Lindsay et al: The Timed Up and Go Test: Unable to predict falls on the acute medical ward Short Report

The Timed Up and Go Test: Unable to predict

falls on the acute medical ward
Robyn Lindsay1, Erica L James2 and Sandra Kippen2
1
Department of Physiotherapy, Bendigo Health Care Group 2LaTrobe University, Bendigo

Prevention of falls amongst older people is a high priority in health care. The aim of this study was to evaluate the ability of the
Timed Up and Go Test to predict those older people who will fall whilst admitted to an acute hospital. The medical records of 160
older patients who were admitted to the medical ward of a large regional hospital were accessed retrospectively. The Timed Up and
Go Test, used in isolation, was unable to identify those patients who were likely to fall. However the co-morbidity of incontinence
was identified as a falls risk factor (OR = 8.7, p = 0.001). The Timed Up and Go Test alone does not possess predictive validity for
acutely unwell older patients. It is therefore recommended that it not be used to identify those people who may fall. [Lindsay R,
James EL and Kippen S (2004): The Timed Up and Go Test: Unable to predict falls on the acute medical ward. Australian
Journal of Physiotherapy 50: 249251]

Key Words: Physical Therapy (Specialty); Movement Disorders; Geriatric Assessment; Accidental Falls

Introduction of balance and function (Podsiadlo and Richardson 1991) it

has potential for assessing falls risk in the hospital setting
The impact on health care resources created by Australias objectively.
ageing population can be expected to be felt for decades to
come, as longevity is often accompanied by substantially The aim of this study was to establish the effectiveness of the
declining health and increasing chronic illness and disability Timed Up and Go Test in identifying those elderly patients
(Hill et al 2000). One area of particular concern is falls that who would fall whilst in hospital. If found to be useful the
can result in fracture, extended hospital stays, complications Timed Up and Go Test may be used as a screening tool to
requiring diagnostic procedures and/or surgery, loss of identify those patients that require a more comprehensive
confidence and fear of further falls (Byers et al 1990, Hill et multidisciplinary assessment, as has been recommended as
al 2000). Falls risk does not decrease on admission to best practice by previous systematic reviews (Hill et al 2000).
hospital. Approximately 40% of older people fall whilst in Method
hospital, more than in the community or in residential aged
care facilities (Hill et al 2000). Subjects The study sample included all patients admitted to
the medical ward at the Bendigo Health Care Group (BHCG)
Clinical assessment of a patient to identify individual falls who were referred to the physiotherapist for a mobility
risk is usually performed soon after admission to hospital. assessment within the study period and who met the
Part of this may be a mobility assessment used in conjunction following inclusion criteria: aged 65 years or older, able to
with other clinical tools such as paper-based falls risk perform Timed Up and Go Test independently or with
assessments and specific balance tests such as the Berg assistance, with or without a gait-aid, and able to understand
Balance Scale (Berg et al 1992). A mobility assessment forms verbal commands in English. Patients who were discharged
part of a falls risk assessment as impaired mobility has been and readmitted within the study period, and patients with
identified as a falls risk factor in this population (Salgaldo et confusion or dementia that interfered with their ability to
al 1994). A mobility assessment tool commonly used in the complete the test were excluded. Patients were assessed at
hospital setting is the Timed Up and Go Test. varying times following their admission to the ward, but
always within 24 hours of being referred to physiotherapy.
The Timed Up and Go Test is a modified version of a
Discharge assessments were completed no more than 24
previously established tool, the Get Up and Go Test (Mathias
hours prior to the discharge occurring.
et al 1986). The Timed Up and Go Test is a test of basic
mobility function that is appealing in its simplicity and One hundred and sixty-six subjects were required to allow
objective nature. It involves the participant getting up out of logistic regression analysis, assuming 10% of subjects would
a chair, walking 3 m, turning, walking back to the chair and fall (Evans et al 1998), and to assure sufficient power (80%)
sitting down. Limited equipment and training are required to detect an odds ratio of 2.0 (i.e. a Timed Up and Go Test
and it can be performed at the hospital bedside. Previous score at which patients are twice as likely to fall than those
research suggests that the Timed Up and Go Test has some with a lower score) (Hsieh 1989).
predictive ability in community dwelling people (predicting
ability to walk outside alone) and it is suggested that it may Measurement tools A data collection tool was developed to
have some capacity to predict function in other settings record patient data directly from patient records. Variables
(Podsiadlo and Richardson 1991) but this is yet to be collected were:
established in an acutely hospitalised population. Since the
Timed Up and Go Test has been deemed a reliable measure of Date of admission, gender, age, and admission diagnosis
balance and functional mobility when compared to other tests Relevant co-morbidities. Co-morbidities were deemed

Australian Journal of Physiotherapy 2004 Vol. 50 249

Lindsay et al: The Timed Up and Go Test: Unable to predict falls on the acute medical ward Short Report

relevant if they had been identified in previous the use of a gait-aid, and physiotherapy intervention during
publications as being related to falls (Hill et al 2000) admission.
Timed Up and Go Test scores on initial assessment by a Results
physiotherapist and prior to discharge
Level of assistance required to complete the Timed Up Data were collected from 160 patient medical records. Of
and Go Test (independent, minimal, moderate, or these patients 105 were female and 55 were male. The mean
maximal assistance) age was 81 years (range 6599 years). Admission diagnosis
Gait aid required to complete Timed Up and Go Test was divided into eight categories for data collection with 36%
of patients admitted with a general medical condition, 21%
Details regarding patient falls during the admission with a heart condition, 14% with a stroke, and 13% with
Physiotherapy interventions that the patient received: decreased mobility. Decreased mobility was an admission
gait-aid, strengthening program, on-ward walking diagnosis documented by admitting medical doctors. This
program diagnosis may have included those people admitted with falls
or an abnormal gait pattern. Co-morbidities were also divided
Discharge destination and whether the patient re- into eight categories. Patients may have possessed more than
presented to the hospital within one week of discharge one of the co-morbidities at the time of assessment. Fifty-nine
with mobility related issues. per cent (59%) of patients were taking more than three
These data were deemed necessary to identify potential medications, 31% were experiencing pain, 29% had arthritis,
confounders that may affect both Timed Up and Go Test 18% were confused, 15% had a visual impairment, 13% were
results and whether the patient fell. It was assumed that the incontinent, and 11% were experiencing vertigo or dizziness.
Timed Up and Go Test score was derived using the testing The mean Timed Up and Go Test score on initial assessment
protocol described by Podsiadlo and Richardson (1991) and was 45.7 seconds (SD 36.4, n = 100). There was no
in which all physiotherapy staff at the study hospital were statistically significant relationship between age (p = 0.91),
trained. The Timed Up and Go Test has been found to possess gender (p = 0.25), or diagnosis (p = 0.39) and Timed Up and
both good inter- and intra-rater reliability (Posiadlo and Go Test. There was, however, a statistically significant
Richardson 1991, Lundin-Olsson et al 1998, Hansen et al relationship between the use of a gait-aid and Timed Up and
1999, Shumway-Cook et al 2000). Data related to falls during Go Test score (p = 0.001). If subjects used a walking frame,
an admission were collected directly from the incident their Timed Up and Go Test score was statistically
database that records all falls that occur involving staff, significantly slower than those who used a stick or no aid.
patients, and visitors. The mean Timed Up and Go Test score for a subject who used
Procedure A retrospective audit of medical records was a frame was 59.9 seconds (SD 42.4, n = 37).
used. An audit period of four months was chosen for the Of the 160 study subjects, 11% (n = 17) experienced a fall
anticipated size of the sample. Data were recorded on the data whilst admitted. None of the patients who fell had been
collection tool and a photocopy of the Mobility Assessment admitted with the documented diagnosis of decreased
form (MR40C) was obtained to complete data collection mobility by the admitting doctor.
without use of the identification sticker. A research assistant
(a physiotherapist) audited 10% of the medical records at the No statistically significant relationship was found between
time of data collection to ensure accuracy of collection. An Timed Up and Go Test on initial assessment and risk of
independent random data audit was conducted to assess the falling (p = 0.61, n = 141). There was also no relationship
accuracy of data entry. between age (p = 0.85, n = 160), use of a gait-aid (p = 0.38, n
= 158), or level of assistance required to complete the Timed
Ethical approval was granted by The Bendigo Health Care Up and Go Test (p = 0.22, n = 100) and risk of falling. When
Group and La Trobe University, Bendigo. all the admission diagnoses and co-morbidities, as possible
predictors of falling, were assessed via logistic regression, a
Data analysis Pearsons product-moment correlation statistically significant relationship between the presence of
coefficient was used to determine the association between incontinence and falling was identified (p = 0.001, OR = 8.7,
continuous variables (e.g., age in years and Timed Up and Go 95% CI 2.5 to 30.3). Patients with incontinence were nine
Test score in seconds). The independent samples t-test or one- times more likely to fall than those without. A trend was
way between-group ANOVA, or their non-parametric identified between the diagnosis of cancer and falling (p =
equivalents, were used to compare the means of continuous 0.09), although this was not statistically significant and was
variables. Yates continuity corrected chi-squared test was therefore not included in the logistic regression model.
used to calculate the relationship between two categorical
variables (e.g., gender and whether the patient fell). The Discussion
strength of the relationship between study variables and
falling was quantified by odds ratios and 95% confidence The Timed Up and Go Test, used in isolation, was not able to
intervals estimated with forward stepwise logistic regression. identify those patients who fell whilst admitted to hospital.
This allowed all the variables associated with falling, However, incontinence was identified as a falls risk factor.
including the main variable of interest, the Timed Up and Go This is consistent with previous research that cited altered
Test, to be assessed together and the strongest predictors of elimination patterns (Hendrich et al 1995) or incontinence
falling to be identified. The variables entered into the model (Ashton et al 1989) as a significant risk factor for falling.
included: age, gender, admission diagnosis, co-morbidities Increased falls risk may arise with the sense of urgency
(the presence of vertigo, pain, the taking of three or more associated with special toileting needs and with the
medications, arthritis, depression, incontinence and visual exacerbation of continence problems by diuretics. These
problems), Timed Up and Go Test score on admission, level medications are often altered whilst a patient is acutely
of assistance required to complete the Timed Up and Go Test, unwell, with the resultant effect being rapid changes in the

250 Australian Journal of Physiotherapy 2004 Vol. 50

 Lindsay et al: The Timed Up and Go Test: Unable to predict falls on the acute medical ward Short Report

medical and continence status of the patient. Evans et al Acknowledgments We would like to acknowledge Lisa
(1998) suggests nursing interventions that may decrease falls Hanson for auditing, and the Department of Health and
due to continence problems (Evans et al 1998) including Human Services and BHCG for funding study leave for the
placing patients with urgency closer to the toilets and toileting first author. Thanks also to Professor Meg Morris, Associate
at-risk patients regularly. Professor Helen McBurney, and Dr Keith Hill for their
encouragement to publish these results.
Due to the poor predictive ability of the Timed Up and Go
Test, we suggest it not be used to assess falls risk. These data Correspondence Robyn Lindsay, C/o Bendigo Health Care
perhaps provide support for a multidisciplinary approach to Group, Department of Physiotherapy, PO Box 126, Bendigo,
falls risk assessment (Hill et al 2000) and do not support the VIC 3552. Email: <[email protected]>.
use of the Timed Up and Go Test alone as a screening tool to
identify older patients who will fall during an acute medical References
ward admission. Berg K, Wood-Dauphinee S, Williams I and Maki B (1992):
Measuring balance in the elderly: Validation of an instrument.
Patients on the medical ward were found to be slower in Canadian Journal of Public Health 83: s7s11.
performing the Timed Up and Go Test than reports of those in Byers V, Arrington M and Finstuen K (1990): Predictive risk factors
the community (Posiadlo and Richardson 1991, Thigpen et al associated with stroke patients in acute care settings. Journal
2000, Steffan et al 2002, Wall et al 2000), but this did not of Neuroscience Nursing 22: 147154.
indicate an increased falls risk. Evans D, Hodgkinson B, Lambert L, Wood J and Kowanko I
(1998): Falls in Acute Hospitals. Adelaide, Joanna Briggs
A limitation of this study is that data were collected in only Institute for Evidence Based Nursing and Midwifery: 171.
one institution. The exclusion of patients with an altered Hansen K, Mahoney J and Palta M (1999): Risk factors for lack of
mental state may also have decreased the usefulness of these recovery of ADL independence after hospital discharge. Journal
results, given that this group is at increased risk of falling of the American Geriatrics Society 47: 360365.
(Evans et al 1998). Other limitations were that due to the Hill K, Smith R, Murray K, Sims J, Gough J, Darzins P and
nature of the study, a retrospective medical records audit, the Vransidis F (2000): An Analysis of Research on Preventing
researcher had no control over the completeness or quality of Falls and Falls in Older People: Acute Care Settings. Canberra,
National Ageing Research Institute.
data recording. This study design did however avoid the
Hawthorne effect and experimenter effects including active Hsieh FY (1989): Sample size tables for logistic regression.
Statistics in Medicine 8: 795802.
behaviour (Portney and Watkins 1993). Active behaviour
involves the subtle manipulation of variables by the Lundin-Olsson L, Nyberg L and Gustafon Y (1998): Attention,
frailty and falls: The effect of a manual task on basic mobility.
experimenter to affect outcomes. Another shortcoming was Journal of the American Geriatrics Society 46: 758761.
that the incident database used to document the circumstances
Mathias S, Nayak USL and Isaacs B (1986): Balance in elderly
of the falls suffered by the sample provided limited patients: the get up and go test. Archives of Physical Medicine
information regarding the intrinsic and extrinsic factors and Rehabilitation 67: 387389.
surrounding the falls. Podsiadlo D and Richardson S (1991): The Timed Up and Go: A
test of basic functional mobility for frail elderly persons. Journal
The published studies investigating objective measures of of American Geriatrics Society 39: 142148.
altered mobility and falls in the acute hospital setting are Salgado R, Lord S, Packer J and Ehrlich F (1994): Factors
limited. Paper-based tools are not efficient or effective in associated with falling in elderly hospital patients. Gerontology
identifying falls risk (Evans et al 1998), yet there are no 40: 325331.
validated objective measures to replace their use. A key Shumway-Cook A, Brauer S and Woollacott M (2000): Predicting
recommendation of this study is that further trials of objective the probability for falls in community-dwelling older adults using
measures of mobility in acute hospital settings be conducted the Timed Up and Go Test. Physical Therapy 80: 896903.
and published. Steffan T, Hacker T, and Mollinger L (2002): Age- and gender-
related test performance in community dwelling elderly people:
This study has identified preliminary data regarding normal Six minute walk test, Berg Balance Scale, Timed Up and Go
Timed Up and Go Test score values for older acute medical Test and gait speeds. Physical Therapy 82: 128137.
patients. To identify any clear relationships between Timed Thigpen M, Light K, Creel G and Flynn S (2000): Turning difficulty
Up and Go Test scores and falls, future research needs to characteristics of adults aged 65 years or older. Physical
include a larger sample in a multi-centre design. It is Therapy 80: 11741187.
suggested that the study design be prospective with an ability Wall JC, Bell C, Campbell S and Davis J (2000): The Timed Get-
to contact patients after discharge to gather information up-and-Go Test revisited: Measurement of the component
tasks. Journal of Rehabilitation Research and Development
regarding falls in the first week following separation from the
37: 109113.
hospital. The increased falls risk for patients with the co-
morbidity of incontinence is a relationship that could be
explored further by examining this cohort specifically along
with their Timed Up and Go Test results.
It is important that objective measures of altered mobility and
falls risk are developed to increase the accuracy of falls risk
identification within hospitals.

Australian Journal of Physiotherapy 2004 Vol. 50 251