Transcript Example of HSR project: Cardiac Counselling and Rehabilitation

Example of HSR project:
Cardiac Counselling and
Rehabilitation: RCT of Complex
Interventions
Marie Johnston
‘History’
• Initiation – 1991!
• Expertise + experience
• Grant application – funded by Chief
Scientist Office
• Differences ‘now’
–
–
–
–
Co-applicants
MRC Framework for Complex Interventions
Power calculations
Randomisation
Background
• MI: high frequency, disabling
• Effects on families
• Evidence that cardiac rehabilitation programmes
effective
• Questions
– Timing
• In patient vs outpatient?
– Duration
– Content
• Emotional outcome
• Recovery
• Risk reduction
– Involvement of partners
• Partner emotional outcome
• Effects on patient
A
B
Ewart et al
CARDIAC REHABILITATION
AND COUNSELLING TRIAL
• Patients within a few days of myocardial infarction
• Intervention using cognitive-behavioural
technologies
– increase information e.g. risk reduction
– enhance perceived control
– enhance coping with limitations and with emotions
• Randomly allocated to intervention (in-patient or
extended) or control
• Outcome: changed - thoughts, emotions, activities
Johnston et al., 1999
Research Questions
After a first MI, do patients (and their partners) who
receive an inpatient cardiac rehabilitation
programme demonstrate:
1. Greater benefit than those receiving normal
care?
2. Equal benefit to those receiving an extended
programme?
[benefit = knowledge, satisfaction with care, mood,
disability]
Design
• Patients following first MI and their partners
• Randomised to:
– Normal care
– Inpatient CR
– Extended CR up to 2 months following discharge
• Followed up
2 weeks
2, 6 and 12 months
• Blind assessment
Randomisation
• Simple randomisation not possible
• Randomisation of post CCU wards
• Avoided confounding with wards and retain
blind assessment by changing
randomisation at variable intervals
• Clearance periods
Cardiac Counselling and
Rehabilitation Programme Delivery
• [Normal care – no formal programme]
• CR groups
– within 3 days of admission
– Inpatient up to 5 sessions
[actual average 5.55, 3.69 hours]
– Extended – up to 8 additional sessions
[actual average 9.55, 8.43 hours]
• Nurse counsellor – control for individual by
having two
• Manual
• Non-judgemental counselling
Cardiac Counselling and
Rehabilitation Programme Content
•
•
•
•
•
•
•
•
Aimed to enhance perceptions of control
Information
Action plans
Advice
Coping skills training
Relaxation
Leaflets and videos
Individual tailoring
– menu
Menu
•
•
•
•
•
Explanation of heart attack
Emotional effects
Risk factors and their modification
Recovery period: resumption of activities
Investigations/treatment
Evidence based techniques for
changing behaviour
•
•
•
•
•
•
•
•
•
•
Goal/target
Monitoring
Contract
Planning
Contingencies
Grading task
Skill enhancement
Skill rehearsal
Prompts
Modelling
•
•
•
•
•
•
•
•
•
Stress management
Environmental changes
Social pressure/support
Persuasive communication
Information re behaviour
and outcome
Personalised message
Homework
Personal experiments
Experiential
Inclusion criteria
• All patients admitted to Ninewells CCU ie
all from region
• First MI (WHO criteria)
• <70 years
• Fluent in English
• Able to participate
• Informed consent (13 refused)
Participants
• 117 randomly allocated
• 10 withdrew
• 7 died
• Numbers in groups
– Control 33
– Inpatient 38
– Extended 29
• No significant differences between 3 groups on
demographic or clinical factors
Illustrative Baseline data
[only Misconceptions significant – used as covariate]
Extended
Inpatient
Control
Men/women
19/10
27/11
19/14
age
57
54
57
Norris
4.86
4.81
5.47
Risk Index
45
40
35
Length of stay
8.6
7.4
7.9
4.11
3.67
Knowledge
2.55
misconceptions
Outcomes – I:
no standardised measures
• Knowledge
– New questionnaire
– 19 statements
– Responses: true, false,
don’t know
– Scores
• Correct (α = 0.68)
• Misconceptions (α =
0.57)
• Uncertainty (α = 0.74)
• Satisfaction with care
– 1 item
– ‘how satisfied do you
feel generally about the
advice that you
received after
your/your partner’s
heart attack?’
– Rated 1 to 10
• 1 = not at all satisfied
• 10 = extremely satisfied
Cardiac Rehabilitation and
Counselling: Knowledge
patients
partners
Knowledge: Correct
Significant group by time interaction: I and E > C at discharge and at 2 months
Significant effects for Misconceptions and Uncertainty
Cardiac Rehabilitation and Counselling:
Satisfaction
Patients: significant main effect of group: significant interaction (E>I at 2mths)
Partners: significant main effect of groups
Outcomes II: standard measures
• MOOD:
Hospital Anxiety and Depression Scale
• DISABILITY/RESUMPTION OF
NORMAL ACTIVITES:
Functional Limitations Profile (UK
version of Sickness Impact Profile)
Cardiac Rehabilitation and Counselling:
Anxiety
Significant interaction: I and E lower than C at 2 and 6 months
Cardiac Rehabilitation and Counselling: Anxiety
Patients
Partners
CR anxiety both.jpg
Partners: significant interaction: I < C at discharge and 2 months;
E< C at 2,6,12 months; E< I at 2 and 6 months
Cardiac Rehabilitation and Counselling:
Depression
Significant interaction: I < C at 2mths; E < C at 2, 6 and 12 mths
Cardiac Rehabilitation and Counselling: Depression
patients
partners
Partners: significant interaction: I<C at 6mths; E<C at 2, 6 mths
Cardiac Rehabilitation and Counselling: Functional Limitations Profile
total
physical
psychological
Significant main effect of groups on all 3 measures: C>I, C>E
Discussion: Results
•
•
•
•
Results show benefits of CR
For both patients and partners
Some lasting to 12 months
Some extra benefit of extended programme
– especially in partners
• No differences between 2 counsellors
• Did not have power to examine changes in
risk factors
Discussion
•
•
•
•
Levels of anxiety in partners
Levels of satisfaction in partners
Results on anxiety similar to other studies
Differential effects on women and men
• Lack of CR programmes in UK
• Provided for highly selected patients
• This intervention is implementable
Secondary analyses
Gender effects
Gender and
Activity Limitations at follow-up
FLP Physic al by gender and intervention gr oup
30
control
20
10
Ge nder
male
0
f emale
Control
Inter vention g roup
Inp atient
Exte nded
Cardiac rehabilitation & counseling
Mean Anxiety
Anxiety over Time : MEN
12
10
8
6
Intervention group
Extended
4
Inpatient
2
Control
Recruitment
Disc harge
2 months
6 months
12 months
T ime
Anxiety in Men following MI with and without Intervention
Anxiety over Time: WOMEN
12
10
8
6
Inter vention group
Extended
4
Inpatient
2
Control
Recruitment
Disc harge
2 months
6 months
12 months
T ime
Anxiety in Women following MI with and without Intervention
Designing a Randomised Clinical Trial
(RCT) to test if stress management reduces
blood pressure in patients with hypertension
(1986-1990)
Why do it?
•High BP major risk factor for cardiovascular disease
•Unclear how mildly raised BP should be treated
•Some evidence that relaxation/stress management effective
but previous studies poorly controlled
•Unclear how well results generalised in previous studies
Main Design/measurement issues in this study
Control, stability of BP over time, & Generalisation
Control
•Placebo control group or non-specific intervention i.e. has all
the common components of the therapy but none of the
specific (active) ones.
•Exercise, flexibility training
Stability
•Length of pre-treatment baseline (habituation). Multiple BP
measures before start of treatment
Sample
7 Practices referred patients with 2/3 DBP 90-104
184 referred
3 month baseline (BP measured twice per day)
88 excluded (96 allocated to 2 treatments
32 BP too low
13 BP too high
6 too heavy, too high alcohol
consumption
7 other illness
30 withdrew
Stress Management v Exercise
Clinic DBP (Johnston)
100
DBP mm Hg
94
88
Stress m anagem ent
Exercise
82
76
70
Pre-treatm ent
Pos t-treatm ent
Tightly controlled trial of stress management (like Patel), in approx.
100 mild hypertensives. Flexibility exercises used as control group.
Long baseline (3 months), clinic ambulatory and stress testing of BP
Stress Man. v Exercise ambulatory DBP (Johnston)
100
DBP mm Hg
94
88
Stress management
Exercise
82
76
70
Pre-treatment
Pos t-treatment
No effect on 12 hour ambulatory BP
Generalisation
Is BP measured clinically adequate for evaluating
relaxation?
•Ambulatory BP : Yes
•Enduring effects of successful therapy on CV system
(Left Ventricular Mass (LVM) : Yes
•A clinical outcome: myocardial infarction (heart
attack), death : No
Common issues in designing a RCT
•Power
•Analysis: “Intention to Treat”
•Blind assessment
•Cluster randomisation
Useful reference if contemplating conducting a RCT
Whole issue of Epidemiologic Reviews, 2002, 24, 1.
Edited by PW Lavori & JL Kelsey & covering
•Design
•Management
•Analysis
•Sample size
•Ethics
•More area specific topics
Assignment
Rates of hospital induced infections are too high, possibly
because the staff do not wash their hands. NHS Scotland
proposes to introduce a new training package to improve staff
hygiene but wish to evaluate it before requiring its use across the
country.
Design an RCT to evaluate the effectiveness of the package.
1 page single spaced. For Feb 13th