A prospective study to identify the fall-prone patient

Sm. Sri. Med. Vol. 28, No. I, pp. 81-86. 1989 Printedin Great Britain.All rightsreserved

Copyrifit c

A PROSPECTIVE STUDY TO IDENTIFY FALL-PRONE PATIENT .JANICE M. MORSE,‘*

COLLEEN BLACK,’ KATHLEEN OBERLE’

and

0277-9536189 53.00 + 0.00 1989 PergamonPressplc

THE

PATRICIA DONAHUE”

‘Professor of Nursing and National Research Scholar, University of Alberta, *Director of Nursing, Charles Camsell General Hospital, Edmonton, Alberta T5M 3A4, ‘Director of Research, University of Alberta Hospitals, Edmonton, Alberta T6G 2B7 and 4Research Assistant, Faculty of Nursing, University of Alberta, Edmonton, Alberta T6G 2G3, Canada Abstract-A prospective study in 16 patient care units in two institutions was conducted to assess patients’ risk of falling. Three types of patient care units were utilized (acute medical and surgical, long term geriatric and rehabilitation areas), with 2689 patients over a 4-month period rated daily for fall risk using the Morse Full Scale. Differences in mean fall score were evident between the acute care areas and the long term and the rehabilitation areas, Examination of the scores in the acute care institution by length of stay, showed different patterns of fall risk; the mean score of the long term patients showed less variation and higher scores. Patients’ falls were analyzed by fall score and type of fall. All types of falls (anticipated physiological, unanticipated physiological and accidental), and the severity of injuries increased, with increasing scores indicating clinical validity of the scale. Key words-patient

falls, accidents,

risk assessment

have focused on investigating a single contributing cause and, with few exceptions, have not used control groups. There is increasing evidence to show that patients who fall are likely to fall again. In a study of repeated fallers and matched controls, Morse et al. [25] found that 55% of the repeated falls occurred while the patient was doing the same activity. Gryfe et al. [26] note that falls tend to ‘cluster’ prior to the patient’s death, indicating increasing frailty. In another study, Morse and her colleagues [20] examined 100 patients who fell, and compared these patients with 100 randomized controls. In this study, discriminant analysis revealed significant differences between the fallers and the controls using six significant variables. The weights derived from this analysis were then used to construct a scale [27] to identify the patient at risk of falling. The scale was validated by computer modelling on a normalized hospital population [28]. As such, the Morse Fall Scale was developed and provided a technique for the rapid identification of fall risk [25]. Examination of the results reveals that there are three types of patient falls. First, the scale identifies Anticipated PhysioIogical Falls, falls that occur when the patient has difficulty ambulating and is confused (i.e. 78% of falls). The other two types of falls (not identified by the scale) are Unanticipated Physiological Falls that occur when the patient has a ‘drop attack’, faints or has a seizure, and Accidental Falls, that occur when the patient trips or slips [20]. Each type of fall has distinct characteristics and particular strategies for intervention. The purpose of this study was to clinically validate the Morse Fall Scale in three types of patient care areas (acute medical and surgical units, long term care areas and a rehabilitation hospital). Patients’ fall risk was rated daily and falls that occurred were analyzed by type of fall and risk score to determine the feasibility of using the scale in practice.

INTRODUCI’ION cost of patient falls has been well documented. Mortality statistics indicate that falls are the principle cause of accidental death in those over 65 years, and one-tenth of those falls occur in hospitals or homes for the aged [l]. In the U.S., it has been estimated that patients who fall and suffer a fractured femur cost the health care system billions of dollars annually [2]. After a fall, even if the patient is not seriously injured, fear of a repeated fall may severely interfere with the patient’s quality of life, and the loss of confidence following a fall may impede independent ambulation [3-51. Recently, research examining the causes and consequences of patient falls has escalated. Many researchers have investigated symptoms that may contribute to the fall, such as, orthostatic hypotension [6-91, urinary frequency and urgency [lo], problems with vision [ll], mental confusion [12], dizzyness [ 131, balance and gait [14, 151 or ‘drop attacks’ [16, 171. Other researchers have investigated specific diseases that contribute to patient falls, such as, Alzheimer’s disease [ 1S], Parkinson’s disease [ 121, CVAs (191 and the presence of multiple illnesses [14,20] or advancing age [lo]. Finally, investigators have explored iatrogenic factors, such as, drug interactions or reactions [7,21], the use of psychotropic drugs or sedatives [ 10,221, factors associated with the hospital environment [23], the location of the fall and patient activity at the time of the fall [24] and patient characteristics. such as the inability to communicate with staff due to language barriers or speech impairments [lo]. Although a patient fall is clearly a multifaceted problem, it is surprising that most researchers The

*Address correspondence to: Dr J. M. Morse, of Nursing, Clinical University of Alberta, Canada.

Sciences Building, Edmonton, Alberta

Faculty 3rd Floor, T6G 2G3,

81

JANKE M. MORSE et al.

82

Table

I. Patient xx by unit Male

Total

Female

n

%

n

%

n

%

143 115 260

49.1 49.8 54.7 54.5

148 116 121 217

50.9 50.2 45.3 45.5

291 231 267 417

10.8 8.6 9.9 17.7

139 205

59.4 46.7

95 234

40.6 53.3

234 439

a.7 16.3

Long ,erm Long term care Nursing home

I3 78

28.3 loo.0

33 0

71.7 0.0

46 78

1.7 2.9

Rehabilitation Neuromuscular Head injury I Orthopedics Diabetes Weight control Head injury II Stroke I Stroke II

I9 I5 84 39 28 25 50 35

24 75 5 26 61 38 84 I8

55.8 83.3 5.6 40.0 68.5 60.3 62.7 34.0

43 90 89 65 89 63 134 53

1.6 3.3 3.3 2.4 3.3 2.3 5.0 2.0

Unit Ophthalmoiogy General surgery I General

surgery

I46

II

General medicine

I

General medicine II G.I. and endocrinology

1394

Total

44.2 16.7 94.4 60.0 31.5 39.7 37.3 66.0 51.8

METHOD

Setting

The study was conducted in two institutions. Six units were selected from the acute care division [general surgical (2 units), ophthalmology (1 unit) and three medical units] along with two units from the long term care division (psychogeriatric and nursing home) from a 1100 bed general hospital. Also eight adult units were selected from the 240 bed rehabilitation hospital [i.e. neuromuscular, orthopedic, diabetes, weight control, head injury (two units and a CVA unit]. The average length of stay in the acute care areas of the acute care hospital was 10 days, with a fall rate of 2.5 falls per 1000 patient bed days in the previous year. Patients were frequently transferred to the rehabilitation hospital from other hospitals in the region, and the average length of stay in that hospital was 40 days. The patient fall rate for the rehabilitation hospital for the previous year was 3.2 falls per 1000 patient bed days. Research design

A pilot project to assess the feasibility of the project was initiated in November, 1985. The pilot was conducted for 2 weeks to determine the most effective methods of data collection. Thereafter, one unit in each institution was introduced to the project every few weeks. For the first 1 or 2 weeks, staff were introduced to the project, instructed in the use of the Morse Fall Scale (from a video learning tape*) and fall-prevention strategies were discussed. Nurses rated all patients’ risk of falling daily, documented fall prevention strategies used and, if a fall occurred, noted the time of occurrence, type of fall and any causative factors. Only falls that occurred on the patient’s unit were included in the analysis (i.e. +A IOmin videotape on the use of the Morse Fall Scale is available without charge. Forward a blank tape to Mr R. Robinson, Educational Audiovisual Services, Glenrose Rehabilitation Hospital, 10230-I I I Avenue, Edmonton, Alberta T5G 0B7, Canada.

1295

48.2

2689

100.0

patient falls that occurred in another department, such as X-ray or while the patient was out on a pass, were excluded). The Instrument

The Morse Fall Scale consists of six variables that are easily identified and quick to score. These variables are: history of falling, presence of secondary diagnosis, use of ambulatory aids (cane, wheelchair or walking frame), administration of intravenous therapy, type of gait (normal, weak or impaired) and mental status. As stated, the scale was developed from a previous study [20,27] using 100 patients who fell and 100 controls. Discriminant analysis identified 80.5% of the patients from the control group who fell. Computer modelling techniques, using a simulated normalized hospital population, increased the discriminate function to 82.9%. Analysis of the false positive cases 6 weeks after the collection of data showed that 5 falls had occurred in these 17 patients (one patient had fallen three times), suggesting that these patients were fall-prone but had not had the opportunity to fall earlier. Analysis of the false negative cases indicated that those falls were either ‘accidental falls’ or ‘unanticipated physiological falls’ attributed to unpredictable events, such as, a seizure, unstable knee joint, drop attacks or fainting. Inter-rater reliability scores on the scale, using 21 raters, was R = 0.96. In order to obtain consistency, videotapes of patients ambulating were used for this trial. RFSUI~TS

Data collection extended from 1 December, 1985 to 30 April, 1986. A total of 252 weeks of data were collected from 16 patient care units. 2689 patients were assessed during this period, 41.2% of whom were over the age of 65 years. Patients, by unit and sex are shown on Table 1. A total of 49,946 patient bed days were recorded: 16,563 from the acute care areas, 13,542 from the long term care areas and 19,841 from the rehabilitation hospital. As patients

The fall-pronepatient

83

Table 2. Fallscoresby setting Settine Fall score 0

IO 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 I05 110 II5 I20 125 Missing Total

Acute care n

%

Long term care n

522 37 241 332 63 99 197 66 75 75 41 45 20 35 25 7 20 6 13 6 3 5 0 0 I 5

26.9 1.9 12.4 17.1 3.2 5.1 10.2 3.4 3.9 3.9 2.1 2.3 1.0 1.8 I.3 0.4 I.0 0.3 0.7 0.3 0.2 0.3 0.0 0.0 0.1 0.3

3 0 22 0 4 II I6 I2 0 17

1939

100.1

%

Rehabilitation "

%

Total n

%

Cum. 4;o

7 9 0 13 1 0 7 0 0 I 0 0 0 0 0

2.4 0.0 17.7 0.0 3.2 8.9 12.9 9.7 0.0 13.7 0.8 5.6 7.3 0.0 10.5 0.8 0.0 5.6 0.0 0.0 0.8 0.0 0.0 0.0 0.0 0.0

27 I4 57 IO 81 22 49 114 6 91 I5 26 39 2 52 6 0 I3 I 0 I 0 0 0 0 0

4.3 2.2 9.1 I.6 12.9 3.5 7.8 18.2 1.0 14.5 2.4 4.2 6.2 0.3 8.3 1.0 0.0 2.1 0.2 0.0 0.2 0.0 0.0 0.0 0.0 0.0

552 51 320 342 148 132 262 192 81 183 57 78 68 37 90 I4 20 26 I4 6 5 5 0 0 I 5

20.5 1.9 II.9 12.7 5.5 4.9 9.7 7.1 3.0 6.8 2.1 2.9 2.5 1.4 3.4 0.5 0.7 I.0 0.5 0.2 0.2 0.2 0.0 0.0 0.0 0.2

20.5 22.4 34.3 47.0 52.5 57.4 67.1 74.2 17.2 84.0 86.1 89.0 91.5 92.9 96.3 96.5 91.5 98.5 99.0 99.2 99.4 99.6 99.6 99.6 99.6 99.8

124

99.9

626

100.0

2689

99.8

100.0

1

usually were not discharged from hospital until they could cope without surveillance, the mean length of stay in each institution was used as an indication to compare the short and the long stay patients. The mean length of stay was 10 days in the acute care hospital and 40 days in the rehabilitation hospital. However. the mean length of stay could not be calculated for the long term care area as the patients are rarely discharged. Of 2689 patients in both institutions, 1265 (47.1%) scored as low risk of falling (i.e. <20), 734 (27.3%) scored as medium risk (i.e. 25-40) and 690 (25.5%) as high risk (i.e. 245). The distribution of total fall scores obtained by setting are shown in Table 2. Distinct differences in the distribution of scores between groups were immediately apparent. The mean score in the acute care setting was 24.78 (SD 22.95). with 58.3% scoring as low risk, 21.8% as medium risk and 19.6%. as a high risk. In the long term care setting. a mean score of 44.37 was obtained (SD 23.35) and 20.1% of these patients received a low score, 34.7% a medium score and 45.1% a high score. The scores for the rehabilitation area. with a mean of 41.9% (SD 21.11), showed that 15.6% rated low. 25.8% medium and 57.6% as a high risk. Analysis of each item in the Morse Fall Scale showed that 50.4% of the patients’ scores varied during the hospital stay. The items that increase the mean scores were attributed to changes in ambulatory aids (30.1%) and deterioration in gait from normal to weak, or weak to impaired (23.4%). A fall increased the patient scores in only 12% of the cases. On the other hand, a decrease in patient scores resulted from improvement in gait in 43.1% of the cases. removal of intravenous infusion in 23.9% and an improvement in mental status in 18.1%.

In the acute care setting, differences in fall scores according to patient condition were reflected when the scores of the short stay (i.e. < 10 days) and the long stay (> 10 days) patients were compared. When analyzed day by day, changes in the patients’ scores reflected differences in the patients’ condition (see Fig. 1). For example, many patients are admitted to the eye unit for minor surgery. The pattern of patient scores in this unit peaks on the day following surgery (day 2) for those admitted for minor surgery and the scores decrease when these patients begin to ambulate as they approach discharge. On the other hand, the scores remain elevated for longer-term patients. Unfortunately, the mean length of stay of 40 days did not provide an adequate sample to permit analysis between long and short stay patients in the rehabilitation hospital or for the long term care setting. However, daily patient scores in the rehabilitation hospital did increase the second day after admission, perhaps because the patients after assessment, were no longer on bed rest and were encouraged to ambulate. Nevertheless, in both the rehabilitation and the long term care areas, scores were relatively flat compared with the variability evident in the acute care hospital. Examination of the type of patient fall (i.e. physiological anticipated fall, physiological unanticipated fall and accidental fall) by the patients’ fall scores revealed that of 147 falls, 91 (i.e. 61.9%) were physiological anticipated falls; whereas, only 20 (13.6%) were unanticipated falls and 36 (24.5%) were accidental. 89.4% of patients who fell repeatedly experienced the same type of fall. The largest percentage of fallers, regardless of type, were high scorers (76.9%); this difference was particularly apparent in the anticipated physiological fall category. The association between fall score category and type of fall was

JANICE h4. MORSE Ed al.

84

+

greatest proportion of injuries were incurred by those patients scored as ‘anticipated physiological falls’ (i.e. 22 of 91 falls, or 24.2%). Eight of these 22 falls resulted in serious injuries (see Table 4). The only other serious injury in the study was an accidental fall, also experienced by a patient with a high fall score, who fractured both a hip and an elbow. Thirteen patients from the acute care area, eight from the long term care and 20 from the rehabilitation areas received injuries resulting from the fall.

LONGSTAY

Repeated fallen

01

0

4

2

a

6 DAYS

1 Q +

201. 0

1.Fall

12

I

50

Fig.

10

I.

I. 4

2

scores for patients

stay. Examples

SHORT STAY LONGSTAY

I’ 6 DAYS

I. 8

I 10

.I 12

with long term and short term

from: (a) surgical and (b) medical units.

statistically significant (x’ = 30.2, df = 4, P < 0.01; see Table 3). On the other hand, most of the falls with patient scores in the low or moderate categories were classified as ‘unanticipated’ or ‘accidental’ falls. Of the patients who fell and were injured, the

During the study period, 26 patients fell twice or more for a total of 66 falls (i.e. 26 first falls and 40 repeated falls), with one patient falling five times in the study period. The numbers of falls by patient care area are shown on Table 5. The mean fall score for patients in this category was 80 (SD 17.6). All of the scores of patients who fell repeatedly in the rehabilitation hospital are shown by fall on Table 6. In this group, the scores of six of the ten patients increased considerably over the course of hospitalization. Three of the patients in the acute care hospital also had increased scores with subsequent falls, but the patients in the long term areas, with higher mean scores, tended to have scores that were stable. At the conclusion of the project, the nursing staff were surveyed to ascertain the clinical feasibility of continuing the use of the scale. 175 nurses responded, of whom 82.9% rated the scale as quick and easy to use, and 54% estimated that it took less than 3 min to rate a patient using the scale. 63% thought it should be a part of ongoing nursing assessment. DISCUSSION

Longitudinal evaluation of the Morse Fall Scale shows that the scale appears to be a valid predictor

Table 3. Falls sccxe bv tvae of fall and settme

Fall sc0re

Anticipated (n = 91).

Unanticipated (n = 20)

AccIdental (n = 36)

AtCt

LTC:

RB

A/C

LTC

R

A/C

LTC

LOW Moderate High

0 4 26

0 2 25

1 2 31

5

0

I 3

I I

0 3 6

2 2 5

Total

30

27

34

9

2

9

9

Total R

n

I 5 6

1 4 IO

10 24 II3

6.8 16.3 16.9

I2

I5

1477

100.0

*Number of falls. tAcure care. $Long term care. ERehabihtatmn ’ Includes repeated falls (n = 107 patients) Table 4. Injuries by sccve and type of fall Fall scOl-e Low Moderate Hleh Total

Anticipated (n = 91)’ 0

I

serioust 13 minor 8 serious: 22

Unanticipated (n = 20). 2 minor 3 minor 5 minor

Accidental (II = 36)’

Total (n = 147)’

2 minor 3 minor

4 7 30

I s.erms~ 3 minor

IO

9

41

*Number

of falls. tl fractured ankle. :2 fractured ribs; I fractured hip; I fractured ankle; laceration: I sprain; I concussion. 81 fractured hip and elbow.

I

fractured vertebra-C2;

I

%

The fail-prone Table

5. Numbers

of patients

who

number

fell

repeatedly.

by settmg

and

patient Table

6. Fall

85 scores for patients

of falls

who

Setting Number of falls

T4cute care

Long

Patient Rehabilitatron

First

Second

I

50.

1051

Third

2

7

5

6

2

90’

90’

3

I

0

2

3

6Yt

90‘5

4

0

2

2

4

35)

75)

5

0

I

0

5

75t

75t

17

23

26

6

251

401

7

75

75v

8

7%

75

75t

9

60t

60t

90t

IO

40t

6%

75t

Total

rehabilitatton

Fall

term

care

fell repeatedly:

hospital

=66

of patient falls. Of the total of 107 patients who fell during the study period, I13 of the falls were experienced by the 75 patients who scored in the ‘high risk’ category. The scale was particularly successful in predicting anticipated physiological falls; 90.1% of fallers in this category were identified by the scale as being at high risk. Furthermore, the scale is sensitive to changes in the patient condition, as evidenced by the variability in daily scores, particularly in the surgical areas of the acute care hospital. Items that contribute to the patients’ fall scores vary by unit (i.e. by medical specialty), which suggests that the scale is sensitive to levels of disability. There are distinct differences in the score profiles according to institution, with a greater proportion of patients receiving a high score in the long term care institution and the rehabilitation hospital. Thus, the scale is correctly identifying differences that are intuitively obvious; one would expect rehabilitation and long term care patients to be at greater risk of falling. It is important to note that although more than 57.6% of the rehabilitation hospital patients scored as having a high risk of falling, the fall rate (2.7 falls per 1000 patient bed days) was lower than the long term care area and the acute care area (3.0 and 2.9 falls per 1000 patient bed days respectively). All but one of the serious injuries occurring during the study period were in patients who were identified by the Fall Scale as being at high risk of falling. This suggests that patients in the high risk category are more likely to be injured if they do fail and indicates that those caring for the patients with a high fall score should be increasingly aware of the necessity of fall prevention. Examination of the scores of the patients that fell repeatedly showed that the scores were very high, and in nine cases. the scores increased substantially with each successive fall. indicating increasing frailty. This is consistent with other research [26] which notes that falls tend to cluster prior to death. There were two limitations in this study. Firstly, the patient care areas were not selected randomly. They were chosen because they were considered ‘high risk’ areas where patient falls were problematic. Thus, areas such as obstetrics were excluded, and the sample was not representative of a typical hospital population. Patients under the age of 18 years were *Those patients under 18 years of age were excluded from the research project because (1) regulation for the protection of Human Subjects requires parental consent for the inclusion of minors in research and (2) falls by young children include falls from climbing and tripping which are a part of normal activity and thus are considered a separate phenomenon.

*Same

circumstances

tSame

actiwty

a: ttme

at time of fall

of fall (e.g.

Fourth

75.t

(e.g.

going

transferrmg

90 90 to the bathroom). or reaching

for

an

object).

also excluded, and no data are available on children or younger adolescents.* It is recommended that the study be repeated, and all patients in several institutions be scored so that more representative norms of patient scores may be obtained. The second limitation is a problem that is unavoidable when studying patient falls, and it is related to design. It is not feasible to separate the rating of fall risk from the subsequent possibility of a fall without also trying simultaneously to prevent the fall. Thus, the researcher is in a paradoxical situation of both predicting the criterion variable and preventing its occurrence. There is, however, no moral solution to this problem, and it becomes a choice of conducting imperfect research or not conducting the research at all [29]. Thus, the staff were given fall prevention strategies to prevent falls in patients rated at medium or high risk of falling despite the fact that successfully preventing a fall could interfere with the significance of the results. A related problem with the researcher’s or administrator’s sudden interest in falls is the increased reporting of falls during a study period. As all patient falls that result in injury are reported, it is probable that injury rate is less variable to reporting error than fall rate. It is recommended in future studies that the injury rate during the research period be compared with previous years to identify changes in the standards of reporting. Nevertheless, the major strengths of this project were that the study was conducted in the clinical areas, during normal working days and the ratings were done by clinicians who would normally be using the scale. As other fall scales have been too complex to use without pencil and paper 1301, required a physical examination [31] or taking a clinical history [32] to assess the patient’s fall risk, the ease of use and administrative feasibility of the Morse Fall Scale appears advantageous. However, although the scale assists in the prediction of fall-prone patients, investigation of methods to prevent patient falls must continue. In conclusion, the Morse Fall Scale was found to be an effective predictor of the fall-prone patient, and it appears to be a feasible tool for nurses to use in the clinical area. The ability to identify the fall-prone patient is an essential first-step in implementing a fall-prevention program and in ensuring that unnecessary injuries to the patient are prevented and costs to the health care system reduced.

JANICE M. MORSE Ed al.

86

Acknoi&dgemenrs-The author acknowledges the assistance of the following persons with this project: research assistants R. Tuck, RN. BN. and G. O’Connor RN; S. Tylko. RN, BScN assisted with data analysis and S. Warren. PhD served as statistical consultant. More than 400 staff nurses in two mstitutions were involved with data collection. Funding for this project was provided by the University of Alberta Hospitals Foundation, UAH Special Services Committee. and, in part, by the Glenrose Rehabilitation Hospital and an NHRDP Research Scholar Award to Dr Morse.

REFERENCES I

7.

8. 9.

10.

II. I?. 13.

14.

Kalchthaler T.. Bascon R. A. and Quintos U. Falls in the institutionalized elderly. J. Am. Geriaf. Sot. 26, 424, 1978. Baker S. P. and Harvey A. H. Fall injuries in the elderly. Clin. Geriaf. Med. 1, 501. 1985. Isaacs B. Don‘t get up you’ll fall. Nursing Mirror 149, 40, 1979. Isaacs B. Don’t let her walk she’ll fall. Nursing Mirror 149, 26, 1979. Morris E. V. and Isaacs B. The prevention of falls in a geriatric hospital. Age & Ageing9, 181. 1980. Camobell A. J.. Reinken J.. Allan B. C. and Martinez G. S.* Falls in old age: a study of frequency and related clinical factors. Age & Ageiig 10, 264, 1981. Davie J. W.. Blumenthal M. D. and Robinson-Hewlens S. A model of risk of falling for psychogeriatric patients. Archs Geriar. Psychiat. 38, 463, 1981. Tideikssar R. and Kay A. D. What causes falls? A logical diagnostic procedure. Geriatrics 41, 32, 1986. Overstall P. W. Falls in the elderly-Epidemiology, etiology and management. In Recent Adtiances in Geriarric Medicine (Edited by Isaacs B.), p. 61. Churchill Livingstone. New York. 1978. Warshaw G. A., Moore J. T., Friedman W. S., Currie C. T.. Kennie D. C.. Kane W. J. and Mears P. A. Functional disability in the hospitalized elderly. J. Am. med. Ass. 248. 847. 1982. Cohn T. E. and Lasley D. J. Visual depth illusion and falls in the elderly. C/in. Geriar. Med. 1, 601. 1985. Johnson E. T. Accidental falls among geriatric patients: can more be prevented? J. nam. med. Ass. 77,633, 1985. Davie J. W.. Blumenthal M. D. and Robinson-Hewlens S. A model of risk of falling for psychogeriatric patients. Archs Geriar. Psychiar. 38, 463, 1981. Tinetti M. E.. Williams T. F. and Mayewski R. Fall risk index for elderly patients based on number of chronic disabilities. Am J. Med. 80, 429, 1986.

IS. Wolfson L., Whipple R., Amerman P.. Kaplan J. and Kleinberg A. Gait and balance in the elderly, two functional capacities that link sensory and motor ability to falls. Clin. Geriar. Med. 1, 649, 1985. 16. Brockelhurst J. C., Exton-Smith A. N. and Barber S. M. Fractures of the femur in old age: A two-centre study of associated clinical factors and the cause of a fall. Age & Ageing 7, 7, 1978. 17. Lipsitz L. A. The drop attack: a common geriatric problem. J. Am. Geriat. Sot. 31, 617. 1983. 18. Morris J. C.. Rubin E. H., Morris E. J. and Mandel S. A. Senile dementia of the Alzheimer’s type. An important risk factor for serious falls. J. Geronr. 42, 412. 1987. 19. Grant J. S. and Hamilton S. Falls in a rehabilitation center: a retrospective and comparative analysis. Rehabil. Nursing 12, 74. 1987. 20. Morse J. M.. Tylko S. J. and Dixon H. A. Characteristics of the fall-prone patient. Gerontologist 27, 516, 1987. 21. Botez M. I. and Hausser C. 0. Falls. Br. J. Hosp. Med. 28, 494, 1982. 22. Barbieri E. G. Patient falls are not patient accidents. J. Geront. Nurs. 9, 165, 1983. 23. Halliday P., Gernie G. R. and Lauzon, F. S. Some bio-engineering approaches to the falling problem. Geriat. Med. 1, 161, 1985. 24. Haga H.. Shibata H., Shichita K., Matsuzaki T. and Hatano S. Falls in the institutionalized elderly in Japan. Archs Geront. Geriar. 5, 1. 1986. 25. Morse J. M., Tylko S. and Dixon H. A. The patient who falls. and falls again: Defining the aged at risk. J. Geront. Nurs. 11, 85, 1985. 26. Gryfe C. I., Aimes A. and Ashley M. J. A longitudinal study of falls in an elderly population. Age & Ageing 6, 211, 1977. 27. Morse J. M., Morse R. M. and Tylko S. J. Development of a scale to identify the fall-prone patient. Unpublished manuscript. 28. Morse J. M. Computerized evaluation of a scale to identify the fall-prone patient. Can. J. publ. Hlth Suppl., 77, 21, 1986. 29. Robb S. S.. Stegman C. E. and Wolanin M. 0. No research versus research with compromised results: a study of validation therapy. Nurs. Res. 35, 113, 1986. 30. Rainville N. G. Effect of an implemented fall prevention program on the frequency of patient falls. Qual. Rev. Bull. 10, 287. 1984. 31. Tideikssar R. and Kay A. D. What causes falls? A logical diagnostic procedure. Geriatrics 41, 32, 1986. 32. Pate1 K. P. Falls and faints in the elderly: look to their clinical history for clues. Mod. Geriaf. 6, 28, 1976.