Patient Engagement

Reducing Inpatient Falls and Injury Rates by Integrating New Technology with Workflow Redesign

Case Study · May 23, 2019

By integrating a novel patient-observer technology into a redesigned clinical workflow, the neuroscience unit at Mission Hospital in Asheville, North Carolina, achieved a 23% reduction in falls and a 12% reduction in fall-related injuries over a 1-year pilot period. Furthermore, the technology and clinical workflow was scaled to two additional units and, collectively, the three units achieved a 40% reduction in fall-related injuries. These results supported scaling the solution to a total of five inpatient units, effectively expanding the number of patients capable of being monitored from six (during the pilot period) to 164.

Key Takeaways

  1. A busy neuroscience, medical, or surgical unit can successfully redesign its processes to adopt innovative fall-reduction technology.

  2. Review your institution’s multicomponent fall-prevention program and monthly falls data to identify opportunities for improvement.

  3. Use a multidisciplinary team to come up with creative ways to reduce inpatient falls and involve that team in the implementation phase.

  4. Ensure that technology solutions and redesigned clinical workflows complement each other.

  5. Identify metrics of success, track them continuously, and analyze them at regular intervals.

  6. Design and apply technology to deliver value across multiple use cases.

  7. Plan beyond the pilot in preparation to scale successes.

The Challenge

The reported rate of patient falls in acute-care hospitals ranges from 1.3 to 8.9 per 1,000 patient-days of care.1 Of the hundreds of thousands of patients who fall in U.S. hospitals each year, 30% to 50% are injured.2,3 Such falls frequently prolong or complicate hospital stays, adding 6.3 days to the typical stay, according to one study.4 A single patient fall with injury costs an average of nearly $14,000.5 Risk factors for falls include, but are not limited to, age, history of recent fall, mobility impairment, urinary incontinence or frequency, certain medications, postural hypotension, and cognitive impairment. Multiple nonclinical factors also contribute to acute inpatient falls, including inadequate staff orientation, supervision, staffing levels, or skill mix; lack of adherence to protocols; deficiencies in physical environment; communication failures; and poor room design (e.g., trip hazards, suboptimal chair heights, inadequate lighting, etc.).

Furthermore, depending on the unit within the hospital, the daily number of alarm signals per patient can reach several hundred. As a result, clinicians become desensitized or immune to the sounds because they are overwhelmed by information and suffer from “alarm fatigue.”6 A systematic review of the inpatient fall-prevention literature concluded that inpatient multicomponent programs reduce falls effectively but provided no clear evidence about which components work best.7

At Mission Health, we have long adhered to standard fall-prevention interventions: beds locked in the low position with rails up; call lights, call bells, assistive devices, and personal items within patient reach; nonslip patient footwear; clutter-free rooms; appropriate use of sensory aids; dry floors and adequate lighting; hourly clinician rounds; and patient and family education. We also have used individually tailored interventions such as bed and chair alarms, yellow identification armbands, and alert signage outside of patients’ rooms. Despite our multicomponent fall-prevention program, compassionate care staff, and strong culture of safety, falls on the neuroscience unit still occurred; between August 2014 and July 2015, a total fall rate of 5.74 per 1,000 patient-days of care was observed. We believed that this rate could be significantly reduced.

The Goal

Mission Health collaborated with Cerner Corporation to design, develop, and trial a scalable strategy for reducing falls in the hospital’s 34-bed neuroscience unit. The approach, jointly funded by Mission Health and Cerner, emphasized the use of new technology and redesigned workflows in addition to the standard fall-prevention program. The goal was to conduct a 90-day pilot in order to evaluate the impact of a “virtual sitter” system (described below) as an alternative to using qualified staff sitters in patients’ rooms. The impact of the intervention was measured by comparing the two approaches in terms of fall rates, injury rates, and patient-to-sitter ratios. Both groups received the standard fall-prevention interventions listed above.

The Pilot Execution

“Virtual Sitter” Technology

In October 2014, Cerner Corporation first introduced the Mission Center for Innovation to Microsoft Kinect technology, which uses a depth sensor to visualize full-body 3-D movement within a 20-foot range. The system’s infrared camera, acting as a virtual sitter, can monitor a patient’s movements under any ambient light. With use of open-software programming, a unique application can be created to define and draw virtual zones, trip wires, and other trigger points around a patient within a field of view (Fig. 1). A continuous live video feed is sent across a secure wireless network to a high-definition monitor (Fig. 2). Additional features include two-way audio, voice recognition, patient-privacy modes, and customizable alerting. When clinical leaders and bedside staff were shown a demonstration with mock patients, the nursing unit manager for the neuroscience unit immediately recognized that this system might reduce the fall rate for high-risk patients and volunteered to be the clinical sponsor for a pilot project.

Mission Hospital Virtual Sitter - Drawing of the Virtual Bed Zone and Rails

Figure 1. Click To Enlarge.

Mission Hospital Virtual Sitter - Drawing of the Piloted Central Monitoring Technician Workstation

Figure 2. Click To Enlarge.

Nursing Workflow

The neuroscience unit had the highest historical fall rate at Mission Hospital (excluding behavioral health units). Clinical leaders, supported by the neuroscience unit director and nursing manager, assembled a multidisciplinary project team focused on adopting the virtual sitter technology and redesigning processes to reduce fall rates. Almost all patients admitted to the neuroscience unit are at risk for falls according to the Morse Fall Scale (MFS). Given that we had only six virtual sitter cameras for the pilot, a team of registered nurses from the neuroscience unit used the MFS criteria and clinical expertise to develop an algorithm to assist nursing staff in identifying the highest-risk patients for participation in the pilot (Fig. 3).

If a patient met at least two criteria for high risk, the nurse educated the patient and/or family member about the virtual sitter system, entered a nursing order, and registered the patient with the monitor technician (stationed at a central location on the unit). A wireless camera was placed on a wall mount and was plugged into an electrical outlet in the patient’s room. Within 1 minute, the monitor technician had a full view of the room, could introduce himself or herself via two-way audio, and then could briefly explain the virtual sitter monitoring system to the patient and family. The algorithm’s effectiveness in identifying patients at high risk for falls has not been fully validated.

Virtual Sitter Patient Fall Risk Algorithm - Mission Hospital

Figure 3. Click To Enlarge.

Technician Workflow

A team of six certified nursing assistants from the hospital staffing pool were extensively educated on the system’s software. These trained individuals were called monitor technicians and were responsible for configuring the patient rooms and observing the patients. Configuration consisted of drawing virtual zones and virtual trip wires around the patient’s bed or chair and setting visual and auditory alerts to detect when a patient moves across the virtual boundaries. When a boundary is crossed, a red box outlines the patient’s video display window on the monitor and an easily interpretable audible alert delivers the appropriate message, such as “Patient is getting out of bed.” The highly customizable, user-friendly interface ensures that the Kinect sensor detects risky patient behaviors while minimizing alert fatigue.

Clinical Pilot Phase

After nursing-staff champions and monitor technicians on the neuroscience unit became comfortable with the new workflows and technology, a 3-month, six-camera virtual sitter pilot was launched in August 2015. The aforementioned algorithm was used to identify patients at highest risk for falls, who represented the ideal candidates for the new system. Patients who were at risk for suicide, homicide, or drug overdose and those who were under legal restriction required an in-person sitter (not a virtual sitter), as per hospital policy. In all, 15% of patients in the neuroscience unit were enrolled in the 90-day pilot.

When a patient’s movements generated a virtual sitter alert, the monitor technician intervened using the following escalation pathway:

  1. Use the two-way audio interface to interact with the patient. Either direct the patient to remain in the bed or chair or assess the need for further intervention.
  2. Contact the patient’s registered nurse or certified nursing assistant, depending on the patient’s need.
  3. Contact the unit supervisor for emergency situations.

Monitor technicians used VoIP (Voice over Internet Protocol) phones to communicate with the care team when their involvement was required. Each technician intervention was recorded on a log sheet so that the total number of interventions and the reasons for these interventions could be tracked. The pilot project team met weekly to review these data sets.

The Pilot Metrics

The 98 patients who were monitored with the virtual sitter (during 348 patient-days of care) experienced 0 unassisted falls and 0 injuries during the 90-day period from August through October 2015. All other patients on the neuroscience unit had an unassisted fall rate (per 1,000 patient-days of care) of 4.06 and an injury rate of 2.45 during the same period. Analysis showed that verbal redirection alone (accounting for 50% of all virtual sitter interventions during the pilot) was highly effective in keeping the patient in his or her bed or chair. This finding was significant because the verbal redirection not only averted patient falls, but it also prevented nurses from being interrupted from other tasks.

These highly encouraging results prompted hospital leaders to support the continued use and expansion of the virtual sitter technology and redesigned workflows. Figure 4 compares the rates of falls and injuries for the entire neuroscience unit during the 12 months before and after the virtual sitter program was instituted. Although only about 15% of patients were enrolled in the virtual sitter program during the second 12 months, the overall annual rates of falls and injuries for the unit declined significantly — indicating that, in effect, an intervention involving a small subset of patients was so powerful that it had a clear impact on the unit’s overall performance.

Unassisted Fall and Injury Rates Before and After Virtual Sitter Implementation at Mission Hospital

Figure 4. Click To Enlarge.

Before the virtual sitter pilot, the annual number of hours for nonmandatory bedside staff sitters (one sitter per patient) was approximately 47,000 (cost equivalent: approximately $680,000). The average cost of using one in-room sitter for each patient is approximately $350 per inpatient day. Effectively implementing the virtual sitter solution with a monitor tech-to-patient ratio of 12:1 would reduce this average cost to $29 per patient per inpatient day.

Throughout the 12-month pilot period, nursing unit staff, patients, and family members were asked to share their thoughts on the virtual sitter program. The responses were overwhelmingly positive, with many respondents noting that having a trained monitor technician maintain a watchful eye on the patient 24-7 provided added safety and comfort. Monitor technicians also validated the benefit of the program, citing many examples in which they were able to assist in the prevention of a fall by communicating with the patient and the unit care team. Monitor technicians also reported that they were able to develop closer relationships with their patients, resulting in interactions that were both effective and rewarding.

The Scaled Solution

The results of the virtual sitter pilot project exceeded initial goals and expectations, and executive approval was obtained in January 2017 to scale the virtual sitter solution. The planning process began in March 2017. The goal was to grow the solution from 6 cameras on the neuroscience unit to 21 cameras across three inpatient units (the neuroscience unit, a medical-surgical unit, and a cardiovascular care unit), each with approximately 32 beds. The team of trained monitor technicians would need to grow from 6 to 12 to support 24-7 monitoring of as many as 12 patients each.

In order to scale our virtual sitter solution, we needed a larger monitoring location. In parallel, but unrelated to the virtual sitter project, Mission Health was upgrading and redesigning its cardiac telemetry system. We therefore took this opportunity to develop an off-site centralized monitoring unit (CMU) that would monitor both cardiac telemetry and virtual sitter patients. We believed that this step would enable economies of scale, achieve standardization, optimize staffing, and ensure professional staff satisfaction.

A state-of-the art CMU was constructed in an existing nonclinical Mission facility located approximately 8 miles away from the main hospital. A new CMU monitor technician role, requiring skills and competencies in both cardiac telemetry and virtual sitter procedures, was created. Although centralizing patient telemetry is considered to be evidence-based best practice, Mission believed that colocating and cross-training staff in this manner represented an innovative approach.

There were two significant operational requirements to scale the virtual sitter solution. The first involved standardization of the nursing clinical workflows and processes. To that end, the Mission nursing education and research team led the creation of virtual sitter clinical workflows and the development of educational materials. Staff members on each nursing unit receiving the virtual sitter solution were then trained and qualified by members of the nursing education team. The second requirement was the deployment of a standard mobile communication device with secure voice and text functionality. In order to deploy the virtual sitter solution to a nursing unit, each member of the staff was required to have the standard mobile device on their person at all times. Having a health system–issued mobile device ensured HIPAA compliance and enabled the monitor technicians to rapidly and reliably communicate to frontline nursing unit staff as needed.

By November 2017, the CMU was fully staffed and operational. All clinical workflows were standardized, nursing unit staff was trained, and 21 cameras were deployed to the three units.

The Scaled Metrics

Table I shows impact on unassisted falls and injury rates for the period from November 2017 through December 2018:

Impact of Virtual Sitter Solution on Unassisted Falls and Injury Rates at Mission Hospital

Table I. Click To Enlarge.

Aggregated data for all three units after the implementation of virtual sitter monitoring demonstrated a 44% reduction in unassisted falls; this finding was statistically significant (p < 0.001). Injuries resulting from falls were reduced by 40%; however, because of the overall low occurrence of fall-related injuries, there were not enough data to demonstrate statistical significance (p = 0.065). In practical terms, the 40% reduction is equivalent to 14 fewer injuries from falls (when normalizing patient-days of care for each period). When applying the referenced average cost of $14,000 per fall with injury, the total cost avoidance is $196,000 (over the 14-month period).

Over this same time period, 145,000 hours of patient-monitoring was conducted by a remote monitor technician staff of 8.4 full-time equivalents (FTEs). In contrast, monitoring these patients with use of the conventional ratio of 1 in-room sitter per patient would have required approximately 60 FTEs.

Work is underway to further expand the virtual sitter solution. The goal is to deploy a total of 164 cameras across five units by late 2019.

Lessons Learned

  • When scaling to multiple units, the key assessment trigger used to identify which patients qualified for virtual sitter monitoring was fall risk according to the Morse Fall Scale. This trigger was less stringent than the criteria that had been piloted on the neuroscience unit. The intent was to ensure that the vast majority of patients would be considered for virtual sitter monitoring; however, the following lessons were learned:
    • Patients who are not redirectable by the central monitor technician voice commands are not ideal candidates for a virtual sitter. Non-redirectable or noncooperative patients are at higher risk for impulsive behavior, which could lead to a fall while being monitored. These patients may be better managed with an alternate intervention.
    • The Morse Fall Scale is not sufficiently sensitive and predictive for the purpose of identifying patients who are appropriate (or not appropriate) for a virtual sitter. To achieve the best possible outcome for all fall-risk patients, an evidence-based patient-assessment algorithm to assess fall-risk patients who are also appropriate for the virtual sitter intervention would be impactful.
  • While every member of the care team carries a mobile device, the response time is highly variable, largely because it can be impacted by (a) resource constraints and workload on the unit and (b) frequent calls to the unit regarding non-redirectable patients, which can lead to alert fatigue.

Future Development

A more predictive and robust patient-assessment algorithm is currently being developed. This algorithm is expected to produce better patient outcomes and reduce the number of calls to frontline staff, in turn reducing alert fatigue. Additionally, technology enhancements that allow “real-time closest caregiver” alerting and by-passing CMU phone calls will also improve staff response times.

We believe that this technology will become standard of care for the at-risk patient; therefore, real-time frontline communication, onboarding, and off-boarding for hundreds of patients weekly will have to be streamlined and automated. Such automation will require additional technologies and solutions to seamlessly integrate with various electronic health records and workflows. With the planned deployment of 164 cameras in patient rooms, we also believe that the technology can be leveraged as a platform and utilized beyond fall-risk monitoring. We are aggressively exploring other patient-safety initiatives for at-risk patient populations, as described below.

Patients at High Risk for Substance Abuse

Hospitalizations for endocarditis and drug dependence in North Carolina increased twelvefold between 2010 and 2015, with an average cost per hospitalization exceeding $50,000;8 given the current opioid epidemic, this trend is likely to increase.9,10 The population of patients who are at high risk for substance abuse (HRSA) is challenging11 and resource intensive, with higher rates of related issues (e.g., leaving the hospital against medical advice, visitors bringing in contraband, workplace violence, other addictions, and mental health disorders, etc.) in comparison with other patients.

To address the unique needs of this population, a multidisciplinary team at Mission Hospital developed and piloted an HRSA care process model. Building on the success of the virtual sitter program for fall-risk patients, this solution is currently being piloted as part of the HRSA patient care plan. Beginning in August 2018, the virtual sitter solution became an option for patient observation when ordered by the provider. The pilot involves a dedicated team of trained monitor technicians who can monitor up to 6 HRSA patients at a time. The pilot unit includes a dedicated certified nursing assistant who receives calls from the CMU on high-risk patient behaviors and rounds on HRSA patients at regular, but unpredictable, intervals.

During the first 4 months of the pilot, at least three potential adverse incidents were prevented after the monitor technician observed high-risk behavior and/or contraband and then rapidly deployed a member of the care team to intervene. While there are insufficient data at this point to determine whether the pilot has had a statistically significant impact, there has been a trend toward lower rates of HRSA patients leaving against medical advice (16.7% for patients with a virtual sitter, compared with 25.3% without). There also has been an increase in the length of stay (8.3 days for patients with a virtual sitter, compared with 5.5 days without), suggesting that patients are more compliant with their care plan when being monitored by a virtual sitter. Patients have commented that they feel more comfortable and private with the virtual sitter as compared with an in-room sitter.

Patients with Behavioral Health Issues

Future pilots involving patients with behavioral health issues, including those with suicidal ideation risk, are being considered. The Joint Commission has provided guidance for managing such patients, including the use of “continuously monitored video.”12,13 The potential application of virtual sitter technology as part of a comprehensive solution for improved care in the behavioral health setting is compelling.

Moving from Reactive Responses to Predictive Responses

With the advancements of data analytics and artificial intelligence, the technology naturally could evolve such that the reactive responses that are provided by monitor technicians today could be augmented by more predictive responses in the near future. We believe that seamlessly and successfully integrating the virtual sitter with other in-room technologies, such as “smart beds” (beds with sensor technology designed to enhance patient safety), telehealth (virtual care), electronic medical documentation, and in-room entertainment, has the potential to significantly improve outcomes and experiences for patients and their care teams.


Acknowledgements: Thanks to Mary Jackson, MBA, BSN, RN, NEA-BC; Nursing Practice, Education, and Research team; Information Technology; Central Monitoring Unit; and the following nursing units’ care team members and clinical support staff: Neurosciences, 3 West Adult Medical Surgical, Cardiovascular Progressive Care, and 8 North Adult Medical Surgical.


1. Oliver D, Healey F, Haines TP. Preventing falls and fall-related injuries in hospitals. Clin Geriatr Med. 2010;26:645–92.

2. Healey F, Scobie S, Oliver D, Pryce A, Thomson R, Glampson B: Falls in English and Welsh hospitals: a national observational study based on retrospective analysis of 12 months of patient safety incident reports. Qual Saf Health Care, 2008;17(6):424–30.

3. Shekelle PG, Wachter RM, Pronovost PJ, Schoelles K, McDonald KM, Dy SM, Shojania K, Reston J, Berger Z, Johnsen B, Larkin JW, Lucas S, Martinez K, Motala A, Newberry SJ, Noble M, Pfoh E, Ranji SR, Rennke S, Schmidt E, Shanman R, Sullivan N, Sun F, Tipton K, Treadwell JR, Tsou A, Vaiana ME, Weaver SJ, Wilson R, Winters BD. Making Health Care Safer II: An Updated Critical Analysis of the Evidence for Patient Safety Practices. Comparative Effectiveness Review No. 211. (Prepared by the Southern California-RAND Evidence-based Practice Center under Contract No.290-2007-10062-I.) AHRQ Publication No. 13-E001-EF. Rockville, MD: Agency for Healthcare Research and Quality. March 2013.

4. Wong C, Recktenwald AJ, Jones ML, Waterman BM, Bollini ML, Dunagan WC: The cost of serious fall-related injuries at three midwestern hospitals. Jt Comm J Qual Patient Saf, 2011;37(2):81–7.

5. Haines T, Hill A-M, Hill KD, Brauer SG, Hoffman T, Etherton-Beer C, McPhail SM: Cost effectiveness of patient education for the prevention of fall in hospital: economic evaluation from a randomized controlled trial. BMC Med. 2013;doi:10.1186/1741-7015-11-135.

6. The Joint Commission Sentinel Event Alert. Medical device alarm safety in hospitals. Accessed June 30, 2016.

7. Miake-Lye I, Hempel S, Ganz DA, Shekelle PG. Impatient fall prevention programs as a patient safety strategy. A systematic review. Ann Intern Med. 2013;158:390–6.

8. Fleischauer AT, Ruhl L, Rhea S, Barnes E. Hospitalizations for endocarditis and associated Health care costs among persons with diagnosed drug dependence — North Carolina, 2010–2015. MMWR Morb Mortal Wkly Rep 2017;66:569–573. DOI:

9. Ronan MV, Herzig SJ. Hospitalizations related to opioid abuse/dependence and associated serious infections increased sharply, 2002-12. Health Aff (Millwood) 2016;35:832–7.

10. Wurcel AG, Anderson JE, Chui KK, Skinner S, Knox TA, Snydman DR, Stopka TJ. Increasing infectious endocarditis admissions among young people who inject drugs. Open Forum Infect Dis. 2016;3:ofw157.

11. Haber PS, Demirkol A, Lange K, Murnion B. Management of injecting drug users admitted to hospital. Lancet. 2009;374:1284–93.

12. The Joint Commission Online. November 2017 Perspectives Preview: Special Report: Suicide prevention in health care settings.
13. The Joint Commission Online. March 2019 Standards FAQ Details. Ligature and Suicide Risk Reduction — Video Monitoring of Patients at High Risk for Suicide. Can video monitoring/electronic-sitters be used to monitor patients at high risk for suicide?

An earlier version of this case study appeared in NEJM Catalyst on January 19, 2017.

Call for submissions:

Now inviting expert articles, longform articles, and case studies for peer review


A weekly email newsletter featuring the latest actionable ideas and practical innovations from NEJM Catalyst.

Learn More »

More From Patient Engagement
PROMs Usage Is Limited but Growing

Patient Engagement Buzz Survey: PROMs Use Is Growing, but Implementation Takes Effort

While clinical outcomes lend themselves to measurement, quantifying how an individual patient experiences symptoms or a loss of function can be more challenging. One promising approach is the use of Patient-Reported Outcome Measures (PROMs).

Taxonomy of the Patient Voice Table

Taxonomy of the Patient Voice

While health care pursues the important trend of putting patients at the center of care, the terms used to describe this goal are proliferating — and muddying the discourse. This taxonomy attempts to classify some of these terms and make some distinctions.

Agarwal01_pullquote handcrafting the patient experience

Handcrafting the Patient Experience

Health care organizations can take cues from consumer-facing companies like Airbnb to creatively insert convenience and surprise into patient encounters.

Screenshot of Tidepool daily diabetes data

A Taxonomy to Engage Patients: Objectives, Design, and Patient Activation

Health information initiatives will succeed only if they focus on patients’ motivation to engage and reflect the type of engagement they seek.

Ghafur01_pullquote -digital health health care consumer patient experience

Engaging Patients Using Digital Technology — Learning from Other Industries

Providers can benefit patients and disrupt health care by learning from the experience of other industries.

Health Care Providers Should Incentivize Patients

Survey Snapshot: Patient Financial Incentives — There Are No Quick Fixes

The NEJM Catalyst Insights Council agrees that while financial incentives are a common strategy to engage patients in healthy behaviors, they are not necessarily effective.

Support of Family and Friends Is More Effective Than Clinician Support in Realizing Health Goals - From the Patient Engagement Insights Report: Why No Single Health Incentive Works.

Patient Engagement Survey: Why No Single Health Incentive Works

Initiatives to improve patient engagement come in a variety of forms. While insurers, employers, and health care providers are all involved in using financial incentives and penalties for engagement efforts, improvement in health outcomes has been elusive. Achieving that ultimate goal will usually require a combination of financial and social approaches.

Health Systems Attending the Nudge Units in Health Care Symposium - Penn Medicine

Key Insights on Launching a Nudge Unit within a Health Care System

Leaders are finding that making higher-value choices easier through subtle changes to choice architecture can have an outsized impact on medical decision-making.

Barriers to Providing an Oustanding Patient Experience

Buzz Survey Report: Patient Experience

An independent NEJM Catalyst report sponsored by University of Utah Health on barriers to achieving an excellent patient experience.

The Patient Engagement Capacity Framework

The Patient Engagement Capacity Model: What Factors Determine a Patient’s Ability to Engage?

Patient engagement assessments often don’t dig deep enough to identify why patients don’t participate in their own health care. We present a new model to help providers pinpoint the reasons for lack of engagement and address them more effectively.


A weekly email newsletter featuring the latest actionable ideas and practical innovations from NEJM Catalyst.

Learn More »


Patient Incentives

75 Articles

Taxonomy of the Patient Voice

While health care pursues the important trend of putting patients at the center of care,…

Ripe for Disruption: Why and How…

For big tech companies like Amazon, Apple, and Google, the health care sector looks ripe…

Patients As Customers

141 Articles

Information Asymmetry: The Untapped Value of…

The knowledge and preferences that patients could — and should — share with clinicians would…

Insights Council

Have a voice. Join other health care leaders effecting change, shaping tomorrow.

Apply Now