The EBQ is a novel behavioral psychometric scale for clinical evaluation of treatment effectiveness in treating overweight and obese patients with diet, lifestyle modification and pharmacotherapy.

1. The
Ea'ng
Behavior
Ques'onnaire
©,
a
Novel
Clinical
Examina'on
ASBP
Spring
Conference
2014
Philadelphia
Ed
J.
Hendricks,
M.D.,
FASBP

2. Background
-­‐
I
When
asked
“What
is
the
effect
of
the
drug
?”
obese
pa'ents
treated
with
an'-­‐obesity
drugs
offer
a
wide
variety
of
answers
such
as:
• “I
don’t
eat
as
much.”
• “I
can
stop
ea'ng.”
• “I
don’t
graze
all
day
and
night.”
• “I’m
not
hungry
as
soon
as
I
stop
ea'ng.”
• “I’m
normal”
(in
respect
to
ea'ng).

3. Background
-­‐
II
1. Obese
pa'ents
have
ea'ng
behaviors
that
have
led
to
weight
gain.
2. An'-­‐obesity
drugs
change
ea'ng
behaviors
inducing
compara've
hypophagia.
3. Treatment-­‐induced
ea'ng
behavior
changes
are
proximate
to
weight
loss.
4. Simplis'c
descrip'ons,
but
if
true
we
hypothesize
that
a
metric
of
ea'ng
behavior
could
be
a
useful
clinical
tool.

4. Measurements
of
Behavior
• Method:
provide
s'mulus
–
measure
reac'on.
• Measurement
of
reac'on
can
be
done
either
by
tes'ng
administrator/observer
or
by
test
subject.
• The
la`er
method,
termed
psychometric
scale
tes'ng,
is
more
widely
used.
• Psychometric
tes'ng
can
be
confounded
because
the
measurements
depend
on
a
subjec've
assessment
by
the
person
tested.

5. Test
Desirable
Characteris'cs
• Discriminate
between
untreated
and
treated
pa'ents.
• Good
test
re-­‐test
reliability.
• Rapid
test
comple'on
by
pa'ent.
• Rapid
test
assessment
by
clinician.
• Real
number,
parametric
data;
not
ordinal
or
non-­‐parametric
data.
• Ques'ons
relate
to
treatment-­‐induced
changes.

6. Some
Scales
of
Ea'ng
Behavior
• Three
Factor
Ea'ng
Ques'onnaire
(TFEQ)
– Stunkard
1985
• Food
Preference
Ques'onnaire
(FPQ)
– Geiselman
1998
• Food-­‐Craving
Inventory
(FCI)
– White
2002
• Power
of
Food
Scale
(PFS)
– Lowe
2009

7. Disadvantages
of
Previous
Scales
• Ques'ons
do
not
necessarily
relate
to
treatment-­‐induced
changes.
• Designed
for
laboratory
tes'ng.
• Lengthy
tes'ng
process.
• Evalua'on
of
results
'me-­‐consuming.
• Lickert-­‐like
answer
structure
producing
ordinal,
non-­‐parametric
data.
• Non-­‐parametric
sta's'cal
analysis.

8. EBQ
Design
• Ques'ons
taken
from
pa'ent
descrip'ons
of
drug
effects.
• Ques'ons
phrased
in
simple
sentences.
• Visual
Analog
Scale;
parametric
data.
• Pa'ents
answer
ques'ons
by
marking
a
100
millimeter
line
under
each
ques'on.
• Scored
by
measuring
mm
from
lej
end.
• Ques'on
8,
reverse;
measured
from
right
end.

9. VAS
Scale

13. EBQ
Ques'ons
• 1.
Are
you
preoccupied
with
thoughts
of
food
or
ea'ng?
• 2.
Do
you
eat
to
comfort
yourself?
• 3.
Do
you
crave
any
specific
foods?
• 4.
Once
you
start
ea'ng,
do
you
find
it
hard
to
stop?
• 5.
Do
you
find
it
difficult
to
s'ck
to
an
ea'ng
plan?

14. EBQ
Ques'ons
• 6.
Do
you
eat
rapidly,
more
rapidly
than
those
around
you?
• 7.
Do
you
“graze”
or
eat
con'nually
during
any
part
of
a
24-­‐hour
day?
• 8.
Are
you
in
control
of
your
ea'ng?
(Reverse)
• 9.
Do
you
eat
more
when
under
stress?
• 10.
Do
you
eat
more
during
highly
emo'onal
'mes?

15. Study
Design
• Observa'onal
prospec've
study.
• Non-­‐randomized;
pa'ents
allowed
to
select
treatment
program.
• Non-­‐blinded;
physician
and
pa'ent
completely
aware
of
treatment
details.
• Sta's'cal
analysis:
– Normally
distributed
data
è
T-­‐test.
– Non-­‐normal
data
è
Wilcoxen
signed
ranks
test

16. Study
Treatment
Methods
• Diet
–
VLCKD,
Very
Low
Carbohydrate
Ketogenic
Diet
– Protein
1.5-­‐2.0
g/ideal
wt./day
– ≤
40
g
carbohydrate/day
• Behavior
Modifica'on
– One-­‐on-­‐one
pa'ent
and
prac''oner
at
every
encounter
– Focus
on
ea'ng
and
exercise
behaviors
• Pharmacotherapy
– Phentermine
mono-­‐therapy

17. TREATMENT
EXPECTATIONS
Study
treatment
methods
are
standard
prac'ce
in
the
private
prac'ce
seqng
where
this
study
was
conducted.
Treatment
results
with
this
method
have
been
published
previously:
Hendricks
EJ,
et
al.
Obesity
(Silver
Spring)
2011;19:
2351-­‐2360.

18. -­‐45.0%
-­‐40.0%
-­‐35.0%
-­‐30.0%
-­‐25.0%
-­‐20.0%
-­‐15.0%
-­‐10.0%
-­‐5.0%
0.0%
5.0%
1
8
15
22
29
36
43
50
57
64
71
78
85
92
99
106
113
120
127
134
141
148
155
162
169
Percent
Weight
Loss
Individual
PaBent
Weight
Loss
at
52
Weeks
Hendricks,
Obesity
2011;
19:2351-­‐2360.

19. -­‐120
-­‐100
-­‐80
-­‐60
-­‐40
-­‐20
0
20
1
5
9
13
17
21
25
29
33
37
41
45
49
53
57
61
65
69
73
77
81
85
89
93
97
101
105
109
113
117
121
125
129
133
137
141
145
149
153
157
161
165
169
173
1
Year
Wt.
Loss
Pounds
N
=
175
Mean
Weight
Loss
=
40
pounds
Std.
Dev.
=
25
Hendricks,
Obesity
2011;
19:2351-­‐2360.

20. 0
5
10
15
20
25
30
35
40
45
Number
of
PaBents
Loss
-­‐
Pounds
1
Year
Weight
Loss
DistribuBon

21. Variability
of
Response
to
Roux-­‐en-­‐Y
Gastric
Bypass
Hatoum,
J
Clin
Endocrinol
Metab
2011;
96:
E1630.

22. -­‐30.0
-­‐25.0
-­‐20.0
-­‐15.0
-­‐10.0
-­‐5.0
0.0
week
1
2
3
4
8
12
26
40
52
%
Weight
Loss
Phentermine
+
VLCKD
Treated
1
Year
%
Weight
Loss
Hendricks,
Obesity
2011;
19:2351-­‐2360
Systolic
BP
mm
Hg

23. -­‐30.0
-­‐25.0
-­‐20.0
-­‐15.0
-­‐10.0
-­‐5.0
0.0
%
Weight
loss
Phentermine
+
VLCKD
then
LCD
Treated
8
Years
Systolic
BP
mm
Hg
%
Weight
Loss

24. Average
Weight
Loss
vs.
Rx
Week
Week
%
Lbs.
1
-­‐3.2
-­‐6.8
2
-­‐5.0
-­‐10.6
3
-­‐6.4
-­‐13.8
4
-­‐8.0
-­‐17.2
8
-­‐12.0
-­‐25.8
Week
%
Lbs.
12
-­‐15.1
-­‐32.6
26
-­‐18.9
-­‐41.5
40
-­‐18.7
-­‐42.0
52
-­‐17.6
-­‐39.7
104
-­‐12.7
-­‐28.8

25. STUDY
DATA

26. Study
Pa'ent
Selec'on
• Type
A:
New
pa'ents
star'ng
VLCKD
and
phentermine.
• Type
B:
Previous
pa'ents,
restar'ng
VLCKD
and
phentermine
ajer
a
treatment
hiatus.
• Type
C:
Current
pa'ents,
LCD
+
drug,
treatment
sa'sfactory,
no
change
needed.
• Type
D:
Current
pa'ents,
LCD
+
drug,
treatment
unsa'sfactory,
change
needed.

27. Criteria
for
Rx
Altera'on
• Rx
Change
Needed:
– Weight
loss
less
than
expected
– Weight
plateau
reached
sooner
than
expected
– Weight
increase
on
maintenance
– Drug
“doesn’t
work
as
well
as
before.”
• No
Rx
Change
Needed
– Expected
weight
loss
achieved
– Stable
maintenance

28. Study
Demographics
• Pa'ents
Tested:
374
• Female
86%;
Male
14%
• Weight
196.2
(±45.4)
pounds
• BMI
33.2
(±6.0)
Kg/m2
• Race
%:
White/Hispanic/Black/Asian:
92/6/1/1

29. Long-­‐term
Phentermine
Rx
Dura'on
PaBent
N
MEAN
(YRS)
RANGE
(YRS)
A
58
N.A.
N.A.
B,
C,
D
316
6.0
0.1
–
20.5
Prior
Report
(1)
117
8.4
1.1
–
21.1
Prior
Report
(2)
269
-­‐
0.25
–
12.0
(1) Hendricks,
Int
J
Obes
2014;
38:
292-­‐298.
(2) Hendricks,
Obesity
2011;
19:2351-­‐2360

30. Normal
Distribu'on
Small
Overlap
ê

31. 0
5
10
15
20
25
30
35
5
15
25
35
45
55
65
75
85
95
Number
of
PaBents
DistribuBon
of
EBQ
Scores
Untreated
N
=
217
Mean
(SD)
62.0
(13.6)
Treated
N
=
197
Mean
(SD)
36.9
(15.7)

32. Ini'al
EBQ
Scores
(P1)
PaBent
Type
N
Mean
(SD)
T-­‐Test:
vs
Type
A
A
–
New,
Untreated
58
60.8
(10.4)
B
-­‐
Restart
ajer
Treatment
hiatus
159
62.4
(14.5)
0.4305
C
-­‐
Treated
No
change
needed
92
39.3
(14.7)
1.8
x
10-­‐17
D
–
Treated,
change
needed
65
55.0
(14.2)
0.0114

33. EBQ
Scores
P1
vs
P2
PaBent
Type
N
P1
Mean
(SD)
P2
Mean
(SD)
Δ1-­‐2
(P1-­‐P2)
T-­‐test
P1
v
P2
P
=
A.
New
43
61.3
(±11.0)
28.1
(±15.9)
33.2
(±17.4)
1.6x10-­‐18
B.
Restart
60
65.1
(±14.2)
40.1
(±14.8)
24.9
(±18.4)
4.7x10-­‐16
C.
no
change
29
37.4
(±11.5)
39.5
(±11.9)
-­‐2.1
(±8.9)
0.4970
D.
need
change
24
59.8
(±13.8)
40.3
(±15.8)
19.5
(±15.2)
3.8x10-­‐5

34. Days
Between
P1
&
P2
EBQ
PaBent
Type
Interval
(SD)
Rx
Plan
A.
New
Pa'ent
11.4
(±
7.2)
7
B.
Old
pa'ent,
previously
treated
20.1
(±
13.1)
7
-­‐
14
C.
Under
Treatment,
no
change
needed
56.6
(±
23.8)
90
D.
Under
Treatment,
change
needed
22.5
(±
12.3)
30

35. Single
Ques'on
T-­‐test
P1
vs
P2
QuesBon
P
1
2
x
10-­‐9
2
4
x
10-­‐14
3
1
x
10-­‐13
4
1
x
10-­‐13
5
1
x
10-­‐11
6
4
x
10-­‐6
QuesBon
P
7
2
x
10-­‐11
8
1
x
10-­‐9
9
4
x
10-­‐10
10
1
x
10-­‐10
Rejected*
0.03
Rejected*
0.20
*These
two
ques'ons
from
ini'al
EBQ
were
deleted.

36. Example
Case:
Type
A
• J.M.
49
year-­‐old
W
male
• Wt.
275.2
pounds
• Ht.
69”
• W.C.
46.5”
• Fat
%
41
• BMI
41
• VLCKD
+
Phentermine
37.5
mg/day
• Rx
Dura'on:
5
months,
-­‐56.8
lbs.,
-­‐20.6%

37. 0
-­‐8.8
-­‐10.8
-­‐14.3
-­‐17.2
-­‐20.6
0
-­‐8.2
-­‐12.2
-­‐15.1
-­‐17.2
-­‐18.1
-­‐25
-­‐20
-­‐15
-­‐10
-­‐5
0
0
1
2
3
4
5
%
Weight
loss
Months
J.M.
vs.
Avg.
%
Weight
Loss
through
5
months
JM
AVG
EBQ
61
EBQ
22
EBQ
19
EBQ
12
EBQ
17
EBQ
12
1
Week=
22
Average
Pa'ent
Wt.
Loss
by
month
from:
Hendricks,
Obesity
2011;
19:2351-­‐2360.

38. Example
Case:
Type
B
• M.L.
46
year-­‐old
H
female
• Wt.
157.6
pounds
• Ht.
61”
• W.C.
36”
• Fat
%
52
• BMI
29,
(Prior
high
31)
• VLCKD
+
Phentermine
37.5
mg/day
• Rx
Dura'on:
3
months,
-­‐30
lbs.,
-­‐17.4%

39. 0
-­‐8.0
-­‐12.2
-­‐15.1
0
-­‐3.5
-­‐12.9
-­‐17.4
-­‐25
-­‐20
-­‐15
-­‐10
-­‐5
0
0
1
2
3
%
Weight
loss
Months
M.L.
vs.
AVG
%
Weight
loss
through
3
Months
AVG
Pt
ML
EBQ
60
1
Week=
32
EBQ
43
EBQ
42
EBQ
50
Average
Pa'ent
Wt.
Loss
by
month
from:
Hendricks,
Obesity
2011;
19:2351-­‐2360.

40. Example
Case:
Type
C
• S.A.
70
year-­‐old
W
female
• Wt.
172.2
pounds
• Ht.
60”
• W.C.
42”
• Fat
%
52
• BMI
34
• VLCKD
+
Phentermine
37.5
mg/day
• Rx
Dura'on:
8
months,
-­‐22.3
lbs.,
-­‐13.5%

41. 0
-­‐3.4
-­‐5.3
-­‐7.5
-­‐9.6
-­‐9.9
-­‐12.0
-­‐12.3
-­‐13.5
0
-­‐8.0
-­‐12.2
-­‐15.1
-­‐17.2
-­‐18.1
-­‐18.8
-­‐18.8
-­‐18.9
-­‐25
-­‐20
-­‐15
-­‐10
-­‐5
0
0
1
2
3
4
5
6
7
8
%
Weight
loss
Months
S.A.
vs.
Avg.
%
Weight
Loss
through
8
Months
EBQ:
31
35
33
-­‐22.3
#
Average
Pa'ent
Wt.
Loss
by
month
from:
Hendricks,
Obesity
2011;
19:2351-­‐2360.

42. EBQ
Comments
• Scores
dependent
on
pa'ent’s
observa'ons.
• Some
pa'ents
poor
at
self-­‐observa'on.
• Inappropriate
in
our
hands
for
5%
of
pa'ents.
• Some
untreated
pa'ents
present
with
low
scores.
• Low
EBQ
score
may
occur
in
untreated
pa'ents
who
have
dieted
previously.
• Pa'ents
treated
with
diet
alone
some'mes
have
high
Δ1-­‐2

43. Clinical
usefulness
of
EBQ
• Useful
as
ancillary
metric
of
treatment
effec'veness.
• Scores
<
50
suggest
treatment
is
effec've
• Large
EBQ
Δ1-­‐2
suggests
good
Rx
effect.
• Scores
>
50
suggest
no
or
ineffec've
treatment
• Increases
pa'ent
awareness
of
Rx
effects.
• Could
improve
long-­‐term
Rx
compliance?

44. EBQ
Summary
&
Conclusions
• Discriminates
treated
from
untreated
pts.
• Good
test-­‐retest
reliability.
• Low
scores
persist
for
years
in
con'nuously
treated
pts.
with
good
response.
• High
or
increasing
scores
are
one
indica'on
treatment
altera'on
should
be
considered.
• Tes'ng
and
scoring
can
be
accomplished
usually
in
<
3
minutes.

45. Comments
• These
data
suggest
the
EBQ
deserves
further
inves'ga'on.
• The
EBQ
has
not
yet
been
validated.
• Prac''oners
are
encouraged
to
use
the
EBQ,
and
to
assist
with
further
inves'ga'ons.
• The
Ea'ng
Behavior
Ques'onnaire©
is
available
from
the
Obesity
Treatment
Founda'on.

46. Some
Ques'ons
for
Future
Research
• Does
high
EBQ
Δ1-­‐2
indicate
a
good
6
month
weight
loss
(i.e.
is
the
pa'ent
a
responder)?
• What
are
EBQ
Δ1-­‐2
values
for
other
drugs?
• Why
do
some
pa'ents
have
low
ini'al
scores?
• Can
the
EBQ
be
used
to
jus'fy
drug
or
dose
changes?
• Does
high
EBQ
Δ1-­‐2
occur
with
all
treatments?

47. This
EBQ
study
was
funded
by
the
ASBP.
With
Thanks
to
Study
Collaborators:
Frank
L.
Greenway,
MD
Professor
and
Director
of
Outpa'ent
Clinic
Pennington
Biomedical
Research
Center
Louisiana
State
University
Baton
Rouge,
LA
Stacy
L.
Schmidt,
PhD
Director,
Obesity
Treatment
Founda'on
Yelena
Istra'y
Student,
Pre-­‐medicine
Sierra
College,
Rocklin,
CA
Margaret
(Mia)
J.
Hendricks
Student,
Psychology
Pepperdine
University
Malibu,
CA