Professor Stephen Lord
Neuroscience Research Australia




Validity – component and total scale items predict
falls
Reliability – items have good test-retest and interrater reliability
External validation or bootstrap adjustment
Published in peer-reviewed journals

Have you fallen in the past 12 months?
Degree of difficulty – easy
Sensitivity and specificity – reasonable
Information gained about fall prevention – nil

Recommended by American and British Geriatrics
Societies

Surprisingly little validation as a predictor of falls

Varying methods
Usual vs. fast performance
Differing walk distances: 3m vs. 8ft
Differing instructions about turning: walk to line,
walk past line, walk around a cone

Varying cut-points: 10s, 14s, 15s, 22s
J Am Geriatr Soc 61:202–208, 2013.





TUG is not useful for discriminating fallers from non-fallers in
healthy, high-functioning older people
TUG is of more value in less healthy, lower-functioning older
people
Overall, the predictive ability and diagnostic accuracy of the TUG
are at best moderate
No cut-point can be recommended
Quick, multifactorial fall risk screens should be considered to
provide additional information for identifying older people at risk of
falls

Berg Balance Scale

Tinetti Performance Oriented Mobility Assessment
(POMA)

Physical Performance battery (Guralnik et al)

Elderly Fall Screening Test

Gait Abnormality Rating Scale

De Morton Mobility Index (DEMMI)
Short Physical Performance Battery

Has been used in very large studies, and
normative values have been established

Comprises
timed standing balance
repeated chair stands
timed walk

Each item scored /4 with combined score / 12
Sit-to-stand 1x & 5x
 Alternate step test (19cm high step)
 Turn test (n steps to turn 180 degrees)
 6m walk – normal speed
 Pick-up 5kg weight test
 Stair ascent (8 steps - 5cm high, 27cm deep)
 Stair descent (8 steps - 5cm high, 27cm deep)
 362 community-dwelling people aged 75+
 Compared w.r.t. validity, reliability and feasibility


Ability to predict multiple fallers
Test
STS – 1
STS- 5
Pick up
Turn
Alt step
6m walk
Stair asc
Stair des
Criterion
Sensitivity
Specificity
RR (95% CI)
≥1 s
≥12s
Unable
0.40
0.63
0.11
0.65
0.56
0.93
1.2 (0.7-2.0)
1.8 (1.2-2.7)
1.5 (0.8-2.6)
≥4
0.77
0.28
1.3 (0.8-2.0)
≥10s
≥6s
≥5s
≥5s
0.69
0.50
0.54
0.63
0.56
0.69
0.58
0.55
2.3 (1.4-3.5)
1.8 (1.2-2.6)
1.4 (1.0-2.1)
1.7 (1.2-2.6)


Poor performances
in two tests did
increase risk
Poor performances
in 3+ tests did not
increase risk further
Impairments
Odds Ratio
0
1
1
1.9
2
4.7
3
5.0
4
4.7
Test
STS- 5
Alt step
6m walk
Stair des
Stair asc
Turn
STS – 1
Pick-up
Validity
Reliability
Feasibility
Total
10
10
10
10
5
0
0
0
5
4
4
5
5
4
2
1
5
4
3
0
0
5
5
4
20
18
17
15
10
9
7
5
FallScreen –
physiological profile
CSRT
assessment
INHIB
PPA
TMT
TUG
verbal fluency
Oliver D et al, BMJ 1997;315:1049-53
Papaioannou A et al,
BMC Medicine 2004 1741-7015-2-1




Prospective study of 263 inpatients
Impaired balance, age 80 years and older and drug
and alcohol problems identified as additional falls risk
factors to the STRATIFY Tool
The Northern Hospital Tool had the greater accuracy
(J) (0.44 vs. 0.28, P = 0.006)
Inter-rater agreement of The Northern Hospital
Modified STRATIFY Tool was fair (j = 0.34) and low
for the STRATIFY Tool (j = 0.19)
Barker A et al, J Advanced Nursing 2010;67:450-7





Prospective cohort study involving 533 inpatients
Possible predictors of falls were collected from medical records,
interview and physical assessment
Fourteen percent of participants fell
A multivariate model to predict falls included:
 male gender (OR 2.70, 95% CI 1.57–4.64)
 CNS medications (OR 2.50, 95%CI 1.47–4.25)
 fall in the previous 12 months (OR 2.21, 95% CI 1.07–4.56)
 frequent toileting (OR 2.14, 95% CI 1.27–3.62)
 tandem stance inability (OR 2.00, 95% CI 1.11–3.59)
With 1 point allocated for each predictor the area under the
curve was 0.73 (95% CI 0.68–0.79)
Sherrington C et al, J Rehab Med 2010; 42: 482–488




Prospective study two hospital EDs in Sydney
Participants aged 70+ years who presented after falling or
a history of 2+ falls in past yr and subsequently discharged
Development study, n= 219 people (31% fell): external
validation study, n= 178 (35% fell)
Two-item screening tool:



AUC was 0.70 (0.64–0.76),


2+ falls in the past year (OR 4.18, 95% CI 2.61 to 6.68)
taking 6+ medications (OR 1.89, CI 1.18 to 3.04)
similar to FROP-Com (AUC 0.73, 0.67–0.79, p=0.25) and PROFET
screens (AUC 0.70, 0.62–0.78, p=0.5)
No physical measure predicted falls
Tiedemann A et al, EMJ 2013;30:11 918-922
The CahFRiS Screen
 Prospective observational cohort study –
254 residents
 Measures taken using care/medical
records and by talking to care staff
 Measures taken:
 Barthel, MMSE, neuro-psychiatric inventory,
impulsivity, medical conditions, medication
use, sit-to-stand, standing balance ratings
Falls in the six month f/u
 Of the 240 residents who completed the
follow-up, 121 (50.4%) fell one or more
times
 The fallers sustained a total of 281 falls
(range 1-16, mean=2.3)
 Overall, this equated to 2.8 falls per
person per year
Whitney J et al, Arch Gerontol and Geriatrics 2012;55:690-5
Absolute risk of falling
In-home assessments
Single assessment for stepping and cognition
• Conflicting stimuli are
presented (shape and word)
• The task is to “Step by the
word”
• Selectively attend to word and
inhibit the shape
• Recording time and errors
• Slow responses and stepping
errors discriminate fallers from
non-fallers
Schoene DS et al. Age and Ageing. 2013: doi: 10.1093/ageing/aft157
Automated performance recording
35
30
25
20
15
10
5
0
game
test
KS van Schooten, SM Rispens, PJM Elders, P Lips, JH van Dieën, M Pijnappels
VU University Amsterdam, The Netherlands, 2VU University Medical Center, The Netherlands
The price is $270 per 11 sq feet or $27,000 for full coverage of an
appartment.
With volume sales the cost could be reduced to $68 per 11 sq feet

There are several setting-specific evidence-based
fall risk screen and assessments

Degree of validation varies

Predictive value for most screens and assessments
are reasonable

New technologies may allow for accurate at-home
fall risk assessments, continuous monitoring and
telehealth care