Predictive Analytics and Computer Visualization Enhances Patient Safety to Prevent Falls
- Conditions
- Falls and Falls With Injury
- Interventions
- Other: Fall prevention algorithmOther: Inspiren camera visualization
- Registration Number
- NCT06339125
- Lead Sponsor
- Massachusetts General Hospital
- Brief Summary
Annually, in the United States there are 700,000 - 1,000,000 inpatient falls reported, and one-third of patients sustain an injury. The average estimated cost per fall is $6,694, resulting in over $1.4 -1.9 billion dollars in losses each year (AHRQ, 2017). This study aims to compare the impact of different fall prevention strategies on the rate of occurrence of falls and falls with injury in an academic medical center on three adult medical units. While maintaining the usual standard of care for fall prevention, each unit will add one of the following: (1) use of a fall risk alert to nurses using an algorithm based on electronic health record data or (2) computerized camera visualization or (3) a combination of both.
- Detailed Description
To decrease falls in the hospital setting, and building on previous nursing fall research, as well as the MFS and the Fall TIPS program, MGH developed a decision support algorithm to identify changes in clinical factors as they occur to alert nurses to the need to adjust fall prevention interventions. MGH Nursing, through a collaboration with RGI Informatics, then deployed the MGH algorithm on one clinical general care unit. The RGI software uses the MGH algorithm live streaming EHR data from Epic to identify patients whose risk of falling may have increased and provide clinical decision support to nurses through an alert on their hospital issued cell phones. Preliminary results demonstrated feasibility and a statistically significant reduction (p \<0.01) in falls with injury over an 11-month period.
Mutually exclusive preliminary work, on a second inpatient general care unit, involving a computerized patient visualization system also yielded reduction in falls. Combined usage of the two technologies may yield a synergistic effect thereby further reducing the incidence of falls in the acute care setting. To date, there is no evidence derived from evaluation of patient outcomes from simultaneous testing of the two technologies. Thus, the purpose of this study is to determine the impact of three different fall prevention interventions (RGI/MGH Algorithm only, Inspiren only and combined RGI/MGH Algorithm and Inspiren) on patients at risk for falls and falls with injury on three adult general care units in a large academic medical center.
Our proposed solution is the only known strategy that extracts and synthesizes physiologic and physical data from multiple sources, to create a dimensional view of a patient's safety profile related to fall risk. Timely alerts will inform nurses of patient's fall risk, reason for risk and their clinical decisions regarding fall prevention strategies. This initial proposal focuses on patients at risk for falls and we are confident that this innovative approach is adaptable to address other critical safety issues for example, pressure injuries and catheter associated urinary tract infections. Detailed information about RGI Analytics and Inspiren is provided below.
Methodology: An observational cohort, mixed-methods study design will be conducted to determine the impact and effectiveness of usual care and three different fall prevention strategies that exceed the standard of care on three inpatient units at MGH over one year. Unit 1 will employee streaming analytics and the MGH algorithm only, Unit 2 will employee Inspiren's AUGI computer visualization only and Unit 3 will employee the combined streaming analytic/MGH algorithm and Inspiren's AUGI device. Unit 4, the control unit, will serve as an internal comparison group from the same institution. In addition to the study interventions all four units will continue to maintain usual MGH evidence-based practice, standards of care for fall prevention.Patient, unit, and nurse demographic data collected for the study currently can be accessed from or calculated from existing sources. Sources include the ADT, PCS financial, acuity, and quality data stored in the PCS Datawarehouse. Unit patient demographic data in the aggregate will include age, gender, and race. Nurse demographic data will include the number of fulltime equivalents, years of experience as a nurse, years of experience at MGH, and highest level of education. Unit data will include counts of patient admissions, patient days, length of stay, nursing acuity, patient type by gender, age, race, ethnicity, number of unit falls and unit falls with injuries, and nurse staffing indicators. Nurse perceptions of the three interventions units will be measured in association with the intervention using real time feedback from cell phone alerts (helpful/not helpful), nurse feedback, and quarterly surveys. The Fall Prevention Efficiency Scale (Dykes, et al., 2021) is a peer reviewed 13-item tool that focuses on four key areas: saves time, does not waste time, is worth the time and is helpful in preventing falls. The survey questions will be adapted to meet the needs of this study and will be administered via REDCap, a Harvard Catalyst secure, web application for managing on-line survey tools.
Research questions
1. In the acute care, inpatient hospital setting, is there a difference in rate of occurrence of falls and injurious falls, comparing three distinct methods of alerting nurses at the point of care to a change in a patients risk of falling while maintaining all other current standards of care for fall prevention and adding these new standards during the study: (1) use of streaming analytics and a fall risk algorithm that alerts nurses to a change in fall risk, (2) computer visualization and artificial intelligence interpretation of patient movement and (3) a combination of both technologies?
2. What are the perceptions of nurses related to:
1. The impact of three study technologies implemented to assist with the identification of increased fall risk.
2. The reduction of nurse burden on the assessment of fall risk and the recommendation for additional interventions to prevent falls.
Research aims:
1. Compare the impact of the three fall prevention innovations, within and between units and to one control unit (all four units using same usual standard of care) on falls and falls with injury.
2. Determine the perceived effectiveness of fall prevention innovations and alerts on clinical decision support and nurse burden using nurse surveys, responses to alerts and focus groups.
Recruitment & Eligibility
- Status
- NOT_YET_RECRUITING
- Sex
- All
- Target Recruitment
- 4500
Adult medical patients admitted to the study units. All nurses working on the study units.
- None
Study & Design
- Study Type
- INTERVENTIONAL
- Study Design
- PARALLEL
- Arm && Interventions
Group Intervention Description Unit 3 Inspiren camera visualization Usual care and live streaming electronic health record driven Algorithm alerts nurses to possible increase in fall risk for review of interventions in place. AND Computer camera visualization detects and anticipates patient movement for patients at risk for falls and alerts nurses with fall risk potential. Unit 1 Fall prevention algorithm Usual care and live streaming electronic health record driven Algorithm alerts nurses to possible increase in fall risk for review of interventions in place. Unit 3 Fall prevention algorithm Usual care and live streaming electronic health record driven Algorithm alerts nurses to possible increase in fall risk for review of interventions in place. AND Computer camera visualization detects and anticipates patient movement for patients at risk for falls and alerts nurses with fall risk potential. Unit 2 Inspiren camera visualization Usual care and computer camera visualization detects and anticipates patient movement for patients at risk for falls and alerts nurses with fall risk potential.
- Primary Outcome Measures
Name Time Method Falls Measured monthly/quarterly over one year Rate of patient falls per 1000 patient days
Falls with injury Measured monthly/quarterly over one year Rate of falls with injury per 1000 patient days
- Secondary Outcome Measures
Name Time Method Nurse perceptions three, six, nine and twelve months Focus groups of nurse perceptions