Journal of Gerontological Nursing

Why Do They Fall? MONITORING RISK FACTORS in Nursing Homes

Yukie T Gross, MS, RN; Yoshiko Shimamoto, PhD, RN; Charles L Rose, PhD; Beth Frank, BS, RN

Abstract

With the rapid increase in the number of frail elderly, the problem of falls has taken on everincreasing importance. The elderly are especially prone to falls because of age-related host factors such as muscle weakness, gait and balance, and mental deficits.1 The expansion of nursing homes in response to the burgeoning number of elderly, and further fueled by public funding, has brought the problem of falls increasingly into the nursing home. This has recently been exacerbated by federal cost containment legislation associated with the diagnosis related groups of the prospective payment system. Such provisions are aimed at decreasing hospital stays but result in increasing' acuity in nursing homes, thereby magnifying the problem of falls in these institutions. As a protected environment, the nursing home is expected to keep falls at a minimum and obviate legal liability. The issue becomes one of assessing risk factors as a base for remedial steps toward reduction and prevention of falls.

The risk factors for falls have been well-recognized in general terms. These risk factors are of two kinds: host factors, already referred to, and agent factors. The latter include a broad range of environmental hazards from architectural sources, such as lighting, furnishings, clutter, and improper or missing safety and assistive devices; staff and management problems; and problems arising from medical regimens. However, host and agent characteristics differ by setting, whether domestic, hospital, or institutional. In addition, there are host and agent differences within the institutional settings. Rirthermore, changes in resident populations and agent factors occur over time. Therefore, the monitoring of risk factors must be continuous and setting-specific, so that remedial interventions may be adjusted accordingly. This article reports a pilot project for a larger study with the above objectives in mind.

This study reports on falls in a Hawaii nursing home based on secondary data sources, ie, incident reports and available institution-wide data. The purpose of this initial approach was to determine the efficacy of such data for determining risk factors in particular settings. This would lead to the design of an improved data base that is also cost effective and feasible within existing resources of the institution.

The literature on risk of falls is largely based on hospital settings. Within hospital settings, there are differences in findings on risk factors that reflect variation in population, environment, and the pervasive variation in study procedures. There are even greater discrepancies when hospital findings are compared with available nursing home studies. Core risk factors are common to all fallers regardless of setting; however, any remedial action in a particular setting should be tailored to risk knowledge empirically derived nom that institution.

Another problem is that sometimes elaborate methodologies and primary data collection are expended on a onetime basis to determine risk factors with little consideration to the need for parsimonious monitoring as circumstances change. In addition, three other issues emerge: the use of a control group of nonfallers for the identification of risk factors; the special problems of the patient who falls repeatedly; and the design of prevention procedures based on knowledge of onsite risk factors as well as the testing of the efficacy of such remedial action. This pilot study and its subsequent phases incorporate in their framework the issues gleaned from this background review.

METHOD

The study was based on incidents reported over a 12-month period in a 178-bed nursing home composed of 43 skilled nursing and 135 intermediate care beds in Honolulu. The staffing consisted of 13 RNs, 12 LPNs, and 68 aides. The following variables were coded from the incident reports: gender, age, sex, medical diagnosis, mental sensory and mobility status,…

With the rapid increase in the number of frail elderly, the problem of falls has taken on everincreasing importance. The elderly are especially prone to falls because of age-related host factors such as muscle weakness, gait and balance, and mental deficits.1 The expansion of nursing homes in response to the burgeoning number of elderly, and further fueled by public funding, has brought the problem of falls increasingly into the nursing home. This has recently been exacerbated by federal cost containment legislation associated with the diagnosis related groups of the prospective payment system. Such provisions are aimed at decreasing hospital stays but result in increasing' acuity in nursing homes, thereby magnifying the problem of falls in these institutions. As a protected environment, the nursing home is expected to keep falls at a minimum and obviate legal liability. The issue becomes one of assessing risk factors as a base for remedial steps toward reduction and prevention of falls.

The risk factors for falls have been well-recognized in general terms. These risk factors are of two kinds: host factors, already referred to, and agent factors. The latter include a broad range of environmental hazards from architectural sources, such as lighting, furnishings, clutter, and improper or missing safety and assistive devices; staff and management problems; and problems arising from medical regimens. However, host and agent characteristics differ by setting, whether domestic, hospital, or institutional. In addition, there are host and agent differences within the institutional settings. Rirthermore, changes in resident populations and agent factors occur over time. Therefore, the monitoring of risk factors must be continuous and setting-specific, so that remedial interventions may be adjusted accordingly. This article reports a pilot project for a larger study with the above objectives in mind.

This study reports on falls in a Hawaii nursing home based on secondary data sources, ie, incident reports and available institution-wide data. The purpose of this initial approach was to determine the efficacy of such data for determining risk factors in particular settings. This would lead to the design of an improved data base that is also cost effective and feasible within existing resources of the institution.

The literature on risk of falls is largely based on hospital settings. Within hospital settings, there are differences in findings on risk factors that reflect variation in population, environment, and the pervasive variation in study procedures. There are even greater discrepancies when hospital findings are compared with available nursing home studies. Core risk factors are common to all fallers regardless of setting; however, any remedial action in a particular setting should be tailored to risk knowledge empirically derived nom that institution.

Another problem is that sometimes elaborate methodologies and primary data collection are expended on a onetime basis to determine risk factors with little consideration to the need for parsimonious monitoring as circumstances change. In addition, three other issues emerge: the use of a control group of nonfallers for the identification of risk factors; the special problems of the patient who falls repeatedly; and the design of prevention procedures based on knowledge of onsite risk factors as well as the testing of the efficacy of such remedial action. This pilot study and its subsequent phases incorporate in their framework the issues gleaned from this background review.

METHOD

The study was based on incidents reported over a 12-month period in a 178-bed nursing home composed of 43 skilled nursing and 135 intermediate care beds in Honolulu. The staffing consisted of 13 RNs, 12 LPNs, and 68 aides. The following variables were coded from the incident reports: gender, age, sex, medical diagnosis, mental sensory and mobility status, and activities of daily living (ADL). Other variables related to falls were length of stay at the time of the incident, time (day, month, hour), location, the type of fall (ie, walking, from bed, chair, commode, or toilet), whether tails occurred during transfer, and the assistive and safety devices used. Data on the total nursing home population were retrieved from computerized institutional records and summaries of daily nursing assessments of patients' ADL and mental status.

Table

TABLE 1COMPARISON OF FALLERS WITH OVERALL INSTITUTIONAL POPULATION*

TABLE 1

COMPARISON OF FALLERS WITH OVERALL INSTITUTIONAL POPULATION*

FINDINGS

During the study year, 1 IS incidents were reported, 40 of which (35%) were falls involving 29 patients. The most common non-fall incidents were patient's assaultive behavior, medication errors, and wandering off the premises. Thus, the 40 falls which occurred in 178 patients, the average patient census over the year, yielded a falls rate of 22.5%. On the basis of the number of patients who fell (29), the rate was 16.3%. Thus, 29 patients were involved in 40 falls. Seven (24. 1%) of the 29 fallers fell more than once accounting for 18 falls or 2.6 falls per multiple faller.

The following sections compare the characteristics of patients involved in falls with those of the total nursing home population (Table 1 ); factors associated with falling incidents (Table 2); factors associated with repetition (Table 3); data from the institutional population, single falls, and repeated falls (Table 4).

Risk Factors

As shown in Table 1 , the mean age of the fallers was 81.9 years, slightly higher than the average age of 80 years in the institutionalized population. The risk factor of age has been consistently documented.2 There was a higher percentage of men who fell (5 1 . 7%) when compared with the total percentage of men in the overall institutional population- (32.9%). When the fallers were compared with nonfallers (the overall population minus the fallers), the percentage of males in the nonfallers was still lower (36%; X2 = 3.74; p<.05).

The major diagnoses of fallers were cardiovascular and cerebrovascular accidents (65.5%), hypertension (44.8%), organic brain syndrome (37.9%), and diabetes mellitus (24.1%). These were generally higher than in the overall population, with a greater difference for hypertension.

Table

TABLE 2FACTORS ASSOCIATED WITH FALLING INCIDENTS*

TABLE 2

FACTORS ASSOCIATED WITH FALLING INCIDENTS*

Sixty-nine percent of the patients had some mental impairment: 27.6% were disoriented; 10% were senile; and a catchall group of 31% were reported as confused, forgetful, or demented. The proportion of impaired patients in the total population was substantially less (31.3%). This finding is also consistent with the literature.3-5

Wheelchairs were the only form of mobility for 45% of the fallers, which was similar to the overall nursing home population (49%). Of the 40 falls, a wheelchair was used as an assistive device in 70% of the falls. Lund and Sheafor found that assistive devices, including wheelchairs, walkers, and canes, increased the risk of falling.3 Overall population data on the general use of wheelchairs were not available.

The additive ADL scores of fallers and the overall population were similar.

The mean length of stay associated with falls was considerably shorter than that of the overall population (11.1 months versus 20 months). A number of hospital studies found that most falls occurred within the first week of hospitalization,6·7 which was attributed to patients' confusion of being in a new environment. In this study, the patients who stayed longer were also more apt to be confined to none of the fallers were bedfast. shorter stay may be a risk because patients staying longer more apt to be confined to bed and therefore at less risk for falling.

Tinetti has also speculated that disabilities may be well comfor and less hazardous than a disability to which the elderly has not yet adapted.7 This, may be less relevant for nurshome patients who, upon admisalready suffer from chronic disaYet, one cannot rule out that in cases, recently admitted patients be less familiar with and more in their new environment. In staff may be less familiar the idiosyncrasies of recently adpatients.

Other Factors

Three fourths of the falls took place the day between 8:00 AM and PM with peaks at 10:00 AM and It is not surprising that most falls place during the day when there more patient activities. The peaks mid-morning and noon are break for staff, while those not on breaks are busy feeding patients. Lund Sheafor attributed the highest freOf falls by month and hour to newly graduated nurses who assumed and night duty.3 Other studies that most falls occurred during the day when staffing density and assopatient activity was highest.8,9

Three fourths of the falls occurred in patients' rooms, which was also supported by other studies.3,4,9 One explanation is that although patients are in their rooms, where much of their waking time is spent, they are less apt to be by staff. This is suggested by the finding in this and other studies that staff members were not present in 85% of the falling incidents.8,9 This was formally tested by a crosstabulation between falls taking place in the patients' rooms and the presence of staff. No relationship was found, however, which weakens the notion that absence of staff observation is a factor 4 in falls that occur in the patient's room.

In a very few cases, falls occurred when restraints were ordered but not used (7.5%). However, patients fell despite the use of restraints in 47.5% of the instances, either from loosening the restraints or because of ineffective restraints. A hospital study found no difference in the use of restraints between fallers and nonfallers.3

Table

TABLE 3FACTORS ASSOCIATED WITH REPEATED FALLING*

TABLE 3

FACTORS ASSOCIATED WITH REPEATED FALLING*

There were no injuries in a majority of the falls (57.5%). The type and incidences of injury (including multiple injuries) in the remainder was as follows: lacerations, 25%; swelling, 22.5%; fractures, 10%; and ecchymoses, 7.5%. This distribution is similar to that found in other nursing home studies.8,9 Fractures appear to be the most serious injury; although they occurred in only 10% of all falls, they were the result in 47% of the falls where injuries occurred (8 of 17 falls). Although the physician was called in 95% of the incidents, he only attended 15% of the time. It may be inferred that he attended all the fracture cases (10%) in addition to another 5% where the injuries were less serious. This suggests that on-site nurses were accurate in their assessments for the need of physician attendance.

An analysis of the relationship of age to occurrence of injury revealed that older patients were more apt to have an injury: fallers with no injuries had a mean age of 79. 1 years, whereas injured fallers had a mean age 86.2 years (F= 11.9,P=. 001).

Repeated Falls

Multiple fallers (Table 3) are of special interest in the investigation of risk factors for falling since they represent patients who are even more prone to falling than single fallers.6,10

The risk factors of falling were compared with the risk factors of repeated falls (Tables 1 and 3). The base for this analysis was the number of falls (40) rather than the number of fallers (29). This was because some of the possible risk factors varied with the incident, such as mental status and location of fall, so they could not have been picked up with a "faller" base. In addition, the faller base would have markedly reduced the statistical adequacy of the number. Organic brain syndrome, mental impairment, and snorter length of stay were significant risk factors in both analyses. ADL was not significant in either analysis. Age, male sex, and hypertension were significant factors in falls but not in repetition of falls. Incontinence was less prevalent in falls but not in repeated falls. There were more cerebrovascular accidents among patients who fall than in the general nursing home population, but there were more wheelchair-bound patients involved in single than repeated falls. In fact, none of the repeated falls occurred in wheelchairbound patients.

Finally, trends in risk factors were examined by a three-way comparison of institutional data, single falls, and multiple falls (Table 4). Progressive increase in risk was found in proportion to male sex, organic brain syndrome, hypertension, incontinence, and mental impairment; progressive decrease in risk was found in incontinence and length of stay. There was also a variation from a linear pattern.

Table

TABLE 4A COMPARISON OF SELECTED CHARACTERISTICS OF THE OVERALL POPULATION, SINGLE FALLS, AND MULTIPLE FALLS

TABLE 4

A COMPARISON OF SELECTED CHARACTERISTICS OF THE OVERALL POPULATION, SINGLE FALLS, AND MULTIPLE FALLS

Age was greater in single falls than in the overall institutional population, but there was no such relationship comparing single with multiple falls. None of the wheelchair-bound patients fell more than once, but roughly half of those who fell once and likewise of the general patient population were wheelchair-bound. There were no differences in the three groups for ADL score. With respect to cerebrovascular accidents, multiple falls had least, the overall population somewhat more, and single falls had the most. The findings are particularly important for decreasing or preventing falls, when one considers that cerebrovascular accident was found in about half of all falls in this study.

DISCUSSION

This study was undertaken as a pilot project for a more extensive study based on a larger number of nursing homes. This will involve development of a data base sensitive to institutional differences and change for the design of a nursing intervention for the prevention of falls.

The planned research focuses on the nursing home since the majority of care of the frail elderly occurs in this setting. The literature, on the other hand, has focused more on the hospital setting. This can be understood in view of the greater incidence of falls in hospitals, which is partly due to the greater mobility of hospital patients. An example of this differential may be gleaned from a comparison of fall rates in one hospital with the fall rate found in the present study. As reported in an Executive Summary,11 907 falls occurred over 13 months in a largely geriatric, 810-bed unit in the Montefiore Hospital, New York. Prorating to 12 months, the fall rate was 1.12 falls per bed per year. By comparison, the present nursing home study yielded a fall rate of 0.22 falls per bed per year, or one fifth of the hospital rate. The rationale for monitoring and remediation of falls proposed for nursing homes would clearly apply to hospitals, which are even more vulnerable.

Archived data in this pilot study proved useful for identifying the risk of falls and, in particular, clarified the relationship between nonfallers, single falls, and repeated falls.

This study identified the need for amplifying the incident reports to serve as a basis for a falls assessment tool. Additions included the acuity level of patients; refinement of the mental impairment measure to indicate increasing stages of dementia, gait, and balance data; re-coding of sensory and mobility data; and history of previous falls. A particularly interesting finding was the higher risk of falling in patients with greater mobility. This helps explain the greater rate in hospitals, where patients are generally more mobile. By the same token, one would expect a greater incidence of falls in community dwellers of comparable frailty.

The association of dementia with falls takes on special significance in view of the increasing number of dementia patients in nursing homes. This pinpoints the need for special surveillance of these patients for the prevention of falls.

The greater incidence of falls in men may be related to a number of factors. Men are more apt to have cerebrovascular accidents (men 73.7%, women 38. 1 %; x2 = 3.77; p = .05) and to be mentally impaired (men 42.1%, women 9.5%; x2 = 4.04). Both cerebrovascular accident and mental impairment were found to be risk factors for falling.

The finding that use of restraints does not prevent falls should discourage the use of restraints for that purpose. In any case, patients (as well as those who are not patients) react negatively to being restrained. Avoidance of restraints is in line with nursing goals that aim to foster independence.

Aside from the fact more falls occur during the day than at night, the peak times of daytime falls coincided with nursing personnel being busy with feeding and dispensing medications, and also with break times. This suggests a number of actions: me use of families and volunteers to assist with feeding, modification of medication schedules, and the staggering of breaks. However, it cannot be assumed that availability of staff is sufficient for minimizing falls. Inservice - education is also needed for staff to anticipate any behaviors that may lead to falls and take steps to prevent them.

This study suggests that archived data are a feasible base for developing risk factor data. The next steps are development of a falls assessment tool based on a modest expansion of archived data coupled with the development of an intervention for reduction and prevention of falls. Both risk factors and remediation should be monitored because of changes in the environment, staffing, and patient characteristics of nursing homes. Such changes may also be affected by societal changes in the older population toward more acuity and by changes in long-term care financing. Finally, it is proposed that those procedures be replicated in a number of nursing homes in Hawaii to determine the extent to which institution-specific risk factors and remediation vary from institution to institution, and to determine their differences in patterns of change. This may then provide a model for monitoring risk factors and developing a preventive intervention based on those factors that are setting and time sensitive.

REFERENCES

  • 1. Rubenstein ZL, Robbins AS, Schulman BL. et al. Falls and instability in the elderly. J AmGeriatrSoc. 1988;36(3):266-278.
  • 2. Janken JK, Reynolds BA, Swiech K. Patient falls in the acute care setting: Identifying risk factors. Nurs Res. 1986; 35(4):215-219.
  • 3. Lund C, Sheafor ML. Is your patient about to fall? Journal of Gerontological Nursing. 1985; 11(4):37-4I.
  • 4. Schested P, Severin-Nielsen T. Falls by hospitalized elderly patients; causes, prevention. Geriatrics. 1977; 32(4): 101-108.
  • 5. Hernandez M, Miller J. How to reduce falls. GeriatrNurs. 1986; (3,4):97- 102.
  • 6. Catchen H. Repeaters: Inpatient accidents among the hospitalized elderly. Gerontologist. 1983; 23(3):273-276.
  • 7. Tinetti ME, Williams TF, Mayewski R. Fall risk index for elderly patients based on number of chronic disabilities. Am J Med. 1986; 80(3):429-434.
  • 8. Morse JM, Prowse MD, Morrow N, et al. A retrospective analysis of patient falls. Can J Public Health. 1985; 76(3,4):1 16-118.
  • 9. Pablo RY. Patient accidents in a long term care facility. Can J Public Health. 1977; 68(5,6):237-247.
  • 10. Morse JM, Tylko SJ, Dixon HA. The patient who falls - And falls again. Journal of Gerontological Nursing. 1985; 11(11): 15-18.
  • 11. Jackson BS, Krasnoff L, Regan B, et al. Executive Summary Patient Slip/Falls: Evaluation and Prevention. Springfield, Va: National Technical Information Service; 1986:1-17.

TABLE 1

COMPARISON OF FALLERS WITH OVERALL INSTITUTIONAL POPULATION*

TABLE 2

FACTORS ASSOCIATED WITH FALLING INCIDENTS*

TABLE 3

FACTORS ASSOCIATED WITH REPEATED FALLING*

TABLE 4

A COMPARISON OF SELECTED CHARACTERISTICS OF THE OVERALL POPULATION, SINGLE FALLS, AND MULTIPLE FALLS

10.3928/0098-9134-19900601-08

Sign up to receive

Journal E-contents