standard vital signs, such as heart rate, respiratory rate, noninvasive blood pressure, temperature, and pulse oximetry, to base clinical decisions. Use of cVSM, instead of eVSC, may help identify patient deterioration from HACs and improve patient outcomes.3

To evaluate the benefits of cVSM in improving patient outcomes, a 12-week quality improvement (QI) project was conducted with the primary aim of comparing the prevalence and incidence rates of postoperative complications, rapid response team (RRT) calls, transfers to an intensive care unit (ICU), and death rates of patients receiving cVSM, after admission to an inpatient postoperative unit compared with patients who received the usual standard of care. The purpose of this project was to prevent FTR events through detection of early signs of complications, by combining cVSM data with standard of care physiologic data, and to evaluate nurse satisfaction during and after implementing cVSM. The primary aim focused on whether cVSM could detect patient deterioration in its early stages to prevent FTR events in adult postoperative inpatients. The secondary aim evaluated nurse satisfaction regarding the usability, accuracy, dependability, and level of customer support of cVSM.

Back to Top | Article Outline


Detecting patient deterioration early is problematic on general care units. Nurses may be assigned up to 6 patients and often have to make care decisions based on eVSC taken 1 to 3 times during a single shift. This is compounded, according to the results of the Casey-Fink Survey, by new graduate nurses' self-reported lack of confidence, fear of not being able to handle the workload, and a deep fear of harming their patients.4 To address the disconnect between nursing staff skill sets and variability in having reliable and valid real-time vital sign data, one common option is to incorporate the Modified Early Warning System (MEWS). Churpek et al5 reported that the maximum MEWS score was the best predictor of cardiac arrest, followed by maximum respiratory rate, heart rate, pulse pressure index, and minimum diastolic blood pressure vital signs. However, they found that the MEWS system omits significant predictors such as diastolic blood pressure and pulse pressure index.5 The authors also reported that the MEWS algorithms are based on expert opinion rather than being derived from unit vital signs, and most MEWS systems have not been scientifically validated.5 Others have compared the MEWS to the current standard of care and found no statistically significant changes in staff response to patient deterioration.6

The literature suggests that isolated single vital signs are ineffective because they do not reflect the patient's current status.3 Notably, respiratory rate has been reported as the most inaccurately measured, yet among the most important vital signs to detect deterioration.6

It should be noted that cVSM is not the same as telemetry monitoring. cVSM includes continuous trending of multiple parameters (eg, heart rate, noninvasive blood pressure, respiratory rate, temperature, and blood oxygen saturation), whereas telemetry monitoring captures cardiac rate, rhythm, and sometimes peripheral blood oxygen saturation (SpO2).

When telemetry is used, the monitor sounds alarms as soon as a parameter breaches a rate, rhythm, or oxygenation level, which results in frequent nonactionable alarms. However, cVSM has programmed delays to allow vital signs to self-correct, thus preventing unnecessary alarms, which can lead to staff distraction and alarm fatigue.7 cVSM uses standard vital signs as a core measure for all patients, can be used on any unit, and does not require any changes in staffing patterns. Health care organizations such as the American Heart Association8 have recently published new practice guidelines delineating who should be monitored, thus placing more emphasis on closer monitoring of standard vital signs.

Back to Top | Article Outline


This project was reviewed by the authors' institutional review board (IRB00078654) and was exempted based on not meeting Federal requirements for human subject research. It was deemed as being QI.

Back to Top | Article Outline

Participants and setting

The project was conducted on a 32-bed orthopedic, orthopedic-spine, and trauma general care ward, at an urban mid-Atlantic academic medical center. A QI team, which included nurses, physicians, clinical engineers, and information technologists, collaborated on the project design. This nonmonitored postoperative unit was chosen because of the opportunity to improve care due to the volume and complexity of patients with postsurgical complications (eg, sepsis, obstructive sleep apnea, and pulmonary embolisms). During the project, cVSM was deemed the unit standard; thus, consent was obtained through assent to participate. If the patient declined, it was noted in the electronic medical record (EMR), and the patient was excluded from the project. In addition, patients were excluded if they were unable to wear the components of cVSM either due to disease, deformity, or the location of their injuries.

Back to Top | Article Outline


On admission to the general unit, patients were asked to wear the ViSi Mobile9 (Sotera Visi Mobile, San Diego, California) wireless cVSM for at least the first 48 hours of unit admission. This was the minimum amount of time to identify postoperative complications, based on the literature.10 Patients were informed of the risks and benefits associated with cVSM. If the patient verbally agreed, the monitor was applied by the nursing staff. Three chest leads were attached by standard electrodes, an accelerometer (to determine their position in the bed) was applied to the sternum, and the monitor itself was secured to the patient's wrist. A single lead from the monitor, which connected a continuous pulse oximetry probe within a continuous noninvasive blood pressure sensor wrap, was attached around the base of the patient's thumb. ViSi Mobile determines the patient's blood pressure by calculating the amount of time it takes for blood to flow from the user's chest to branches of the ulnar and radial arteries in the thumb, taking into account the patient's vascular tone. Since every patient's vascular tone is different, the device has to be calibrated once a shift to each individual.

The preset default alarm limits were set widely to reduce the likelihood of frequent, nonactionable alarms while still notifying staff of real-time, early signs of patient deterioration. Where permissible, the alarm parameters included preset delays intended to allow the patient's vital sign parameter to self-correct without any intervention (eg, heart rate had a delay of 5 seconds, systolic and diastolic blood pressure were set at a delay of 120 seconds each, pulse oximetry delay was set at 60 seconds, and respiratory rate had a delay of 120 seconds). If the patient was unable to self-correct within the preset delays, the alarm was dispatched to a central monitor and subsequently to the patient's nurse via their Wi-Fi phone. Alarms did not sound in the patients' room.

Back to Top | Article Outline

Data collection—preintervention

A power analysis showed this project needed a sample size of 306 (α = 0.05 and power = 0.80).11 Since the outcome variables are relatively rare events, a larger sample size was necessary to capture measurable event happenings.

A preintervention (control) group included a retrospective chart review of 427 patients spanning 12 weeks from August 11, 2015, through November 8, 2015. These hand chart reviews documented the number of complications noted in the patient's medical record, number of RRT calls, number of transfers to an ICU, and number of documented FTR events on the project unit.

Back to Top | Article Outline

Data collection—intervention

The intervention portion of this project was conducted for 12 weeks from December 11, 2015, through March 8, 2016. The total number of patients who assented to participate in the project was 422. The demographics of both groups are summarized in the Supplemental Digital Content, Table 1, available at: These outcomes were de-identified, entered into a secure database, and imported into IBM SPSS Statistics for Windows, Ver 21.0 (IBM Corp, Armonk, New York) for statistical analysis.

To measure Aim 1, patients in the intervention group wore a 5-parameter (ie, heart rate, continuous noninvasive blood pressure, respiratory rate, pulse oximetry, and temperature) ViSi Mobile wireless fidelity (Wi-Fi) monitor (Sotera Wireless, San Diego, California) for a minimum of 48 hours, and alarm data were collected every day by a coinvestigator. Postsurgical complications were based upon Maryland Hospital Acquired Conditions (MHACs)10 that were population specific to this unit, and the MHAC reference list used in this project was agreed upon by the project team.

Nurses on the project unit received cVSM training and were instructed to check the patient for all alarm notifications that transmitted to their Wi-Fi device. This included correlating the alarm with the patient's physical assessment and laboratory values, and documenting any actions taken. Alarm events were collected every day by the study team, de-identified, and entered into a secure database along with the patient's demographic data, which was then imported into SPSS for statistical analysis.

Because the hospital was preparing for conversion to a new EMR system, the improved workflow of having cVSM data automatically download into the EMR was not achievable during the project. Nursing staff manually entered patient vital signs at standard unit times and as needed per the patient's condition. Unit-trained cVSM experts and vendor-supplied clinical experts conducted daily rounds for the duration of the intervention time frame to assist the nursing staff with becoming familiar and comfortable with incorporating cVSM into their clinical decision-making.

To measure Aim 2, unit staff were asked to complete a 21-item Likert scale (1 = strongly disagree to 6 = strongly agree) cVSM Staff Satisfaction survey. It was distributed during the first 6 weeks of project implementation and upon project completion (after 12 weeks). Areas covered included device dependability, workload, decision-making, patient care/safety, training, troubleshooting, and overall satisfaction with the technology. Measures of central tendency were compared between the beginning of the pilot and postproject completion.

Back to Top | Article Outline


The demographic data were similar between the preintervention and intervention groups. Age and service were statistically significant but were ruled out as confounding factors by computing the Cochran-Mantel-Haenszel (CMH) test.

The incident rate of complications was 34.300 per 1000 patient days in the preintervention group compared with 9.600 per 1000 patient days for the intervention group. This was statistically and clinically significant (P < .05). The incident rate (1.110 per 1000 patient days) of RRT calls was equal in the preintervention and intervention groups and was not statistically significant. The incident rate of transfer to an ICU was also not significant; however, there was a decline in the number of transfers to the ICU in the intervention group. The incidence of FTR events declined to zero in the intervention group as compared with the preintervention group (Table 1). However, this is a rare occurrence, and although clinically significant, it is not statistically significant. Using the CMH test, there were no confounding effects from age or service based on the results (Table 2).

Table 1

Table 1

Table 2

Table 2

The first staff satisfaction survey was completed and returned by 17 of the 34 nurses (50%), and the second survey was completed and returned by 15 of the 29 nurses (52%). All surveys were completed by females with a mean length of employment of 7.7 years (range, 7 months to 33.3 years). Most of the nurse respondents rotated shifts, had a bachelor's degree, and were 29 years or younger.

Comparing the results from the 6-week implementation survey with the postimplementation (12 weeks) survey, 9 of the 21 questions had a net positive increase. One question (if it was my choice to deploy this system on my unit, would I say yes, no, or not sure) was not only statistically significant, but institutionally significant as the nurses responded that they wanted to deploy the system as the standard of care. Nine questions showed a net negative decline, and 3 questions showed no net change (see the Supplemental Digital Content, Table 2, available at:

Back to Top | Article Outline


The percent compliance with using cVSM increased from an initial 38% to a sustained average of 62%, indicating improved staff comfort level with using the new technology. This project initially revealed that staff did not always recognize a deterioration pattern or value cVSM. Research indicates that nurses have gaps in synthesizing a patient's vital sign data to recognize a pattern of deterioration.12 Project leaders worked daily with staff, helping them correlate cVSM data with patient assessments. This ongoing education helped staff understand the benefits of cVSM and its usefulness in critical thinking and clinical decision-making. By the end of the 12-week intervention time frame, the nursing staff's clinical confidence and their ability to recognize early signs of deterioration through vital sign changes improved.

Churpek et al5 reported that the patient's respiratory rate, heart rate, pulse pressure index, and minimum diastolic blood pressure were the most accurate vital sign parameters in early detection of patient deterioration, yet respiratory rate has been reported as inaccurate in several studies,12–15 and nurses rarely consider pulse pressure index when making care decisions.5 In this project, the most accurate vital sign parameter was systolic blood pressure, which had a positive predictive value (PPV) of 97%, followed by high respiratory rate (PPV of 85%) and low SpO2 (PPV of 76%), indicating high sensitivity and reliability and a low false alarm rate. As a result of incorporating cVSM into their patient assessment, nurses were able to identify signs of clinical deterioration early (eg, sepsis, autonomic dysreflexia, pulmonary embolism, and new-onset atrial fibrillation) and intervene more quickly thereby improving patient outcomes.

Back to Top | Article Outline

Analysis of age and service for potential confounding

According to Lamorte and Sullivan,16 the relative risk or risk ratio is the probability of an event happening in an exposed group versus the event happening in a control group. Applying this definition to the project, the null hypothesis was that there was no treatment effect of cVSM on the complication rate by age or service. Since it was impossible to control for confounding in the project through the usual methods of randomization or matching, the CMH test was used to compute a weighted average of the risk ratios across the subgroups.14 The CMH test showed that there was no confounding by age or service because all but one of the relative risk ratios fell between the lower and upper bounds of the confidence interval for each risk factor.13 It can be concluded, therefore, that the effect size of cVSM was not impacted by either age or service. From Table 2, the CMH result was 0.271, and when that is subtracted from one, it results in a 73% decrease in the risk of patients developing a complication in the cVSM group compared with the control group without cVSM.

Mushta et al3 advocated to make FTR a nurse-sensitive indicator and empower nurses to detect patient deterioration earlier through employing cVSM. The outcomes of this project support the literature, as the baseline complication rate was measured to be 22% with eVSC, and it decreased to 5.9% with cVSM. McGrath et al17 reported that cVSM must be accompanied by systems that will accurately identify high-risk patients for FTR events on general care wards.

Although not statistically significant, staff were positive about deploying cVSM. The literature also reflected that cVSM has been positively received by nurses in other institutions because it has been shown to be effective in detecting patient deterioration earlier than eVSC.18,19 Important staff perceptions obtained from the survey included increased trust in the cVSM system, improved patient safety using cVSM, and improved communication among team members. Staff reported being positive about use of technology in their job. The survey results supported staff finding the monitor easy to work with and were committed to the successful use of the cVSM system. It took 24/7 clinical support for the first 6 weeks of cVSM to orient each staff member and provide them with a solid foundation through real-time, hands-on practice and validation. By project completion, most staff (73%) reported that they preferred using cVSM to eVSC, and most (67%) wanted to deploy it as a permanent standard of care. The literature suggested that cVSM may lead nurses to initiate interventions to mitigate FTR events, resulting in nurses perceiving that cVSM enhanced patient safety.6,12 Other authors have found that cVSM was seen as a positive tool by nursing staff to detect early deterioration prior to serious events developing.14,15

Back to Top | Article Outline


This project was conducted for a short period in a single postoperative general unit with a specific population of patients; therefore, the results are not generalizable for practice. Not all patients on the unit had cVSM during the intervention time frame. The staff survey questions were not validated prior to use, but demonstrated reliability and internal consistency.

Back to Top | Article Outline


This project demonstrated that cVSM can detect early deterioration in adult postoperative inpatients to prevent FTR events. The complication rate decreased from the baseline of 22% to 5.9%, the number of RRTs was unchanged, the transfer rate to an ICU decreased 0.09%, and there were no FTR events, which accomplished the purpose of the project.

As staff competency improved in using and incorporating cVSM into their daily decision-making, the nurses were strongly positive about the value and importance of keeping cVSM as the unit standard. Using high fidelity continuously available data for decision-making began addressing the new graduate nurses' greatest concerns of lacking confidence, harming their patients, and being unable to handle the workload by providing a more robust picture of the patient's clinical status. Having continuous data also helped the nurses prioritize patient care more accurately and thus helped the nurses make effective workload adjustments, by alerting the providers to acute early deterioration in their patients and requesting further workup.

Back to Top | Article Outline


1. Jones S, Bottle A, Griffiths P. An Assessment of “failure to rescue” derived from routine NHS data as a nursing sensitive patient safety indicator for surgical inpatient care. Report from the National Nursing Research Unit, Florence Nightingale School of Nursing and Midwifery at King's College London. Published 2011. Accessed April 6, 2018.
2. Reed K, May R, Nicholas C, eds. HealthGrades Patient Safety in American Hospitals Study. Denver, CO: HealthGrades, Inc; 2013.
3. Mushta J, Rush KL, Andersen E. Failure to rescue as a nurse-sensitive indicator. Nurs Forum. 2017;53(1):84–92.
4. The Johns Hopkins Hospital Department of Nursing. The Johns Hopkins Hospital Adult and Pediatric Nurse Residency Programs: CY 2014 & 2015 Annual Reports Comparison for the Johns Hopkins Hospital. Baltimore, MD: The Johns Hopkins Hospital Department of Nursing; May 18, 2017.
5. Churpek MM, Yuen TC, Huber MT, Park SY, Hall JB, Edelson DP. Predicting cardiac arrest on the wards a nested case study. Chest. 2012;141(5):1170–1176.
6. Kyriacos U, Jelsma J, Jordan S. Early warning scoring systems versus standard observations charts for wards in South Africa: a cluster randomized controlled trial. Trials. 2015;16:103.
7. Winters BD, Cvach MM, Bonafide CP, et al Technologic distractions (part 2): a summary of approaches to manage clinical alarms with intent to reduce alarm fatigue. Crit Care Med. 2015;40(1):130–137.
8. Sandau KE, Funk M, Auerbach A, et al Update to practice standards for electrocardiographic monitoring in hospital settings: a scientific statement from the American Heart Association. Circulation. 2017;136(19):e273–e344.
9. Sotera Wireless. The ViSi Mobile® System. Accessed April 18, 2018.
10. Colmers JM, Wong HS, Kinzer D, et al State of Maryland Department of Health and Mental Hygiene, Health Services Cost Review Commission; 2016. Accessed September 15, 2016.
11. Fleiss JL, Levin B, Paik MC. In: Shewart WA, Wilks SS, eds. Statistical Methods for Rates and Proportions. 3rd ed. Hoboken, NJ: John Wiley & Sons, Inc; 2003. Accessed March 23, 2017.
12. Cardona-Morrell M, Prgomet M, Turner RM, Nicholson M, Hillman K. Effectiveness of continuous or intermittent vital signs monitoring in preventing adverse events on general wards: a systematic review and meta-analysis. Inter J Clinl Prac. 2016;70(10):806–824.
13. Zimlichman E, Szper-Kravitz M, Shinar Z, et al Early recognition of acutely deteriorating patients in non-intensive care units, assessment of an innovative monitoring technology. J Hosp Med. 2012;7(8):628–633.
14. Cardona-Morrell M, Prgomet M, Turner RM, Nicholson M, Hillman K. Effectiveness of continuous or intermittent vital signs monitoring in preventing adverse events on general wards: a systematic review and meta-analysis. Int J Clin Prac. 2017;70(10):806–824.
15. Haahr-Raunkjaer C, Meyhoff CS, Sorensen HBD, Olsen RM, Aasvang EK. Technological aided assessment of acutely ill patient. The case of postoperative complications. Eur J Intern Med. 2017;45:41–45.
16. LaMorte WW, Sullivan L. Confounding and Effect Measure Modification. Accessed March 23, 2017.
17. McGrath SP, Taenzer AH, Karon N, Blike G. Surveillance monitoring management for general care units: strategy, design, and implementation. Jt Comm J Qual Patient Saf. 2016;42(7):293–302.
18. Taenzer AH, Pyke JB, McGrath SP. A review of current and emerging approaches to address failure-to-rescue. Anesthesiology. 2011;115(2):421–431.
19. Menendez ME, Ring D, Bateman BT. Preoperative opioid abuse is associated with increased morbidity and mortality after elective othopaedic surgery. Clin Orthop Relat Res. 2015;473(7):2402–2412.

continuous vital sign monitoring; episodic vital sign collection; failure to rescue; patient deterioration; postoperative complications

Supplemental Digital Content

Back to Top | Article Outline
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved