Abstract: In healthcare facilities, patient falls have been a major contributing factor associated with patient injuries that result in increased costs and increased length of stay. Recent actions by the Department of Health and Human Services, enacted through the Department of Medicare and Medicaid Services, now hold healthcare facilities financially responsible for injury to patients that result from falls (CMS Hospital Acquired Conditions). Because of these rulings, costs associated with patient falls are now a greater threat to the survival of rural healthcare facilities. The rate of patient falls at Wise Regional Health System was unacceptably high. As a result of statistically analyzing the demographics of previous fall victims, and initiating additional interventions for identified patients, Wise Regional Health System has been successful in drastically decreasing their inpatient fall rate.
Healthcare delivery in Wise County Texas was very limited until 1973. At that time Decatur Community Hospital, a not for profit facility, opened its doors with 50 beds and two operating rooms. Creation of the hospital was supported by the city of Decatur, and public donors, to enhance community healthcare delivery to the residents of Wise County. The name was changed to Wise Regional Health System in 2001. To match population growth, and increase availability of local health care for residents, a new facility was constructed and bed size increased to 148 in 2004.
Wise County's population is aging, as is the general population. Patient demographics in rural settings include a disproportionate number of patients over the age of 65 and a disproportionate number of uninsured patients. Wise County estimates are 11% and 16%, respectively, as reported by the Texas State Data Center, with the closest urban market estimated at 8.4% over 65 and 12% uninsured. All hospitals are constantly working to reduce cost and maintain the quality and value of services to their patients. With a larger percentage of elderly and uninsured patients, rural facilities must focus greater attention on prevention of injuries to patients, which result in higher cost of care.
Assessing the Patient Population for Fall Risk
Wise Regional Health System nursing staff routinely assess all admissions for fall risk based on multiple factors incorporating the Hendrich Fall II Risk Assessment (Hendrich, Bender, & Nyhuis, 2003). The resulting fall risk score is used to establish appropriate fall interventions for the risk level assessed. These interventions range from applying a “smiley face” to the patient's door, identifying the patient as a fall risk, to assigning a full-time sitter.
The Patient Fall Prevention Committee monitors the outcome of interventions designed to prevent patient falls. Monthly statistics are presented as patient falls per 1,000 patient days. A benchmark of 2.4 falls per 1,000 patient days is the goal that the facility was striving for, but monthly statistics continued with rates as high as 5.6 falls per 1,000 patient days. The average fall rate for the period of January 1, 2008 to June 30, 2008 was 4.5 falls per 1,000 patient days. The outcome was 3–10 patient falls per month with a mean of 7.8. There were discussions of changing the benchmark from 2.4 to 4.5. References from the Texas Nurses Association for benchmark change were sited, as were references suggesting increased fall rates and injury expectations in the rural setting, due to patient demographics.
The Fall Prevention Committee decided that the fall rate had become a systemic long-term problem, due to the high incidence mentioned above, that needed to be evaluated and addressed aggressively. Improving patient safety, and the possible costs associated with treating patients who may be injured due to a fall, were the driving forces.
Fall Prevention Process Evaluation
The Quality Management Department was reluctant to change the benchmark for patient falls from 2.4 to 4.5 per 1,000 patient days. Rather than lower the standard, a decision was made to statistically evaluate demographic information collected from the medical records of all patients that had suffered a fall for the past 12 months. These data would be reviewed to determine if quality improvement indicators could be developed and applied to our patient population that would result in better identification of patients at risk for falls, allowing more proactive intervention.
All medical record demographic data that could enhance the description of and show possible correlation with a patient that had suffered a fall was collected.
The original dataset included:
•the Hendrich II Fall Risk assessment,
•the Braden Scale, which is routinely used to predict pressure ulcer risk, and is a good indicator of a patient's overall physical ability or limitations,
•the patient age,
•the patient sex,
•the location of the fall (bed, bathroom, bedside commode, etc.),
•the service caring for the patient,
•the diagnosis and/or treatment,
•number of medications,
•the number of days in the current hospitalization,
•the number of all previous hospital inpatient encounters, and
•the time of day.
Using NCSS Statistical Analysis and Graphic Software, the dataset was subjected to all possible regression analysis to reveal which subset of the data would best explain the regression formula when compared with the fall risk score assigned by nursing assessment. Analysis of variance (AOV) was applied to patient data determined most relevant by the all possible regression analysis. The AOV was expected to provide valuable insights into the demographics of our patients who fell.
During data analysis, patient sex was ruled out as a determining metric as there were an equal number of male and female patient falls. Medications were rejected as a metric as medication evaluation is included in the Hendrich II Fall Risk Assessment.
It has been reported, and generally accepted, that patients are at the greatest risk of suffering a fall during shift change. This was not the case in our facility. A review of the AOV histogram of fall times shows our patient fall rate is highest between 24:00 and 04:00 hours(see Figure 1), with additional increased falls between 20:00–24:00 and 08:30–12:00 hours. Shift change is at 07:00 and 19:00 hours, the times of minimum patient falls.
After evaluating the fall risk AOV (see Figure 2), there was concern that our patient assessment was flawed. Patients that had been assessed with a fall risk between 10 and 13 were suffering the majority of falls. The Hendrich II Fall Risk Assessment is reported to have a validity of 98% (Department of Defense, 2005). We compared our staff patient assessments to predicted fall risk values, using the collected dataset. Predicted values for fall risk scores, from the correlation formula, were generated, using NCSS software. It was determined, from the predictive analysis, that the assessment of the nursing services for patient fall risk score using the Hendrich II Fall Risk Assessment was valid. The staff assigned fall risks were within 2 standard deviations (SD) of the predicted fall risk for 89% of the data tested and 94% within 3 SD.
The interpretation that patient care services are doing a good job of assessing patients for fall risk and preventing falls of patients with the higher risk scores was also valid. The majority of the patient falls were suffered by patients with a mid-range fall risk scores of 10–13, not the identified very high-risk patients.
From the dataset, the Fall Risk Score, the Braden Scale, the number of times the patient had had inpatient contact with the facility, the age of the patient and the number of days the patient had been in the hospital displayed the best correlation with patient falls. These metrics were selected to develop a fall patient identification model.
By evaluating AOV for these metrics, a scenario was developed. The patients who suffered the highest number of falls, are in their sixties (see Figure 3), have had limited contact with the facility (see Figure 4), are early in their current admission (see Figure 5), have a high Braden Scale (see Figure 6) and a Fall Risk Score of 10–13, as presented earlier. Patients who are in their 60s and have had limited contact with our facility are routinely admitted from a self-sufficient social environment. These patients have been independent, are not accustomed to being assisted by others in their activities of daily living and less likely to seek assistance, even when instructed to do so by attending nurses.
The times associated with patient falls, and the fact that most falls occur in attempts to go from the bed to the bathroom support this theory. These patients are rising for nighttime and morning toileting without calling nursing service for assistance.
Fall Prevention Process Improvement
The nursing services of our facility continue to follow established prevention methods for all high-risk patients. An additional layer of patient identification and fall prevention tools was implemented, and additional focus placed on fall prevention training and staff involvement.
•Toileting rounds were established.
•Walking reports at change of shift were initiated.
•Run Charts, specific for each service, were incorporated into fall prevention posters and distributed monthly to increase staff awareness.
•Enhanced fall prevention training was included in nursing orientation with the development of a new PowerPoint presentation.
•Documentation of involvement of the patient and family in fall prevention, patient safety, and involvement in their own care was developed (Ward, Candela, & Mahoney, 2004).
•Additional signage was placed in the patient's room. A “Yield” sign instructing the patient to call for assistance, reinforcing safety and involvement, was prominently displayed near the patient's bed (Ward et al., 2004).
•A method of assigning relative fall risk based on the patient scenario presented earlier was developed. Twice a week, patient data from the census of the main Medical/Surgical services is evaluated. A relative fall ranking is assigned from highest to lowest and forwarded to each service director. The director then uses the information, coupled with patient evaluation, to increase fall interventions.
•Investigation of each new patient fall, following each occurrence, was initiated. The unit director and risk management meet after each fall to determine “what could have prevented this patient fall,” and generate a fall investigation report.
•Fall investigation reports are presented at each Fall Prevention Meeting for education and discussion.
The evaluation and improvement process resulted in reduced patient falls for the last 3-month period, July through September, 2008. Patient falls/1,000 patient days have steadily decreased, from 4.37 falls per 1,000 patient days, in July, to 1.29 falls per 1,000 patient days, in August, to 0 falls per 1,000 patient days, for the last reporting period, September, 2008.
By using all statistical data available to develop a demographic picture of the patient population in question, focusing efforts on patient monitoring, involvement of the patient in his/her care, and increased involvement of the caregiver staff, a dramatic decrease in patient falls can be accomplished.
Sustaining the improvement made at Wise Regional Health System will be a challenge and require a change in our culture with increased and continued patient and staff involvement.
With increasing costs of operation and the changes in reimbursement for costs related to CMS identified Hospital Acquired Conditions, all facilities must develop and improve their processes to prevent patient falls and increase the safety of those in their charge.
Larry Wayland, MHA, CPHQ, is the quality improvement coordinator for Wise Regional Health System, Decatur, TX. Larry assists in improvement activities by mediation and statistical evaluation of current and proposed processes.
Lynn Holt, LVN, is the administrative director of physician relations, quality management director for Wise Regional Health System, Decatur, TX. Lynn reviews and approves quality improvement activities for all services and insures regulatory data are collected, evaluated, and submitted.
Sue Sewell, MSN, is the Chief Nursing Officer for Wise Regional Health System, Decatur, TX. Sue monitors the activities of the nursing staff at Wise Regional Health System and works to improve services to the patient.
Joan Bird, CMSRN, is the 5th floor director of medical/surgical service of Wise Regional Health System, Decatur, TX. Joan manages the 5th Floor medical/surgical service and is active in implementing and documenting the results of newly identified opportunities for improvement.
Laurie Edelman, RN, is the 4th floor director of Medical/Surgical services of Wise Regional Health System, Decatur, TX. Laurie is a member of several Quality Improvement teams.
By participating in this independent study offering, the reader will be able to
1Identify methods of evaluating demographic criteria used in determining Fall Risk in a rural hospital population.
2Develop a monitoring program based on facility-specific statistics for tracking patient falls and devise a program of interventions based on these findings.
3Recommend ways to improve effectiveness through inter-departmental involvement in Fall Prevention.
1Healthcare facilities nation-wide are now held responsible for injury to patients that result from falls, due to a recent ruling from which agency?
a.Texas Hospital Association
b.The Joint Commission
c.American Medical Association
d.Department of Health and Human Services
2Demographic information collected from a Patient's medical record that was statistically relevant in determining the individual's fall risk included:
a.Braden scale, patient age, diet order, time of day.
b.Patient age, patient sex, number of days in admission, number of visitors.
c.Location of fall, time of day, Hendrich II Fall Risk Assessment, Patient age.
d.Medications, ordered treatments, service caring for patient, origination of admission (ER vs. MD office).
3According to the article, the majority of falls occurring in this facility were documented:
a.Between 59 and 69 years of age.
b.Late in the hospitalization.
c.At a low Braden scale ranking.
d.At a low Fall Risk Score.
4A committee assigned to monitor, evaluate, track occurrences, and make recommendations on Patient Falls would be:
a.Nurse Manager's Committee
b.Patient Fall Prevention Committee
5Interventions that have proved successful in deterring patient falls include the following:
a.Sad face stickers on Patient's door.
b.Part-time sitter in Patient room.
d.Eliminating discussion on fall risks in employee orientation.
6Shift change, generally accepted as the time of day for greatest risk for Patient falls, was ruled out in this study. The time of day identified by the AOV histogram (analysis of variance) which included the most falls in this facility was:
7A dramatic decrease in Patient falls can be accomplished by all of the following except:
a.Not involving the Patient in his/her care.
b.Using all statistical data available to form a Patient demographic.
c.Focusing efforts on Patient monitoring.
d.Increasing involvement of health care staff.
8In the article, Wise Regional Health System saw the following results in their process improvement:
a.Patient falls/1,000 days increased over a 4 month period.
b.Patient falls/1,000 days increased over a 2 month period, and then leveled off.
c.Patient falls/1,000 days decreased over a 3 month period.
d.Patient falls/1,000 days stayed the same over a 3 month period.
9Injury prevention and thus healthcare costs are becoming increasingly important in rural facilities due to:
a.Larger number of patients age 35–50.
b.Larger number of patients age 50–60.
c.Larger percentage of uninsured patients.
d.Larger pediatric population.
10In the study, it was determined that the majority of falls in this facility occurred when Patients attempted nighttime and morning toileting without calling for assistance. This was addressed by:
a.Staff spending more time in the Patient's room.
b.Establishing toileting rounds and educating staff.
c.Having different speakers at Fall Prevention meetings.
d.Asking relatives and friends to spend the night in the Patient's room.