BJMB
Brazilian Journal of Motor Behavior
Special issue:
“Control of Gait and Posture: a tribute to Professor Lilian T. B.
Gobbi”
!
Batistela, Rinaldi,
Moraes
2023
VOL.17
N.4
126 of 133
Mini-BESTest cutoff points for classifying fallers and non-fallers female older adults
ROSANGELA A. BATISTELA
1
| NATALIA M. RINALDI
2
| RENATO MORAES
1,3
1
Biomechanics and Motor Control Lab, School of Physical Education and Sport of Ribeirão Preto, University of São Paulo, Ribeirão Preto, SP, Brazil
2
Center of Physical Education and Sports, Department of Sports, Federal University of Espírito Santo, Vitória, ES, Brazil
3
Graduate Program in Rehabilitation and Functional Performance, Ribeirão Preto Medical School, University of São Paulo, Ribeirão Preto, SP, Brazil
Correspondence to:!Rosangela Alice Batistela, Sc.D., Escola de Educação Física e Esporte de Ribeirão Preto, Universidade de São Paulo, Av. Bandeirantes, 3900,
Ribeirão Preto, SP, 14040-907, Brasil, Phone: +55 16 99617-8613
email: batistela.rosangela@gmail.com
https://doi.org/10.20338/bjmb.v17i4.354
HIGHLIGHTS
• We examined the accuracy of the Mini-BESTest for
identifying falls in older women.
• We established the Mini-BESTest cutoff scores for
fallers in different age groups.
• The Mini-BESTest is a highly accurate tool for
identifying falls in older Brazilian women.
• The Mini-BESTest cutoff score to classify the fallers
was 26 for 65-69 years old.
• The Mini-BESTest cutoff score to classify the fallers
was 24 for 70+ years old.
ABBREVIATIONS
AUC Area under the ROC curve
Mini-BESTest Mini-Balance Evaluation Systems Test
MMSE Mini-Mental State Examination
PD Parkinson’s disease
ROC Receiver operating characteristic
-LR Negative likelihood ratios
+LR Positive likelihood ratios
PUBLICATION DATA
Received 23 02 2023
Accepted 19 06 2023
Published 20 06 2023
BACKGROUND: The Mini-Balance Evaluation Systems Test (Mini-BESTest) is an efficient
screening tool healthcare professionals use to predict the risk of falls in older adults. However,
the Mini-BESTest cutoff scores to classify fallers and non-fallers were established using men
and women in the same sample. Considering the higher number and prevalence of falls in
older women, it is important to know the Mini-BESTest accuracy and the cutoff score
specifically for this population.
AIM: We examined the capability and accuracy of the Mini-BESTest for identifying fallers and
non-fallers female older adults without neurological impairments and established the cutoff
scores according to different age groups.
METHOD: Eighty-one female older adults were classified into fallers (n=40) and non-fallers
(n=41) groups according to their retrospective history of falls in the last 12 months. Fallers and
non-fallers were divided into three age groups according to the following ranges: 65-69 years,
70-74 years, and 75+ years. We used the receiver operating characteristic (ROC) curves to
determine the relative performances of the Mini-BESTest score for classifying participants with
and without a history of falls.
RESULTS: The Mini-BESTest is a good and highly accurate tool for identifying female
Brazilian fallers and non-fallers. The Mini-BESTest cutoff scores established to classify fallers
and non-fallers female older adults in the different age groups were 26 points for 65-69 years
and 24 points for 70-74 years and 75+ years.
INTERPRETATION: The Mini-BESTest is an important tool that health professionals in clinical
practice can use to estimate the risk of falls for older Brazilian women.
KEYWORDS: Older women | Falls | Accuracy | Sensitivity | Specificity | Mini-BESTest
INTRODUCTION
Falls are a major health problem among older adults worldwide and have been recognized as the second leading cause of
accidental or unintentional injury deaths in the older population
1,2
. Indeed, approximately one-third of individuals aged over 65 years are
reported to fall each year, with a greater risk for women, becoming even more recurrent with increasing age
1,3,4
. In Brazil, a recent
population-based study reinforced this prevalence showing that 25.1% of older adults (60-75 years or older) reported at least one fall in
the last 12 months, with a higher prevalence among older women (30.2%) than older men (18.4%)
5
. Previous studies have also
evidenced that older women do not have just an increased risk of falls but experience more falls than men
1,3,4-7
, showing a worse quality
of balance
8,9,10
and a greater perception of their risk of falling than men
10
.
Studies have evidenced that, after an initial fall, older adults have a higher risk of falling again
6,7
. The occurrence of a first fall is
a predictor of future falls
6,11
, and these falls may result in fractures
12
, frequent injury-related hospitalizations, physical disability, reduced
functionality, less independence, fear of new falls, loss of the ability to perform daily living activities, poor quality of life, reduced survival
of those who experienced falls, burdens on caregivers and society as a whole
13,14
. Falls represent a leading cause of hospital admission
in older adults leading to substantial healthcare costs
2,4
. Thus, avoiding the first fall can contribute to maintaining functional
independence since roughly half of older adults fall recurrently after the first fall
1,6
, which can help decrease health care costs and,
ultimately, death in older adults
6,15
. These severe consequences and the adverse impacts of falls emphasize the need to identify and
BJMB! ! ! ! ! ! ! ! !
Brazilian(Journal(of(Motor(Behavior(
(
Batistela, Rinaldi,
Moraes
2023
VOL.17
N.4
127 of 133
Special issue:
“Control of Gait and Posture: a tribute to Professor Lilian T. B. Gobbi”
classify older people at risk of falls, apply an appropriate clinical assessment tool with specific cutoff scores, and then select appropriate
prevention strategies
14,16
.
The literature on falls epidemiology and risk factors for falls among older adults has grown considerably in recent
decades
3,7,17,18,19
. Many factors, including female gender, advancing age, cognitive deficits, reduced physical activity level, and obesity,
have been associated with a higher risk of falling
20
. Thus, screening is essential to identify these factors associated with an increase in
the number of falls in older people
3,7
.
Balance and gait deficits are other significant predictors of falls in older adults
1,21,22
. Studies have suggested to healthcare
professionals the Mini-Balance Evaluation Systems Test (Mini-BESTest) as an efficient screening tool to identify older adults with higher
fall risk and assess the components of the postural control system, functional balance and gait stability responsible for the occurrence of
falls in older adults
14,15,23
. The literature contains cutoff scores for the Mini-BESTest for individuals with Parkinson’s disease (PD)
24,25
,
individuals with stroke
26
, and healthy older adults
14,15
.
Yingyongyudha and colleagues
14
compared the areas under the receiver operating characteristic (ROC) curves of the Mini-
BESTest, BESTest, Berg Balance Scale, and Timed Up and Go Test to identify older adults with a history of falls without neurological
problems. The authors suggested a single score of 16 (out of 28) as the cutoff score for the Mini-BESTest for identifying older adults with
a history of falls. In addition, the sample was composed of male and female community-dwelling older adults in Thailand. However,
whether these values would be generalizable to the Brazilian population, specifically for female older adults with different chronological
ages, is unknown. For the Brazilian people, Magnani et al.
15
also analyzed the areas under the ROC curves of the BESTest and Mini-
BESTest to identify the reference values of these tests to identify fallers in community-dwelling Brazilian older adults of different age
groups (60-102 years). Their results showed that the cutoff scores to identify older adults with fall risk according to the Mini-BESTest in
different age groups were 25 points for 60-69 years of age, 23 points for 70-79 years of age, 22 points for 80-89 years of age, and 17
points for 90 years of age or older. However, although in a lower number, this study also included males in the sample and generalized
the same cutoff scores for both genders. Considering the higher number and prevalence of falls in female older adults and the lack of
specific cutoff scores for Brazilian older women, it is essential to know the Mini-BESTest accuracy and the cutoff score specifically for this
population. Therefore, we examined the capability and accuracy of the Mini-BESTest for identifying fallers and non-fallers female older
adults without known neurological impairments. Based on this, we established cutoff scores for classifying Brazilian fallers and non-fallers
female older adults in age groups 65-69 years, 70-74 years, and 75+ years. This knowledge is fundamental for an appropriate clinical
application of the Mini-BESTest for fall prevention and balance rehabilitation in this population.
METHODS
Participants
Eighty-one female older adults volunteered for this study. They were identified as fallers if they experienced a fall in the 12
months preceding the data collection. We used the fall definition proposed by Beauchet et al.
27
“as unintentionally coming to rest on the
ground, floor, or other lower level”. Participants signed the informed consent form approved by the local ethics committee. Before the
Mini-BESTest
15,28,29
assessment, the participants were screened by filling out a questionnaire to check their history of falls, health status,
physical activity level (Modified Baecke Questionnaire)
30
, cognitive functions (Mini-Mental State Examination, MMSE)
31,32
, and
anthropometric parameters. We included community-dwelling older adults who could walk independently without using any assistive
device (cane or walker). Participants were excluded if they had a stroke, neurological disease, or other diseases that could compromise
their stability.
Procedures
The Mini-BESTest assesses functional balance, specifically the transitions/anticipatory postural control, reactive postural
control, sensory orientation, and gait stability. The Mini-BESTest comprises fourteen tests with a maximum score of 28 points. For each
test, the score varies from 0 (the lowest functional level) to 2 (the highest functional level).
Statistical analyses
Data were analyzed considering the entire sample and separate age groups: 65-69 years, 70-74 years, and 75+ years. The
receiver operating characteristic (ROC) curves were used to determine the relative performances of Mini-BESTest scores for classifying
participants with and without a history of falls. The accuracy of the Mini-BESTest for discriminating participants with and without a history
of falls was assessed using the area under the ROC curve (AUC). An AUC value of 0.9 and greater indicates high accuracy, between 0.7
to 0.9 indicates moderate accuracy, between 0.5 to 0.7 indicates low accuracy, and 0.5 and less indicates a result due to chance
14,33
.
The cutoff point was defined by selecting the best score between high sensitivity and high specificity
34
. Sensitivity and specificity values
were used to calculate positive and negative likelihood ratios according to methods used in previous studies
35,36
. Positive likelihood ratios
BJMB! ! ! ! ! ! ! ! !
Brazilian(Journal(of(Motor(Behavior(
(
Batistela, Rinaldi,
Moraes
2023
VOL.17
N.4
128 of 133
Special issue:
“Control of Gait and Posture: a tribute to Professor Lilian T. B. Gobbi”
(+LR) were calculated as Sensitivity/1 − Specificity. Negative likelihood ratios (-LR) were calculated as 1 − Sensitivity/Specificity. A
positive likelihood ratio (+LR) and a negative likelihood ratio (-LR) were also calculated for each age group. An +LR higher than 5 and an
-LR lower than 0.2 indicate that the test is useful due to its high probability of correctly identifying participants with and without a history of
falls, but LR values very close to 1.0 indicate that the test is useless, as the probability of correctly and incorrectly identifying participants
with a history of falls is the same
14,34
. A likelihood ratio of 1.0 means the true-positive and false-positive (or true-negative and false-
negative) rates are the same, making the test results useless
31
. For this reason, a +LR indicates the clinical usefulness of a positive test
result: the larger the +LR value above 1.0, the more valuable the positive result. The −LR indicates the usefulness of a negative test
result: the smaller the value below 1.0, the more valuable the negative result
35,36
. We provided the power for the ROC curves analyses
based on the calculator available at http://riskcalc.org:3838/samplesize/
37
. The power calculation determined that the minimum sample
size was 18 for an AUC of 0.95 (3 fallers and 15 non-fallers) and 25 for an AUC of 0.99 (4 fallers and 21 non-fallers) participants. These
values were obtained with an alpha level of 0.05 (2-tailed), a power of 0.8, a null hypothesis of AUC ≤ 0.5, and a prevalence (ratio of
positive cases/total sample size) of 0.16 (i.e., +LR = 13.01 and total sample size = 81)
37
.
Using the data of the entire sample, we ran a multiple regression analysis (stepwise method) with the number of falls as the
dependent variable. The following independent variables were included in the regression analysis: age, height, body mass, MMSE, Mini-
BESTest, and physical activity level. For all analyses, the significance level was set at p≤0.05.
RESULTS
Entire sample
Participants were divided into two groups based on their retrospective history of falls: non-fallers (n=41) and fallers (n=40).
Table 1 presents the data for both groups. Except for the Mini-BESTest, no statistical difference existed between groups for all other
variables assessed. The Mini-BESTest score was smaller for the fallers than for the non-fallers. For the entire sample, 51.2% did not fall
in the last six months, 18.6% felt once, 16.3% felt twice, and 14.0% felt three or more times.
Table 1. Mean and standard deviation (in parentheses) of the assessed variables for non-fallers and fallers across age groups.
Variables
Non-Fallers
(n=41)
Fallers
(n=40)
p-value
Age (years)
72.1 (4.6)
73.8 (5.0)
0.125
a
Body Mass (kg)
64.7 (12.4)
67.3 (12.2)
0.343
a
Height (m)
1.56 (0.06)
1.54 (0.06)
0.176
a
Mini-Mental State Exam (score)
28.1 (1.7)
26.9 (3.1)
0.072
b
Mini-BESTest (score)
26.5 (1.8)
19.6 (3.4)
≤0.0001
b
Physical Activity Level (score)
5.1 (2.4)
4.9 (2.5)
0.688
a
Number of Falls
---
2.2 (1.7)
---
a
One-way ANOVA;
b
Mann-Whitney Test; Bolded p-values indicate statistical significance.
The ROC curve analysis indicated a high accuracy in classifying fallers from non-fallers, as the area under the curve was
higher than 0.9 (Figure 1A and Table 2). Both sensitivity and specificity were high, and the cutoff point for the Mini-BESTest to classify
fallers from non-fallers was 24 points (Table 2).
Table 2. Parameters of the receiver operating characteristic (ROC) curve for the Mini-BESTest as a classifier of older adults with a
history of falls.
Groups
AUC
95% CI
p-value
Sensitivity
(%)
Specificity
(%)
+LR
-LR
Cutoff
Point
All sample
0.97
0.93-1.00
≤0.0001
95
93
13.01
0.05
24
65-69 y
1.00
1.00-1.00
≤0.0001
100
100
---
a
0.00
26
70-74 y
0.99
0.96-1.00
≤0.0001
92
93
13.78
0.08
24
75+ y
0.95
0.88-1.00
≤0.0001
100
85
6.49
0.00
24
AUC: area under the curve; CI: confidence interval; +LR: positive likelihood ratio; -LR: negative likelihood ratio;
a
It was not possible to
compute as the denominator was zero; Bolded p-values indicate statistical significance.
The regression analysis identified age (β=-0.046; p=0.029) and Mini-BESTest (β=-0.235; p≤0.0001) as predictors for the
number of falls (R
2
=0.57). Based on β, Mini-BESTest exhibited a larger predictive power than age for multiple falls risk. Thus, the smaller
the Mini-BESTest score, the higher the risk of multiple falls (Figure 2).