When & How do Patients Change? Wolfgang Lutz Outcome Presentation

1
Calgary, December2014
Why, whenand howdo patientschange?
Identifyingandpredictingprogressand outcomein psychotherapy
Wolfgang Lutz
(University ofTrier)
http://www.kpplutz.uni-trier.de

2
•Therapistsestimatethemselfon the80th percentil.
•Nobody sawhimself/herselfbelowthe50th percentile.
0%50%100%
Walfish, McAlister, O‘DonnellandLambert (2012)
Hannan, Lambert, Harmon, Nielsen, Smart, Shimokawa, et al. (2005)
An average engineer rated his performance to be at the 78th percentile (Meyer, 1980)
Better-than-average Effect (BTA)

3
•PsychotherapyResearch could/shouldbecomepartofclinicalpracticein ordertosupportthedeliveryofpsychotherapy. (Treatment tracking)
•Modern toolsdevelopedin thecontextofeMentalHealth/Feedback researchcanhelptorealizethis.
•But thedecisionabouttheclinicalvalidityoftheso deliveredadditional informationhastostayin thehandsofa scientificallywell-trainedtherapist.
•Replication: Severaldatasetscouldbeusedforvalidation–wehavetodeal withlarge patientandsettingvariation.
Statement

44
When, howandwhydo peoplechangethroughpsychologicalinterventions--Human Change Through PsychotherapyProgram(HCTPP)
•Research topicswithintheClinical Psychology and PsychotherapySectionat theUniversity of Trier /European Center forPsychotherapyand PsychotherapyResearch
•Fromthemacro-tothemicro-level ofchangein psychotherapy
–1. macro-level:
–Patient orclient-focusedpsychotherapyresearch/predictionofchange/feedback
–2. meso-level:
–Discontinoustreatmentcoursesandunderlyingprocesses/factors
–3. micro-level:
–Therapeuticmicro-strategies
•Outpatient centerandclinicaltrainingprogram, PhDprogram„PsychotherapyResearch“ andresearchorientedfocusin themasterprogram„Clinical Psychology“

8
Feedback Portal (signalclient)
8

99
Outline
1.A 3-level researchprojecton individual patientchangeandpatient-focusedresearch
2.Therapisteffects, Disaggregation
3.Macro-level: New projectsandapplications, NN
4.Meso-level: Shapes ofchangeandsuddengainsandlosses
5.Discussion
What does it mean?
How to improve it?
How to extend it?
How to deal with different shapes and discontinuity?
What makes a difference?

1100
Evaluation of the
TK-Project
• Pilot project „Quality monitoring in
outpatient psychotherapy“
of the Techniker Krankenkasse
• Three regions of Germany: Hessen,
Westfalen-Lippe and Südbaden.
> Duration 2005-2010 (IG: modified
review system, Feedback, Long-term,
structured diagnostic IDCL)
> Evaluation by WGs Trier (Lutz)
and Mannheim (Wittmann)
> Full report see website of the TK
T-values
middle
strong
mild/no middle strong
Measure
Session
Date
Change*
mild/no
The GSI (Global Severity Index) measures the general symptomatic status. It merges the itensity of perceived
burden in all of the 53 symptoms.
General
Symptomatic
Status (BSI)
Value 1.38
T-Value 97
Interpersonal
problems (IIP-D)
Value 1.84
T-Value 61
Disorder Specific
(BDI)
Value 1.38
T-Value 97
T-values
How stressful does your patient experience the problem?
How does your patient asses his/her global psychological distress?

11
Descriptionofthesample
N
age
M(SD)
female
N (%)
male
N (%)
IG
1060
40,2 (11,3)
716 (67,5%)
344 (32,5%)
CG 1
614
41,5 (11,0)
432 (70,4%)
182 (29,6%)
CG 2
33210
41,2(11,6)
23592 (71,0%)
9618 (29,0%)
CG 3
27563
40,6 (11,2)
19852 (72,0%)
7711 (28,0%)
NP
2778
42,4 (11,6)
1937 (69,7%)
841 (30,3%)
11
IG
CG
NP
CG2
CG3
CBT
71667,5%
41367,3%
159957,6%
1635049,2%
1116640,5%
PD
32931,0%
19732,1%
114541,2%
1576347,5%
1458652,9%
PA
151,4%
40,7%
341,2%
10973,3%
18116,6%
Total
1060100,0%
614100,0%
2778100,0%
33210100,0%
27563100,0%

12
Whatdo therapistsdo withfeedback? -dependingon feedbacktype
149; 30%
352; 70%
Due to Feedback...
...wurden keinerlei Veränderungen in derBehandlung vorgenommen
...wurde mindestens eine der genanntenVeränderungen vorgenommen
.

13
Patients´ evaluations of outcome monitoring
13
Question n
Completely
right
Rather
right
neither
/ nor
Rather
wrong
Completely
wrong
I like the idea of having a project
monitoring the quality of
outpatient psychotherapy.
597
374
(62,6%)
177
(29,6%)
41
(6,9%)
3
(0,5%)
2
(0,3%)
I find it important to monitor the
results of psychotherapeutic
treatments.
597
399
(66,8%)
156
(26,1%)
30
(5,0%)
8
(1,3%)
4
(0,7%)
The time I needed to answer the
questions was appropriate.
597
389
(65,2%)
181
(30,3%)
14
(2,3%)
12
(2,0%)
1
(0,2%)
92.2% 0.8%
92.9% 2.0%
95.5% 2.2%
Lutz, Böhnke, Köck, Rubel, 2011,2012,2013

14
Effectsize
Pre-Post
Duration oftreatment
(sorted by length of treatment by therapist in decendingorder)
BSI:
Symptoms
atthebeginning
BSI
Symptoms
after therapy
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Therapist
Effectson
Treatment
Length

15
Level 1:Symptomsij= β0i+ β1i*Symptoms_preij+ eij
Level 2:β0i= γ00+r0i´;β1i= γ10+ r1i
InterceptResidual (95% CI)
Therapistrankedfrommosttoleast effective
Therapisteffecton outcome(correctedafter initrialimpairment); 9.8%, d=.66
N= 751 Patienten
N= 177 Therapeuten
Multilevel-Model
Saxon& Barkham , 2012, JCCP.;
Baldwin & Imel, 2013

16
Study Sample
Npatients= 349
Ntherapists= 44
CompleterandStudy Sample
CompleterSample
Npatients= 751
Ntherapists= 177
Study Sample requirement
Diagnoses
Early alliance
Early feedback
5 patients per therapist

17
TherapisteffectTotal VariancePatient variance
Level 2
Level 1
Therapisteffecton treatmentoutcomeStudy Sample: N = 349
6.9%
6.9%
variance
unexplained
variance
unexplained
Level 1 predictors:
•Initial impairment(BSIpre)
•Early alliance(HAQpre)
•Numberofdiagnoses
•Early feedback
34.47%
34.47%

18
Therapist and patient attitude towards and usage of feedback
Attitude towardsfeedback
Howsatisfiedareyouwiththe QM project?
Amountof modificationsdue to Feedback

19
Therapisteffectson treatmentoutcomeStudy Sample: N = 349
Explainedvarianceon level1
•Initial impairment(BSIpre)
•Early alliance(HAQpre)
•Numberofdiagnoses
•Early feedback
Model A:
34.47%
Model B:
37.88%
Model C:
35.95%
Therapistattitudetowardsfeedback
Patient attitude
towardsfeedback

20
Therapist effects on Outcome, Treatment Length, Drop-out (TK-Study) in ES
No correlation between therapist
Effects in outcome and length
larger ES =
betteroutcome, less
drop-out, shortertreatments

21
21
46; 28%
119; 72%
Outpatient clinic
universityTrier
*
What do therapists do with feedback?
…discussedwiththepatienthis/her answersin thequestionnaire.
…triedtoassistthepatientsressources.
…triedtoadjustmytherapeuticinterventions.
…discussedwiththepatienthis/her interpersonal problems.
…preparedtheend ofthetherapy.
…triedtoenhancethepatientsmotivationfortherapy.
…variedtheintervallsbetweensessions.
…triedtoenhancethetherapeuticalliance.
…consultedadditional sourcesofhelp
(e. g. supervision, intervision, literature, furthereducationetc.).
…triednewhomeworkwiththepatient.
…other.
Outpatient clinicTrier
TK-project
Ifmodificationsweremade: Duo tothefeedback, I…

22
Therapisteffectson Outcome, Treatment Length, Drop-out (Outpatient Center Trier) in ES
22
larger ES =
betteroutcome, less
drop-out, shortertreatments

23
Outlook: Aggregated DatasetN= 48,648 (patients); N=1800 (therapisits)
23
Dataset 1
Outpatient clinic Southwest Germany
Npatients= 668
Ntherapists= 97
BSI
Dataset 2
German TK project
Npatients= 636
Ntherapists = 120
BSI
Dataset 3
Outpatient clinic Midwest Germany
Npatients= 752
Ntherapists = 71
BSI
Dataset 8
IAPT data set
UK
Npatients= 5,639
Ntherapists= 119
PHQ
Dataset 7
CORE data set
UK
Npatients= 25,842
Ntherapists= 789
CORE
Dataset 4
CelestHealth data set USA
Npatients= 11.356
Ntherapists= 401
BHM
Dataset 5
Compass data set
USA
Npatients= 1,194
Ntherapists= 60
MHI
Dataset 6
University Counseling Center
USA
Npatients= 2,561
Ntherapists = 143
OQ
Total dataset
Npatients= 48,648
Ntherapists= 1,800
ManythankstoMichael Barkham, Jaime Delgadillo, Michael Lambert, Dietmar Schulte, Ken Howard, & Mark Kopta

2424
Most effective
Ø
Least effective
16.8% (302)
66.9%(1204)
16.3% (294)
0%50%100%
AggegatedDatasetN= 48,648 (patients); N=1800 (therapists)

25
Parameterschätzer
Variable
DF
Parameter- schätzer
Standard- fehler
t-Wert
Pr>|t|
Intercept
1
29.82503
2.34412
12.72
<.0001
patprother
1
-0.66114
0.11963
-5.53
<.0001
therprodat
1
-0.09526
0.01721
-5.53
<.0001
n
1
0.00203
0.00088245
2.30
0.0235
3
4
N of therapists
N patients per therapist
Bootstrap Analysis on differencesbetweenstudies:
About47% ofthein variationin CI fortherapisteffects
areexplainedbynumberoftherapists, patientsper therapistandnumberofpatients

26 26
Outline
2.TherapisteffectsandDisaggregation
5.Discussion
What does it mean?
How to improve it?
How to extend it?

27 27
ExpectedTreatmentResponse
Predictor Variables:
Current Well-Being, Current Symptoms, Current Life Functioning, Psychotherapy in the past, Duration of Problem,Treatment Expectations,Global Assessment of Functioning
Lutz, W., Martinovich, Z., & Howard, K.I. (1999).Journal of Consulting and Clinical Psychology, 67, 571-577.

28 28
http://www.slf.ch/lawineninfo/lawinenbulletin/regionale_lawinenbulletins/rbbob/rbbob_c_DE
1 gering
2 mässig
3 erheblich
4 gross
5 sehr gross
NearestNeighborsinAvalancheResearch

29 29
•Problem: Sample specificity
•New: Individual predictions based on their nearest neighbors
•Two homogeneous subsamples of the 30 nearest patients were selected for a CBT oriented treatment group and an integrative interpersonal and CBT oriented treatment group and Growth Curve Modeling was conducted on those two groups for each patient
DisaggregationandNearestNeighbors
Lutz, W., Saunders, S., Leon, S. C.et al. (2006). Psychological Assessment. Lutz, W., et al. (2005). Journal of Consulting and Clinical Psychology. 73, 904-913
N=619 (Inventory of Emotional Distress (EMI)
Site 1: N= 359 Outpatient Clinic at the University of Berne(Integrative Cognitive-Behavioral and Interpersonal Focus)
Site 2: N=260 Outpatient Clinic at the University of Bochum(Cognitive-Behavioral Focus)

3300
30
35
40
45
50
55
60
0 5 10 15 20 25 30 35 40
Observed Score
Expected CBT
Expected Integ. Therapy
T-Score
Session
0
A Patient with a Diagnoses of Anxiety & Depression
– Treated with Integrative Therapy

3311
A Patient with a Diagnoses of Anxiety & Depression
– Treated with CBT
30
35
40
45
50
55
60
0 5 10 15 20
beob. Werte
bewält. Vorgehen
kominiertes Vorgehen
T-Werte
EMI
Sitzung
0
Observed
CORE-SF Scores
Prediction Based
on Intake Scores
Core
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
Sessions
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Prediction Based
on Session 1-3 Scores
Wakefield Metropolitan
District (UK)
204 clients,session-by-session
with the CORE-SF (18 items)
Lutz, Leach, Barkham,Lucock, Stiles,
Evans, Noble, Iveson (2005). JCCP.
73,904-913

3322
Decision rules and Outcome (Feedback
between session 2-8 & Outcome between
session 17-28, N=389)
0
10
20
30
40
50
60
70
80
90
100
3 2 1 0 0 1 2 3
amount of negative feedback amount of positive feedback
% of no change/ failure
0
10
20
30
40
50
60
70
80
90
100
% of reliable improvement.
67.0 75.0 84.0 90.0 95.0 97.5 99.5
0
10
20
30
40
50
60
70
80
90
100
1 4 7 10 13 17 20 24 30 33 36 39
Sessions
LSQ-30
LSQ-30 Predicted Score 67%
75% 84% 90%
95% 97.5% 99.5%
Lutz, W., Lambert, M.J., et al..
(2006). The probability of
treatment success, failure and
duration.Clinical Psychology
& Psychotherapy, 13, 223-
232.

33 33
N´s(all pre= 5627;
all post=2838: all FU=711)
Bochum
Pre=1572
Post=1089
Bern
Pre=1181
Post=545
Trier: ES=.85
Pre=1175
Post=348,
FU(1/2-1y)=129
NIMH: ES=.1.56
Pre=122
Post=106
FU(1y)=98
TK:
Pre=1577
Post=760
FU(1y)=485
Many thanks to Dietmar Schulte, Klaus Grawe, Irene Elkin for providing their data
M(nor)+SD=.54
M(clin)=1.14
BSIComparisondifferent sitesNaturalisticandRCT datasets-all

34
DensityPlot ofTreatment Effects
34
~ 60%

35
NIMH Pre--BSI Scores(Trier Outpatient Center N=335, AM=1.2)
NIMH: AM=1.35
Trier: AM=1.21
Propensity Score Matching (PSM)

36
Propensity Score Matching (PSM)
NIMH andPre-- BSI ScoresPSM (Trier Outpatient Center N=228, AM=1.4)

37
ComparableES andDistributionsPre-andFU
NIMH andPre--BSI ScoresPSM (Trier Outpatient Center N=228, AM=1.4)
NIMH (CBT) andTrier PSM Scores
at1-year FU
NIMH:AM=.57,ES=1.56
Trier: AM=.59,
ES=1.68, ES(FU)=1.72
But different treatmentlength!

38 38
Outline
2.TherapisteffectsandDisaggregation
5.Discussion
What does it mean?
How to improve it?
How to extend it?

3399
Patterns of early change in efficacy
and effectiveness studies
Discontinous
0
4
8
12
16
20
24
28
32
36
1 2 3 4 5 6
Sitzung
CORE-SF
Continous
0
4
8
12
16
20
24
28
32
36
1 2 3 4 5 6
Sitzung
CORE-SF
Stulz, N., Lutz, W., Leach, C., Lucock, M., Barkham, M. (2007).
Journal of Consulting and Clinical Psychology, 75, 864-874
Lutz, W., Stulz, N., & Köck, K. (2009). Journal of Affective Disorders.
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
1 2 3 4 5 6
Sitzung
CORE-SF
Rapid early response (n = 24)
Low impairment (n = 26)
High impairment (n = 48)
Continous (n = 56)
Discontinous (n = 38)

40 40
TreatmentOutcomeandTreatment Length
Outcome
(reliable change)
Treatment Length (sessions)
Subgroups
-
0
+
7-13
14-30
>30
Rapid early response
4%
0%
96%
47%
34%
19%
Lowimpairment
4%
96%
0%
31%
57%
12%
Highimpairment
20%
58%
22%
2%
28%
70%
Continous
0%
81%
19%
42%
27%
31%
Discontinous
13%
43%
44%
43%
33%
24%

41 41
Lutz, W., Stulz, N., & Köck, K. (2009). Journal of Affective Disorders.
Lutz, W., Hoffmann, S. et al. (2014). JCCP.
What do we know about change patterns in psychotherapy?
Depression: 61.1% „Early responder“
Panic disorder: 20.2% „Early Responder“
Disorderspecificcontrolledsettings

42
Treatmentoutcomeand lengthof the different earlychangegroups
Class 1: Early deterioration
Class 2: Medium symptoms –slowchange
Class 3: High symptoms –nochange
Class 4: Early response
Early respondershowthe highestpre-post effectsizesand the highestprobabilityto completethe treatment. Nonresponder(class3) and deteriorater(class1) showhigh probabilitiesfor drop-out.
Lutz, W., Hoffmann, S. et al. (2014). JCCP.

43
Whatdo weknowaboutearlyresponse?
43
-ER seemstobeconsistentvia different settings, diagnosis, treatmentsandinstruments
-The ER grouphashigh treatmenteffects.
-Rates seemhigherin RCT`sthanin naturalisticsamples.
-in naturalisticstudiesthosealso withshortertreatments/ in RCT`sthosewhichfinish themanual.
-Differential ratesrelatedtodiagnosis(heterogeneous)

44
44
0
10
20
30
0 5 10 15 20
Session
BDI
Sudden Gain
Lutz, W. & Tschitsaz, A. (2007).
Tschitsaz, A. & Lutz, W. (2009).
Lutz, et al., (2013)
About 40% of patients experience a gain, which makes
about 51% of overall change
Change rate witht sudden gain: 79%, without: 41%
Follow-up scores (at 6 or 18 Mt.) are significantly better
(Tang & DeRubeis, 1999; 2005)
Sudden gains occur in CBT & supportive therapy and under
routine clinic conditions (Hardy, 2005; Stiles et al., 2004).
Sudden losses have been rarley investigated.
Frequency of gains and losses by sessions
(N=1500 outpatients, Trier, Bern, Bochum)

45
SuddengainsandlossesAnalysisofvideotapes, N=25 patients, 38 sessions–an example: Alliance ruptures
45
Sudden gainsandlossesdefinedaccordingtoHSCL-11 scoresbeforeeachsession.
Alliance rupturesratedaccordingtothe3RS codingsystembyEubanks-Carter et al. (2009)
withTorsten Ehrlich

46
Percentageofsessionswithalliancerupturesamonggain/loss/neutral sessions
73,91
84,21
55,00
8,70
36,84
25,00
8,70
36,84
12,50
0,00
10,00
20,00
30,00
40,00
50,00
60,00
70,00
80,00
90,00
100,00
gain session
loss session
neutral session
withdrawal
confrontation
w & c
χ²(2)=4.777;
p = .092
χ²(2)=5.654;
p = .059
χ²(2)=6.976;
p < .05
withTorsten Ehrlich

47 47
•Therapist effects exist in clinical practice for treatment outcome, treatment length and drop-out. The influence of the therapists seems to be more important as more impaired patients are.
•PsychotherapyResearch couldbecomepartofclinicalpracticeandsupportthedeliveryofpsychotherapy(Treatment tracking). Feedback on treatment progress on a patient level seems to improve therapy, especially for those with an early negative development. It also seems to have an impact on treatment length. More research and implementation, e.g. what do therapists do with feedback?
•Patients have a positive attitude to the evaluation of treat-mentresults/feedback. The active and self-organised handling of problems is supported. Therapist and patient attitude
towards feedback seem to influence results.
Discussion- Whatdoesthismeanforresearchandpractice?

48 48
•Early response: It seems there are patients, which are coming at the right time to the right place and those respond very fast to therapy. Responsible here is probably a specific patient X life event interaction. More research on inter-individual differences over the course of treatment would be helpful. The goal would be a typology of change patterns with influential process factors (mechanisms of change) and a theoretical backup.
•Differential effects and differential patient progress: It seems a subgroup responds to specific treatment manuals another maybe to extended clinical programs. Extended clinical programs, including combined CBT with a broader focus on e.g. emotion regulation issues/ mindfulness/interpersonal issues –they might be better suited for patients with more problems around comorbidity and depression issues, but this needs further investigation.
Discussion- Whatdoesthismeanforresearchandpractice?

When & How do Patients Change? Wolfgang Lutz Outcome Presentation

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to When & How do Patients Change? Wolfgang Lutz Outcome Presentation (20)

Recently uploaded (20)

When & How do Patients Change? Wolfgang Lutz Outcome Presentation