A comparative study of the use of four fall risk assessment tools on acute medical wards.

Abstract

To compare the effectiveness of four falls risk assessment tools (STRATIFY, Downton, Tullamore, and Tinetti) by using them simultaneously in the same environment.
Prospective, open, observational study.
Two acute medical wards admitting predominantly older patients.
One hundred thirty-five patients, 86 female, mean age+/-standard deviation 83.8+/-8.01 (range 56-100).
A single clinician prospectively completed the four falls risk assessment tools. The extent of completion and time to complete each tool was recorded. Patients were followed until discharge, noting the occurrence of falls. The sensitivity, specificity, negative predictive accuracy, positive predictive accuracy, and total predictive accuracy were calculated.
The number of patients that the STRATIFY correctly identified (n=90) was significantly higher than the Downton (n=46; P<.001), Tullamore (n=66; P=.005), or Tinetti (n=52; P<.001) tools, but the STRATIFY had the poorest sensitivity (68.2%). The STRATIFY was also the only tool that could be fully completed in all patients (n=135), compared with the Downton (n=130; P=.06), Tullamore (n=130; P=.06), and Tinetti (n=17; P<.001). The time required to complete the STRATIFY tool (average 3.85 minutes) was significantly less than for the Downton (6.34 minutes; P<.001), Tinetti (7.4 minutes; P<.001), and Tullamore (6.25 minutes; P<.001). The Kaplan-Meier test showed that the STRATIFY (log rank P=.001) and Tullamore tools (log rank P<.001) were effective at predicting falls over the first week of admission. The Downton (log rank P=.46) and Tinetti tools (log rank P=.41) did not demonstrate this characteristic.
Significant differences were identified in the performance and complexity between the four risk assessment tools studied. The STRATIFY tool was the shortest and easiest to complete and had the highest predictive value but the lowest sensitivity.

Full text

Page 1

BRIEF METHODOLOGICAL REPORTS
A Comparative Study of the Use of Four Fall Risk Assessment
Tools on Acute Medical Wards
Michael Vassallo, FRCP, PhD,?Rachel Stockdale, MRCP (UK),wJagdish C. Sharma, FRCP,?
Roger Briggs, FRCP,zand Stephen Allen, FRCP§
OBJECTIVES: To compare the effectiveness of four falls
risk assessment tools (STRATIFY, Downton, Tullamore,
and Tinetti) by using them simultaneously in the same
environment.
DESIGN: Prospective, open, observational study.
SETTING: Two acute medical wards admitting predomi-
nantly older patients.
PARTICIPANTS: One hundred thirty-five patients, 86
female, mean age ? standard deviation 83.8 ? 8.01 (range
56–100).
MEASUREMENTS: A single clinician prospectively com-
pleted the four falls risk assessment tools. The extent of
completion and time to complete each tool was recorded.
Patients were followed until discharge, noting the occur-
rence of falls. The sensitivity, specificity, negative predictive
accuracy, positive predictive accuracy, and total predictive
accuracy were calculated.
RESULTS: The number of patients that the STRATIFY
correctly identified (n590) was significantly higher than
the Downton (n546; Po.001), Tullamore (n566;
P5.005), or Tinetti (n552; Po.001) tools, but the
STRATIFY had the poorest sensitivity (68.2%). The
STRATIFY was also the only tool that could be fully com-
pleted in all patients (n5135), compared with the Down-
ton (n5130; P5.06), Tullamore (n5130; P5.06), and
Tinetti (n517; Po.001). The time required to complete
the STRATIFY tool (average 3.85 minutes) was significant-
ly less than for the Downton (6.34 minutes; Po.001),
Tinetti (7.4 minutes; Po.001), and Tullamore (6.25 min-
utes; Po.001). The Kaplan-Meier test showed that the
STRATIFY (log rank P5.001) and Tullamore tools (log
rank Po.001) were effective at predicting falls over the first
week of admission. The Downton (log rank P5.46) and
Tinetti tools (log rank P5.41) did not demonstrate this
characteristic.
CONCLUSION: Significant differences were identified
in the performance and complexity between the four risk
assessment tools studied. The STRATIFY tool was the
shortest and easiest to complete and had the highest pre-
dictive value but the lowest sensitivity. J Am Geriatr Soc 53:
1034–1038, 2005.
Key words: falls; risk assessment; comparative study
B
tients likely to benefit from expensive multidisciplinary in-
terventions.4Several fall risk factors have been identified,
and some of them have been compiled into fall risk assess-
ment tools.5,6Such tools are based on the premise that the
higher the number of risk factors, the higher the risk of
falling.7
Although a number of tools have been used to identify
fall risk in hospitalized patients,8not all have been validat-
ed.9Several of those that have been validated have high
accuracy, but when tested outside the specific setting in
which they were originally validated, the predictive accu-
racy is not reproduced.10This may have occurred because
of different patient and staff characteristics, as well as op-
erational differences between the environments. There is
considerable overlap between the characteristics used to
compile the tools, raising the question of whether fall risk
assessment tools are indeed any different. A comparative
study of four assessment tools was therefore conducted to
determine whether there are any differences in performance
of the tools in the same environment. The chosen fall risk
tools were the STRATIFY,10the Tullamore,11a Tinetti-
based assessment,11and the Downton.12They were chosen
because staff were familiar with their use, they had previ-
ously been used on elderly people in various hospital set-
tings to predict the risk of falls,1,8,10,11,13and they are still
being used widely in many hospitals.
By studying four fall risk assessment tools simulta-
neously with the same investigators, under the same ward
ecause there is some evidence that falls in hospital can
be reduced,1–3it is important to identify high-risk pa-
Address correspondence to Dr. Michael Vassallo, Royal Bournemouth
Hospital, Castle Lane East, Bournemouth, BH7 7DW, United Kingdom.
E-mail: michael.vassallo@rbch-tr.swest.nhs.uk
DOI: 10.1111/j.1532-5415.2005.53316.x
From the?Kings Mill Hospital, Sutton in Ashfield, United Kingdom;wRoyal
Bournemouth Hospital, Bournemouth, United Kingdom;zSouthampton
General Hospital, Southampton, United Kingdom; and§University of
Bournemouth, Poole, United Kingdom.
JAGS
r 2005 by the American Geriatrics Society
53:1034–1038, 2005
0002-8614/05/$15.00

Page 2

conditions,itwashopedtodeterminewhethertherearereal
differences in effectiveness between fall risk assessment
tools and whether more complex tools are any better than
simple tools at identifying patients who fall on medical
wards.
METHODS
The study was conducted in two medical wards admitting
predominantly elderly patients. They were admitted for
treatment of a wide range of medical conditions. Approval
was obtained from the North Notthinghamshire ethics
committee. One hundred thirty-five consecutive patients
were studied. None declined to participate. On admission,
patients had medical and nursing assessments. A single cli-
nician prospectively conducted the medical assessment,
which consisted of measurements of vision, depression,
mobility, and a medication review. The nursing assessment
included information on agitation, need of frequent toilet-
ing, activities of daily living, and postural blood pressure.
The clinician completing the medical assessment completed
the risk assessment tool. Fall prevention measures were not
dependent on the score obtained, but measures were taken
to try to correct any individual fall risk factors identified
using the various tools. For example, postural hypotension
would be treated even though the patient might be at low
overall risk of falls.
Information was collected on patients’ age, sex, history
of falls, and medications on admission. Patients had a
physical examination noting the presence of impaired vi-
sion, hearing loss, lower limb abnormalities, gait distur-
bance,back extension,postural
confusion. Patients were deemed to have impaired vision
if they were registered blind or partially sighted or were
unable to see less than 6/60 on a Snellen chart using glasses,
if appropriate. Hearing impairment was defined as the in-
ability to follow a conversation with or without using a
hearing aid. A limb was considered abnormal if there was
any evidence of weakness (Medical Research Council cri-
teria grade 4/5 or less), neuropathy, amputation, joint ab-
normality excluding minor osteoarthritic changes, or any
condition judged to interfere with normal gait such as cell-
ulitis or a deep vein thrombosis. A patient’s gait was as-
sessed using the Get Up and Go Test.14On this basis,
patients were classified into four groups: normal, safe (with
or without using aids), unsafe (with or without using aids),
and unable, if the patient was bedridden. Back extension
was studied after testing the patient’s mobility and was re-
corded with the patient standing as being able or unable to
perform the maneuver. Patients were considered to be con-
fused if they scored less than 7 of 10 on the Hodkinson
Abbreviated Mental Test score.15
The Downton Fall risk tool12was compiled based on a
history of falls, medications (tranquilizers/sedatives, diu-
retics, antihypertensivesexcluding
parkinsonian drugs, and antidepressants), sensory deficits
(visual impairment, hearing impairment), limb abnormal-
ities (such as hemiparesis), confusion, and unsafe gait (with
or without aids). Each of these factors scored a point; scores
of 3 or above identify patients at risk.
STRATIFY consists of five factors, each found to be
independently associated with falling.10These factors are
hypotension, and
diuretics,anti-
presenting with a fall or having a fall on the ward, the
presence of agitation, visual impairment, need for frequent
toileting, and impaired ability to transfer and walk. Scores
of 2 or more were considered to be high risk.
The Tinetti fall risk index is based on number of chron-
ic disabilities.7The higher the number of chronic disabil-
ities, the higher the likelihood of having recurrent falls. The
nine risk factors included in the fall risk index are mobility
score, morale score, mental status score, distance vision,
hearing, postural blood pressure drop, back examination,
medications on admission, and admission activity of daily
living score. To simplify the tool, the Geriatric Depression
Scale score, Get Up and Go test, and Abbreviated Mental
Test score were used instead of the Philadelphia Morale
score, gait and balance assessment, and Mini-Mental State
Examination score, respectively. These modifications were
validated for use16(unpublished data). The subjects’ fall
risk score was the number of index risk factors present.
Scores of 0 to 3 were considered low, 4 to 6 was medium,
and 7 to 9 was high risk. For the purposes of analysis, me-
dium and high risk were considered together.
The Tullamore tool11assesses sex, age, gait, sensory
deficits, falls history, medication, medical history and mo-
bility under various subheadings. Patients are classified as
low (score 3–8), medium (score 9–12) or high (score 13).
Medium- and high-risk scores were considered together.
For all the tools, the cutoff point from low to higher
risk was that suggested by the respective authors. Patients
were followed up to the point of discharge from the ward.
Nursing staff kept a record of falls as they occurred on the
wards in a falls diary. The clinician completing the tool was
blinded to the occurrence of falls that occurred after tool
completion. Patients who fell at least once were classified as
fallers. Other outcomes assessed were the number of risk
assessments that could be completed in their entirety on
initial assessment and how long it took to complete the falls
risk assessment. The time was calculated by estimating the
time taken for each individual assessment. The total for the
tool was then the total of all the individual components
required to complete the tool.
STATISTICS
The sensitivity, specificity, and total predictive accuracy of
thetoolswerecalculated.Sensitivitywasdefinedasthetotal
number of fallers correctly identified as high risk. Specificity
was defined as total number of nonfallers correctly defined
as low risk. The total predictive accuracy was the total
number of patients correctly identified expressed as a per-
centage. The positive predictive value was defined as the
number of high-risk patients who went on to fall whereas
the negative predictive value was the number of low-risk
patients who did not fall. Results were expressed as a per-
centage. Fishers exact probability test was used to compare
the accuracy of the various risk tools by comparing the
numbers of patients correctly identified by the various tools
to the best performing tool. Data was collected on 135 pa-
tients. Assessment items that could not be completed were
identified and recorded. Because we aimed to evaluate the
practical utility of each of the tools, all items were included
in calculating the total score. Incomplete items received a
A COMPARISON OF FOUR FALLS RISK ASSESSMENT TOOLS
1035JAGSJUNE 2005–VOL. 53, NO. 6

Page 3

scoreofzero sothatthetotalscoreforeachtoolwasderived
from the number of positively scored items.
The Kaplan-Meier hazard statistic was used to assess
the likelihood of falls in high- and low-risk patients for each
of the tools for the first week of patient stay using all 135
patients.Significance wasexpressed usingtheLog ranktest.
The number of falls was censored in daily time intervals for
the first week in both the high- and low-risk category for
each of the tools
RESULTS
One hundred thirty-five patients were studied: 86 female,
mean age ? standard deviation 83.8 ? 8.01 (range 56–
100). Almost all patients had an acute illness of varying
severity (e.g., respiratory tract infection, heart failure,
stroke, urinary tract infection) on a background of chron-
ic disease (e.g., arthritis or dementia). The mean length of
stay was 14.6 ? 7.5 days 7.5 (range 6–22 days). Twenty-
two fallers, of whom six hadrecurrent falls, contributing 29
falls in total, were identified. The performance of the var-
ious tools is shown in Table 1. The STRATIFY had the
highest total predictive accuracy but the lowest sensitivity.
The number of patients that the STRATIFY correctly iden-
tified (number of high-risk patients who fell and low-risk
patients who did not fall) was significantly higher than the
Downton (Po.001), Tullamore (P5.005), and Tinetti
(Po.001). In view of the low sensitivity, a separate anal-
ysis with 1 or above as the cutoff point for high risk was
performed. This change gave a sensitivity of 86% and
specificity of 25%. The number of correctly identified pa-
tients was 48 (35%). This was significantly inferior to the
Tullamore (P5.03), which now had the highest number of
correctly identified patients.
It was possible to complete the STRATIFY tool for all
patients evaluated. None of the other tools could be com-
pleted in their entirety for all patients (Table 2). The
STRATIFY performed significantly better than the Tinetti
(Po.001), which could be completed for only 17 patients.
Not all the Tinetti items could be completed because of
inability to perform postural blood pressure measures
(n575) or the Geriatric Depression Scale (n562) because
of severe cognitive function, although it was still possible to
classify 70% of patients as medium- to high-risk for falls
usingthe Tinetti tool. Completion of the STRATIFYdid not
differ from the Downton (P5.06) and Tullamore tools
(P5.06). The Downton and Tullamore were not completed
in five subjects because of an inability to complete a cog-
nitive assessment. It was still possible to classify the re-
spective patients into a high-risk category. The time
required to fill in the STRATIFY was significantly less than
that for the Downton (Po.001), Tinetti (Po.001), and
Tullamore (Po.001).
The predictive value of the tools to identify fallers over
the first week was analyzed using the Kaplan-Meier hazard
test. The STRATIFY (log rank P5.001) and Tullamore (log
rank Po.001) were able to identify fallers from the time of
admission and throughout the first week of patient stay
(Figure 1). The Downton (log rank P5.46) and Tinetti (log
rank P5.42) did not demonstrate a similar ability.
DISCUSSION
It is well recognized that the performance of any given fall
risk assessment tool varies when used in different settings.17
This may result from differences in patient, staffing, and
environmental characteristics. By studying and comparing
the performance of the various tools under the same
conditions in the same environment, this study attempted
to identify any differences in the effectiveness of these tools.
Although there is considerable overlap between the
Table 1. Characteristics of Tools When Identifying Fallers
Characteristic
Downton
(n5135)
STRATIFY
(n5135)
Tullamore
(Medium/high risk)
(n5135)
Tinetti
(Medium/high risk)
(n5135)
Sensitivity
Specificity
Positive predictive value
Negative predictive value
Patients correctly identified, n?
Total predictive accuracy, %
81.8
24.7
17.5
87.5
46
34.1
68.2
66.4
28.3
91.5
90
66.6
90.9
40.7
22.9
95.8
66
48.8
77.3
30.9
17.9
87.5
52
38.5
?STRATIFY vs Downton, Po. 001; Tullamore, P5.005; Tinetti, Po.001.
Table 2. Number of Completed Risk Assessment Tools and Time to Complete
Other Score Characteristic
Downton
(n5135)
STRATIFY
(n5135)
Tullamore
(Medium/high risk)
(n5135)
Tinetti
(Medium/high risk)
(n5135)
Time to complete, minutes, mean ? standard deviation
Number fully completedw
6.34 ? 2.62
130
3.85 ? 1.67
135
6.25 ? 2.56
130
7.40 ? 3.88
17
?STRATIFY vs Downton, Po.001; Tinetti, Po.001; Tullamore, Po.001.
wSTRATIFY vs Tinetti, Po.001.
1036
VASSALLO ET AL.
JUNE 2005–VOL. 53, NO. 6JAGS

Page 4

characteristics compiling the various tools, this study iden-
tified significant differences in their performance.
The total predictive accuracy of the various tools was
low, the highest being the STRATIFY at 66.6%. This is
principally because of low specificity. Specificity refers to
patients who do not fall having been correctly identified as
at low risk of falling. However, the hallmark of effective fall
prevention measures and high-quality care is that high-risk
patients are prevented from falling. A low specificity is ex-
pected in an environment in which patients are prevented
from falling. In addition, not all patients identified as being
at high risk will fall even if left on their own, further re-
ducing a tool’s total predictive accuracy. The ideal tool,
with high sensitivity and specificity, is difficult to develop,
and the most important measure of a fall risk assessment
tool is arguably its sensitivity. A possible way of improving
fall risk assessment is to focus on items in existing tools that
improved their sensitivity, such as a history of falls, confu-
sion, and an unsafe gait. The STRATIFY, despite having the
highest total predictive accuracy, had the lowest sensitivity
because it failed to identify the highest number of fallers as
beingathighrisk.Thisisanimportantweaknessofthetool.
Changing the cutoff point of the tool to 1 improved its
sensitivity by increasing the number of fallers identified as
high risk but reduced the predictive accuracy.
The tools differed in complexity, as reflected in the sig-
nificant differences in the time required to fill them out and
the inability to complete some of the tools in their entirety.
It is likely that the low accuracy of some tools like the
Tinetti resulted from the missing data. Despite this, it was
still possible to categorize the majority of patients into the
medium- to high-risk group. The analyses included all pa-
tients, regardless of whether the tools were completed in
their entirety because it was desired to study the utility of it
in real life, where one needs to decide on the outcome of a
risk assessment regardless of whether a tool is completed.
This is important because fall risk assessment needs to be
accurate, simple, and not time consuming to be implement-
ed effectively on wards without adding a considerable bur-
den on already hard-pressed staff.
This study has a number of limitations. The fall risk
assessment was done only once; therefore a change in pa-
tient condition could explain the low predictive value of the
tools. The results obtained may not be reproducible on
other units because of differing staff and patient character-
istics, including a different sex mix. Another limitation is
that it did not include allavailable fall risk assessment tools.
Some, such as the Morse Fall Scale,18are widely used in the
United States, but significant differences were identified
between the four risk assessment tools studied, with
STRATIFY having the best predictive accuracy but the
lowest sensitivity.
REFERENCES
1. Vassallo M, Vignaraja R, Sharma JC et al. The effect of changing practice on
fall prevention in a rehabilitative hospital. The Hospital Injury Prevention
(HIP) Study. J Am Geriatr Soc 2004;52:335–339.
Downton
Time (Days)
876543210
0.5
0.4
0.3
0.2
0.1
0.0
−0.1
0.5
0.4
0.3
0.2
0.1
0.0
−0.1
0.5
0.4
0.3
0.2
0.1
0.0
−0.1
0.5
0.4
0.3
0.2
0.1
0.0
−0.1
Risk
High
Low
STRATIFY
Time (Days)
876543210
Tullamore
Time (Days)
876543210
Tinetti
Time (Days)
876543210
Log Rank P = 0.42
Log Rank P = 0.46Log Rank P = 0.001
Log Rank P < 0.001
Risk
High
Low
Risk
High
Low
Risk
High
Low
Figure 1. Cumulative hazard to first fall for all the fall risk assessment tools.
A COMPARISON OF FOUR FALLS RISK ASSESSMENT TOOLS
1037JAGS JUNE 2005–VOL. 53, NO. 6

Page 5

2. Haines TP, Bennell KL, Osborne RH et al. Effectiveness of targeted falls pre-
vention programme in subacute hospital setting: Randomised controlled trial.
BMJ 2004;328:676–679.
3. Healey F, Monro A, Cockram A et al. Using targeted risk factor reduction to
prevent falls in older in-patients: A randomised controlled trial. Age Ageing
2004;33:390–395.
4. Capezuti E. Building the science of falls-prevention research. J Am Geriatr Soc
2004;52:461–462.
5. American Geriatrics Society, British Geriatrics Society, and American Acad-
emy of Orthopaedic Surgeons Panel on Falls Prevention. Guideline for the
prevention of falls in older persons. J Am Geriatr Soc 2001;49:664–672.
6. Chu LW, Pei CK, Chiu A. Risk factors for falls in hospitalised older medical
patients. J Gerontol A Biol Sci Med Sci 1999;54A:M38–M43.
7. Tinetti ME, Williams TF, Mayewski R. Fall risk index for elderly patients
based on number of chronic disabilities. Am J Med 1986;90:429–434.
8. Oliver D, Daly F, Martin FC et al. Risk factors and risk assessment tools for
falls in hospital in-patients: A systemic review. Age Ageing 2004;33:122–130.
9. Wyatt JC, Altman DG. Prognostic models: Clinically useful or quickly for-
gotten? BMJ 1995;311:339–341.
10. Oliver D, Britton M, Seed P et al. Development and evaluation of
evidence based risk assessment tool (STRATIFY) to predict which elderly
inpatients will fall: Case-control and cohort studies. BMJ 1997;315:
1049–1053.
11. Cunnington AL, Dhaylan B, Stott DJ. Predicting falls in long term careFare
the Cannard and Tinetti tools useful? Age Ageing 2003;32(Suppl. 1):16.
12. Downton JH. Falls in the Elderly. London: Edward Arnold, 1993.
13. Nyberg L, Gustafson Y. Using the Downton index to predict those prone to
falls in stroke rehabilitation. Stroke 1996;27:1821–1824.
14. Mathias S, Nayak USL, Isaccs B. Balance in elderly patients. The ‘get up and
go’ test. Arch Phys Med Rehabil 1986;67:387–389.
15. Qureshi KN, Hodkinson HM. Evaluation of a ten question mental test in the
institutionalised elderly. Age Ageing 1974;3:152–157.
16. Vassallo M. The prevention of falls in hospital. PhD Thesis. University of
Southampton, 2002.
17. Perell KL, Nelson A, Goldman RL et al. Fall risk assessment measures: An
analytic review. J Gerontol A Biol Sci Med Sci 2001;56A:M761–M766.
18. Morse JM, Morse RM, Tylko SJ. Development of a scale to identify the
fall-prone patient. Can J Ageing 1989;8:366–377.
1038
VASSALLO ET AL.
JUNE 2005–VOL. 53, NO. 6 JAGS