Title: A Complex Intervention in General Practice: Chest Symptoms that Call for Action
1A Complex Intervention in General PracticeChest
Symptoms that Call for Action
25 year relative survival
All cancers - female
All cancers - male
Lung cancer - female
Lung cancer - male
3Advanced disease at diagnosis
Andersen et al. Br J Soc Psychol 19953433-52
4MRC Complex Interventions Framework
Long-term Implement
Definitive RCT
Exploratory Trial
Determine whetherothers can reliablyreplicate
yourintervention and results in
uncontrolledsettings over thelong term
Compare a fully definedintervention toan
appropriatealternative usinga protocol that
istheoreticallydefensible,reproducibleand
adequatelycontrolled, in a study with
appropriatestatistical power
Modelling
Describe theconstant and variablecomponents
ofa replicableintervention ANDa feasible
protocolfor comparing theintervention to
anappropriatealternative
Theory
Identify the components of the intervention, and
the underlying mechanisms by which they will
influence outcomes to provide evidencethat you
can predicthow they relate toand interact
witheach other
Explore relevant theory to ensure best choice of
intervention and hypothesis and topredict
major confounders and strategic designissues
Pre-clinical
Phase I
Phase II
Phase III
Phase IV
Continuum of increasing evidence
Campbell M et al. BMJ 2000321694-696
5Designing trials of complex interventions
Proposed evaluation of a complex intervention
Define and understand the problem and its context
Develop and understand the intervention
Develop and optimise the evaluation
Promising and warrants definitive evaluation
Unlikely to be cost effective
Implement without trial
Campbell, Murray, Darbyshire, Emery, Farmer,
Griffiths, Guthrie, Lester, Wilson, Kinmonth BMJ
2007 334 455-459
6Understand the context
- The NHS
- Gatekeeping role in primary care
- Limited hospital resources
- Therapeutic nihilism
- Specialists not interested
- Older males in deprived areas who smoke
- Patients not interested
- General practice nGMS contract
- GPs not interested
7The underlying problem
Lung cancer almost always advanced. Median time
to presentation 99 days. (gttumour doubling time)
8Smoking, age and lung cancer risk
9The underlying problem
Lung cancer almost always advanced. Median time
to presentation 99 days. (gttumour doubling time)
90 in smokers 92 55 years or over
Psychological models (Andersen et al,
1995) Social theory (Zola, 1973, Pescolido
Levy, 2002) Quantitative qualitative evidence
10Motivation to understand symptoms
Unexpectedness Salience Personal relevance
Perceived Consequences
Delay
Symptoms/signs
Uncertainty
List of possible explanations
Logical comparison with illness prototypes
Optimistic bias
Social interaction
11Time from earliest symptom to consultation
- Patients factors consult later (plt0.05)
- Living alone
- Pack years smoking
- History of COPD
- Patients factors consult sooner (plt0.05)
- Cough
- Haemoptysis
- Loss of appetite
- Shortness of breath
- History of chest infection
- History of renal failure
-
12The underlying problem
Lung cancer almost always advanced. Median time
to presentation 99 days. (gttumour doubling time)
90 in smokers 92 55 years or over
Psychological models (Andersen et al,
1995) Social theory (Zola, 1973, Pescolido
Levy, 2002) Quantitative qualitative evidence
Possibly at several stages
75 of people could present earlier and still
meet guidelines for referral
13Develop optimal intervention
Fewer waiting with haemoptysis or waiting more
than 3 weeks with other symptoms
1. Key processes and outcome
Mapped onto model of underlying problem
2. Mechanisms
3. Barriers
4. Quantify effect
5. Refine target group
6. Best combination?
14Motivation to understand symptoms
Increase salience Personal relevance Social
marketing
Increase motivation
Benefits of early diagnosis/ treatment
Delay
Symptoms/signs
Uncertainty
Greater consideration of illnesses that
benefit from early diagnosis
List of possible explanations
Logical comparison with illness prototypes
Educate on/modify illness prototypes
Optimistic bias
Prompting/sanctioning by family/friends Personalis
ed action coping plans primary care
sanctioning
15Develop optimal intervention
Fewer waiting with haemoptysis or waiting more
than 3 weeks with other symptoms
1. Key processes and outcome
Mapped onto model of underlying problem
2. Mechanisms
Primary care Uptake in at risk groups Barriers
to consulting smoking cessation Anxiety and fear
3. Barriers
Not known needs exploratory trial
4. Quantify effect
Consider smokers vs ex-smokers Effects of COPD
5. Refine target group
Not known need experimental and qualitative
research
6. Best combination?
16 17Take control - Look after number one
Take action - It takes two to tango
Dont wait - Remember the three week rule
18Optimise evaluation
Improved survival
1. Outcomes
2. Randomisation
3. Recruitment and retention
4. Sample sizes
19Motivation to understand symptoms
Increase motivation
Increase Salience Personal relevance
Benefits of early diagnosis/ treatment
Delay
Symptoms/signs
Uncertainty
Greater consideration of cancer etc.
List of possible explanations
Logical comparison with illness prototypes
Educate on/modify illness prototypes
Optimistic bias
Prompting/sanctioning by family/friends Personalis
ed action plans primary care sanctioning
20Optimise evaluation
Improved survival
1. Outcomes
Waiting time to consult (and reasons) Intentions/
self efficacy/knowledge
Anxiety
Patient level intervention patient Practice
level intervention - practice
2. Randomisation
Target group 10 of practice populations ?40-50
recruitment rate Needs pilot study
3. Recruitment and retention
200 for pilot study to answer gaps in knowledge
4. Sample sizes
21Year 1 refining the intervention
- Focus group with primary care professionals
- Focus group with target patients
- Interviews with lung cancer patients
- Developing a self help manual and primary care
nurse consultation - Refining the evaluation
22200 RCT
- 10 enrolled in parallel qualitative
Intervention (10) Clinic and manual
Intervention (100) Clinic and manual
Control (100) Usual care
Questionnaire follow up 1 6/12
Qualitative follow up over 6/12 Qualitative
follow up with 10 professionals
Questionnaire follow up at 1 6/12
Qualitative follow up at 6/12 (10)
23Designing trials of complex interventions
Proposed evaluation of a complex intervention
Define and understand the problem and its context
Develop and understand the intervention
Develop and optimise the evaluation
Promising and warrants definitive evaluation
Unlikely to be cost effective
Implement without trial
Campbell, Murray, Darbyshire, Emery, Farmer,
Griffiths, Guthrie, Lester, Wilson, Kinmonth BMJ
2007 334 455-459