Transcript Marianjoy Rehabilitation Hospital Fall Risk Assessment

Marianjoy Rehabilitation Hospital
Fall Risk Assessment Tool Project
Donna Pilkington, RN, MSML, CRRN
Kathleen Ruroede, PhD, MEd, RN
Nancy Cutler, RN, MS, CRRN
Fall Risk Assessment Literature
• Morse Fall Scale
• Marianjoy Fall Risk Assessment
Morse Fall Scale
• The Morse Fall Scale (MFS) is a rapid and simple
method of assessing a patient’s likelihood of falling.
• The MFS is used widely in acute care settings, both
in the hospital and long term care inpatient settings.
• It consists of six variables that are quick and easy
to score, and it has been shown to have predictive
validity and interrater reliability.
• A large majority of nurses (82.9%) rate the scale as
“quick and easy to use,” and
• 54% estimated that it took less than 3 minutes to
rate a patient.
Morse Fall Scale Indicators
1.
History of falling with in
three months
No = 0
Yes = 25
2. Secondary Diagnosis
No =
0
Yes = 15
3. Ambulatory Aid
Bed rest/nurse assist = 0
Crutches/cane/walker =15
Furniture
= 30
4. IV/Heparin Lock
No = 0
Yes = 20
5. Gait/Transferring
Normal/bedrest/immobile = 0
Weak
= 10
Impaired
= 20
6. Mental Status
Oriented to own ability = 0
Forgets limitations
= 15
Scoring the Morse Fall Scale
Risk Level
MFS score
Action
________________________________________
No Risk
0 – 24
Basic Care
Low Risk
25 – 50
Standard
Fall
Precautions
High Risk
> 51
High Risk
Precautions
Marianjoy Fall Risk Assessment
•
Altered elimination patterns 10
•
Lower extremity hemiparesis
10
•
Unilateral neglect
10
•
Activity intolerance
10
•
Impaired cognition
20
•
Episodes of dizziness/seizures
10
•
Sensory deficits (hearing,
•
Special medications (narcotics,
psychotropic, hypnotic,
antidepressants etc.)
5
sight, touch)
5
•
Agitation
•
Impaired mobility
•
History of previous falls
20
•
Impulsiveness
20
•
Communication deficits
20
20
5
•High Risk: >60 points
•
Diuretics, and drugs that
increase GI motility
5
•
Upper extremity paresis
5
•
Age greater that 65 or less
than 16
Place Patient in Caution Club
5
Guiding Question?
Is the Marianjoy Fall Risk
Assessment a valid and reliable
method for predicting rehabilitation
patient fall events if it is properly
scored at admission?
Description of Research Study
• Pilot study of 50 patients
– 25 patients who had fallen
– 25 matched patients who had not fallen
• Dependent variable fall status
• Independent variables
– Caution Club status
– Admission FIM total score
– Modified admission Berg Balance total score
– Admission fall risk assessment
Pilot Study Results
• Patients significantly differed on Berg,
FIM, and fall risk assessment scale
• Five items found to separate fall groups
– History of falls
– Unilateral neglect
– Episodes of dizziness / seizures
– Special medications
– Diuretics and drugs that increase GI motility
– Sensory deficits
Always be alert
for a new and
creative idea...
You never
know what’s in
your grasp
Replicated Study with a Larger Sample
•
•
•
•
2005 data used
Total N = 450 patients included
125 patients with documented fall status
325 patients who had not fallen were
randomly selected from dataset
• 232 patients were on caution club status
• 218 patients not on caution club status
Replicated Study with a Larger Sample
• Hypotheses tested
– Patients did not significantly differ on fall
status for:
• Fall assessment
• Admission FIM Score
• Modified Berg Balance Score
• Age
Replicated Study with a Larger Sample
• Statistical Procedures
– Descriptive statistics
– Sensitivity and specificity on original scale
– Sensitivity and specificity on converted
dichotomous scale
– Item analysis on dichotomous scale that
separate fallers from non-fallers
– Total of 9 items discriminate groups
Replicated Study with a Larger Sample
• Statistical Procedures
– Validity procedures using factor analysis
(component analysis)
– Reliability analysis using Cronbach’s Alpha
– Logistic regression to develop predictive model
of fall status
– Development of new “Caution Club” threshold
value – New Threshold Cut Score = > 4
Always be ready for any surprises
while working on the project
Results – Descriptive Statistics
Descriptive Statistics
Age
Fall
No
Yes
N
325
125
Mean
65.60
62.27
Std.
Deviation
16.793
17.596
Gender
Valid
Female
Male
Fall
No
Yes
No
Yes
Frequency
179
52
146
73
Percent
55.1
41.6
44.9
58.4
Results – Inferential Statistics
Ranks Original Fall Assessment by Fall Status
Initial Fall Risk Assmnt
Fall
No
Yes
Total
N
325
125
450
Mean Rank
189.19
319.90
Test Statisticsa
Mann-Whitney U
Wilcoxon W
Z
Asymp. Sig. (2-tailed)
Initial Fall
Risk Assmnt
8513.000
61488.000
-9.561
.000
a. Grouping Variable: Fall
Sum of Ranks
61488.00
39987.00
Results – Inferential Statistics
Ranks Original Fall Assessment by Caution Club Status
Initial Fall Risk Assmnt
Caution Club
No
Yes
Total
N
218
232
450
Mean Rank
109.50
334.50
Test Statisticsa
Mann-Whitney U
Wilcoxon W
Z
Asymp. Sig. (2-tailed)
Initial Fall
Risk Assmnt
.000
23871.000
-18.365
.000
a. Grouping Variable: Caution Club
Sum of Ranks
23871.00
77604.00
Results – Inferential Statistics
Ranks
Total Berg
FIM Total Admission
without tubshower
Age
NewCaution
ClubWeig
htCutat3
No
Yes
Total
No
Yes
Total
No
Yes
Total
N
214
236
450
214
236
450
214
236
450
Mean Rank
265.93
188.84
287.94
168.88
223.75
227.09
Berg and FIM Significantly Differ, but Age does not significantly differ
Results from Item Analysis
• Nine items found to discriminate fall groups
– History of Falls (Weight 2)
– Impulsiveness (Weight 2)
– Communication Deficits
– Altered Elimination Patterns
– Unilateral Neglect
– Lower Extremity Hemiparesis
– Upper Extremity Hemiparesis
– Special Medications
– Diuretics and Drugs that Increase GI Mobility
Factor Analysis and Reliability
Rotated Component Matrixa
• Three Components
Extracted
• 55% Total Explained
Variance in Model
Reliability Statistics
Cronbach's
Alpha
.558
N of Items
9
1
UEExtremHemipDichot
LEHemiparDichot
UnilatNeglDichot
HxFalls2
Impuls2
CommunDeficDichot
AlterEliminDichot
SpecialMedsDichot
DiureticsDichot
.841
.826
.710
-.110
.104
.477
.212
-.038
-.048
Component
2
.045
-.008
.211
.739
.722
.504
.354
-.003
.097
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
a. Rotation converged in 5 iterations.
3
-.014
-.092
-.043
-.069
.087
.061
.080
.813
.768
Logistic Regression Model
Variables in the Equation
B
Step
a
1
Total Berg
FIIM Total Adm
New Caution Club
Age
Constant
.015
-.033
-1.398
-.011
1.690
S.E.
.035
.009
.280
.007
.520
Wald
.188
13.646
25.024
2.837
10.555
Sig .
.665
.000
.000
.092
.001
Exp(B)
1.015
.968
.247
.989
5.419
a. Variable(s) entered on step 1: totberg, FIIMtotadm, NewCautionClubWeightCut3, Age.
• R Square Value .253
Results from Crosstabulations
Original Caution Club Status by Fall Crosstabulation
Fall
Caution
Club
No
Yes
Total
Count
% of Total
Count
% of Total
Count
% of Total
No
200
44.4%
125
27.8%
325
72.2%
Yes
18
4.0%
107
23.8%
125
27.8%
Total
218
48.4%
232
51.6%
450
100.0%
New Caution Club Status by Fall Crosstabulation
Fall
NewCaution
ClubWeightCutat3
No
Yes
Total
Count
% of Total
Count
% of Total
Count
% of Total
No
191
42.4%
134
29.8%
325
72.2%
Yes
23
5.1%
102
22.7%
125
27.8%
Total
214
47.6%
236
52.4%
450
100.0%
Sensitivity and Specificity
Fall
+
a
+
Caution
Club
b
102
c
-
134
d
23
125
( a+c )
Sensitivity
Specificity
False Negative
False Positive
PPV
NPV
-
191
236
( a+b )
214
( c+d )
325
( b+d )
= a / (a + c) = 102 / 125 = .82
= d / (b + d) = 191 / 325= .59
= c / (a + c) = 23 / 125 = .18
= b / (b + d) = 134 / 325 = .41
= a / (a + b) = 102 / 236 = .43
= d / (c + d) = 191 / 214 = .89
Odds and Odds Ratio
• True Odds Ratio = 6.25
• This can be interpreted to mean that a
patient who is on caution club status was
6.2 times more likely to incur a fall than a
patient who was not on caution club
status.
Odds and Odds Ratio
• Relative Risk of a Fall = 3.9
• This can be interpreted to mean that the
risk of patients on caution club status are
3.9 times more likely to occur than those
patients who were not on caution club
status.
Don't get off
strategy and
stay focused
Conclusions and
Recommendations