Transcript Psychosocial interventions

Development of Psychosocial Interventions
Steven A. Safren, Ph.D., ABPP
Professor Of Psychology, Department of Psychiatry, Harvard
Medical School
Director, Behavioral Medicine Service, Department of Psychiatry,
Massachusetts General Hospital
Stages of Drug Development and Review (FDA)
1. Preclinical testing (laboratory animals and what
proposed to do for human testing)
2. Phase 1 Trials - Toxicity: Healthy volunteers
looking at side effects, metabolism (n ≈ 20-80)
3. Phase 2 Trials – Preliminary efficacy: Those with
the disease or condition, compare study drug to
placebo or different drug (n ≈ 36-300)
4. Phase 3 Trials: Large Scale Effectiveness:
Different populations, different doses, in
combination with other drugs. Agreement with
FDA (n ≈ 200-3,000)
“Steps” for Psychosocial Intervention Development
(not universally agreed upon like FDA)
1. Conceptual model: Articulate, conceptually, based on
behavioral sciences, why your intervention is appropriate
2. Develop intervention based on that model
3. Open pilot trial (Iterative is good): Refine intervention,
Feasibility and acceptability
4. Pilot Randomized Trial: Further feasibility and
acceptability and initial estimation of a clinical effect
5. Efficacy Study (or studies): Does it work under ideal
research conditions
6. Effectiveness Research: Does it work in real world
settings
7. Dissemination and Implementation Research: Studying
how to bring this research into practice
3
Why are the stages of psychosocial treatment
development not universally agreed upon like FDA?
What is needed next depends on
1. Population
2. Setting
3. Type of intervention and whether it has an existing
evidence base
4. Sometimes there is a need for rapid dissemination
5. Sometimes options for a control group may be limited
6. Behavioral scientists do not always agree – and there is
not always a hard, measurable, definitive outcome (i.e.
blood or scan test for a psychosocial outcome)
4
1. Conceptual Model
1. An overarching model behind your intervention
2. Given existing data (yours and others), and the
population under study why do you think this
intervention will work? The intervention addresses
the components of the model
3. Your measures / assessments assess the aspects of
the model
4. In your grant, it ties together the aims, significance,
innovation, preliminary studies, and the research
strategy
5
1. Developing a conceptual Model for CBT-AD:
Interventions for adherence in HIV that did not
address mental health comorbidity
• Minimal interventions – MGH/Fenway work
• “Life-Steps” – single session adherence intervention;
significant effects but comparison group “caught up” over
time (Safren et al., 2001)
• Pager study – significant but modest effects (Safren et al., 2003)
• Meta analyses of adherence interventions: significant but
modest effects
• Simoni (2006): 19 RCTs
• Amico (2006): 25 studies
Safren et al., 2001, Behaviour Research and Therapy; Safren et al., 2003;
AIDS Care; Amico et al., 2006. JAIDS; Simoni et al., 2006. JAIDS
1. Developing a conceptual Model for CBT-AD:
The prevalence of depression is high in many chronic
illnesses
General
Population
Individuals with
type 2 diabetes
HIV-Infected
Individuals
7% 1
10-15% 2
36% 3
1.
2.
3.
Kessler et al., 2005; Archives of General
Psychiatry
Geffken et al., 1998; Psychiatric Clinics of North
America
Krishnan et al., 2002; Biological Psychiatry, Bing
et al., 2005
1. Developing a Conceptual Model for CBT-AD
Depression is Associated with Non-Adherence
Dimatteo 2000 – various medical conditions:
• Patients with depression 3 X greater odds of being
nonadherent (95% CI = 1.96-4.89) (12 studies)
• Studies did not include HIV or Diabetes
Diabetes (Gonzalez et al 2008)
• 47 independent samples
• Depression significantly associated with non-adherence
(p<.0001; r = -0.21, 95% CI 0.17– 0.25)
HIV (Gonzalez et al 2011)
• 95 independent samples
• Depression significantly associated with non-adherence
(p<.0001; r = 0.19; 95% CI = 0.14 to 0.25)
1. Conceptual Model for CBT-AD
Depression may interfere with social psychology
models of health behavior change
Depression / Anxiety,
substance use dx
Information
Behavioral Skills
Motivation
IMB model from Fisher et al, Health Psychology, 2006.
Health
Behavior
2. Develop Intervention Based on Conceptual Model
1. Once you know the factors that you think are
causing the problem, look at evidenced-based
strategies that address these factors
2. If evidenced based strategies do not exist, are
there any basic behavioral science data that would
inform what your intervention might look like?
10
2. Develop Intervention Based on Conceptual Model:
Integrating CBT for Depression with CBT for
Adherence:
INITIAL OUTLINE CBT-AD
 Each CBT module for depression integrates adherence counseling
1,2. Adherence training and Overview of CBT for
Depression
2 sessions
3. Psychoeducation and Motivation
1 session
4. Behavioral Activation
1 session
5-10 Cognitive Restructuring
5 sessions
11, 12 Problem Solving
2 sessions
13, 14. Relaxation Training
2 sessions
3. Open pilot trial
1. General goal is for feasibility and acceptability
before a pilot randomized trial
2. See Rounsaville et al (2001) for very specific
guidelines for “Stage 1a” and “Stage 1b” behavioral
intervention research , such as:
•
Articulation of manual
•
Procedures for training and supervision of
interventionists
See Rounsaville et al (2001), Clin Psychol Sci Prac, 8:133-142.
12
3. Open Pilot / Refine Intervention
Done with outline of intervention manual N-=5)
DSM-IV SCID; Major Depression primary,
however most had comoribidity
(anxiety disorders, history of
substance abuse or dependence)
MEMS Adherence Outcomes
100
80
Men who have sex with men (MSM)
60
1. African-American male in his mid
40’s who was on disability
2. White male in his early thirties who
worked full-time
3. White male in his late forties who
was on disability
4. African American man in his late
forties who worked part time and
was on disability
5. White male in his early fifties, with a
relatively higher socio-economic
status who worked full time.
40
20
1 2 3 4 5 6 7 8 9 10 11 12
Treatment session
Depression (BDI scores) across treatment sessions
40
35
30
25
20
15
10
5
0
1
Safren et al., 2004. Cognitive and Behavioral Practice,
11, 415-423
2
3
4
5
6
7
8
9 10 11 12
Treatment session
4. Pilot Randomized Trial
1. Additional feasibility and acceptability data for
• Intervention
• Control Condition
• All study procedures
2. Estimation of an effect (to assist with power
analysis)
3. See Rounsaville et al (2001) for very specific
guidelines for “Stage 1c” behavioral intervention
research , such as:
• Number of participants (rule of thumb = 15- 30
per cell)
See Rounsaville et al (2001), Clin Psychol Sci Prac, 8:133-142.
14
4. Pilot Randomized Trial of CBT-AD
Done with more detailed / fleshed out version of
manual outline
Example Changes
• Flexible with order of
modules
• Flexible with number of
sessions
• Reduced target treatment
length to 12 sessions
• Changed language where
possible (e.g. “cognitive
restructuring” = “adaptive
thinking”
Adherence Training / Life-Steps
1 session
Psychoeducation and Motivation
1 session
Behavioral Activation
2 session
Adaptive thinking (cognitive
restructuring)
4 sessions
Problem Solving
2 sessions
Relaxation Training
1 session
Maintenance & Relapse
Prevention
1 session
4. Pilot Randomized Trial of CBT-AD
• Conducted at Fenway
• 2 Arm, cross-over design comparing
CBT-AD to a “Life-Steps”, our single
session of adherence counseling
• 3-month: CBT-AD resulted in
• improved adherence (MEMS=pill
cap)
• IA assessed depression at three
months
• Gains maintained at 6 and 12
months.
• Those who “crossed over” caught up
after completing the full
intervention
• Plasma Viral load: longitudinal
improvements comparing followups to baseline
• Retained 43 of 45 participants
randomized for primary analysis
100
MEMS Adherence outcomes
75
50
25
0
BASELINE
CBT
ETAU
F(1,42) = 21.94, p< .0001, Effect size (Cohen d) = 1.0
25
HAM-D outcomes
20
15
10
5
0
BASELINE
NIMH R21 MH066660 (2003-2007)
Safren et al., 2009; Health Psychology
T2
T2
F(1,42) = 6.32, p < .02, Cohen d = .82
5. Efficacy Studies
1. The “gold standard” – randomized controlled trial
2. Follow CONSORT guidelines
3. All procedures and measures valid, reliable, pilot
tested
4. Power analysis and analysis plan is tight
5. Registered (e.g. clinicaltrials.gov)
6. Follow all “ingredients” for psychosocial trials, such
as
• Blinded assessments when possible and when
done via raters (i.e. not the interventionist)
• Objective measures when possible
• Rating of adherence/competence to the
intervention and contamination in control
group
17
5. Efficacy Studies:
CBT-AD in HIV+ IDU (hybrid efficacy/effectiveness?)
• 2 arm study (ETAU or CBT-AD)
• Participants (N=89) recruited from methadone clinics and
community in Massachusetts and Rhode Island
• History (or current) IDU but in SU treatment
• Acute Outcomes: CBT superior to
ETAU in adherence and depression
• Follow up Outcomes: Depression
gains maintained, adherence gains
not maintained
• No Viral load differences
• CBT group had improved CD4
compared to ETAU
R01 DA018603
Safren et al., 2012 – Journal of Consulting and Clinical Psychology
5. Efficacy Studies:
CBT-AD Extension to Type 2 Diabetes
• 2 arm study (ETAU or CBT-AD)
• Participants (N=87) with Type 2 diabetes, poor glucose control,
stable medications, and diagnosis of depression
• Acute Outcomes: CBT superior to
ETAU in adherence to glucose
monitoring and medications,
depression, and glucose control
(HbA1c)
• Follow up Outcomes: CBT-AD
superior in adherence, and HbA1c.
Depression gains maintained but no
longer differentiate from ETAU
R01 DA018603
Safren et al., 2014 – Diabetes Care
5. Efficacy Studies: Current work - Project
“TRIAD”
NIMH funded efficacy trial (PI: Safren)
R01MH084757-05
3 arm study (2:2:1 randomization) to allow Pre-consent
Screen
for time matched control and costBaseline
effectiveness analysis
Assessment
• Life-Steps plus provider letter
• CBT-AD
Life-Steps
(Weekly Visit One)
• Information/supportive
psychotherapy
CBT-AD
Treatment as Usual
5 Non-Treatment Visits
• 3 site study (MGH, Brown, Fenway)
• Wide inclusion criteria
11 Treatment Visits
Screen out/
Drop out
ISP-AD
11 Treatment Visits
4 month
8 month
12 month
NIMH R-01 MH084757
6. Effectiveness Research
Primary question: Does the intervention work in real world
settings
• Usually done in real world settings
• Wide inclusion/exclusion criteria
• Usual / intended providers of the intervention
• More flexible application of intervention
• Other interventions and cross-over permitted
• Cost-effectiveness may be involved
• Usually uses usual care comparison group or compares
different efficacious interventions (comparative
effectiveness research )
Refs: Singal, Higgins, Waljee (2014), Clinincal and Translational Gastroenterology.
Carroll and Rounsaville, (2013), Psychiatric Services
6. Effectiveness Research
1. Nueva Dia Study: Simoni et al (2013), preliminary RCT in
HIV at U.S. Mexico-Border, Spanish adaptation with
Latino Graduate Students as therapists
2. Ziphamandla: Nurse interventionists in HIV-positive
patients in S. Africa
Increased need for effectiveness studies, but less consensus
in field re: standards for study design = potentially harder
to get through peer review funding
7. Dissemination and Implementation Science
• This is the study of how to best bring discoveries into
clinical practice
• May require adapting the intervention?
• Usually does not involve patient level outcomes, but
rather outcomes are:
• Whether the intervention was used or not
• What factors are related to uptake of the
intervention in systems / settings
Suggested readings / resources
1. Overview of stage model:
• Rounsaville, B.J., Carroll, K.M., & Onken, L.S. (2001). A stage
model of behavioral therapies research: Getting started and
moving on from stage I. Clinical Psychology: Science and
Practice, 8, 133-142
2. Hybrid Efficacy/Effectiveness
• Carroll, K.M., & Rounsaville, B.J. (2003). Bridging the gap: A
hybrid model to link efficacy and effectiveness research in
substance abuse treatment. Psychiatric services, 54, 333339.
3. Dissemination and implementation science:
• Brownson, Colditz, & Proctor. Dissemination and
implementation research in health. NY: Oxford University
Press.
4. NIH view on pilot intervention and services grants - The NIH PAR
for R34 mechanism. http://grants.nih.gov/grants/guide/pafiles/PAR-12-279.html