Post on 21-Dec-2015
yuri.kulikov@warwickorthopaedics.org
Collective Uncertainty Project: Hope for Randomised Clinical Trials in
Trauma and Orthopaedics Kulikov Y11, Brydges S22, Girling A33, Lilford R33, Griffin D11
11 Warwick Medical School, UK22 eLab, University of Warwick, UK33 Public Health & Epidemiology, University of Birmingham, UK
GlossaryCollective equipoise (Freedman, 1987)
justifiable uncertainty within the medical profession about which treatment is most effective for a particular condition
implies that there is no (rational) preference whatever
Collective uncertaintyin reality collective equipoise is rarely exactly evenly balanced, but uncertainty remains about the best treatment option
though individuals may have a preference for one treatment, they are balanced by the others with the opposite view
required before any clinical trial can be approved by ethics committee
Prior probabilityPrior sense of the effects of treatment(s)
Can be based entirely on expert opinion
yuri.kulikov@warwickorthopaedics.org
Randomised controlled clinical trials (RCTs) in Trauma & Orthopaedics
produce the most reliable evidence about the effects of clinical carebut
are (very) expensive and (very) difficult for a number of reasons (McCulloch et al, 2002)
in particular becauselack of surgeons’ individual equipoise in a specific clinical situation, which is often rational, has been shown to be a major obstacle in participant recruitment
Collective Uncertainty Project
• to apply Collective Uncertainty to individual clinical scenarios
• to assess degree to which uncertainty must be present as the fundamental criterion for eligibility for a trial
yuri.kulikov@warwickorthopaedics.org
Design (materials and methods)
Secure website
Clinical data
Clinical images
Interactive Voting Scale
yuri.kulikov@warwickorthopaedics.org
Design (materials and methods)
Integrated into the UK Heel Fracture Trial (UK HeFT) – conservative versus operative treatment for displaced fractures of calcaneumExpert panel of 10 Consultant Trauma Surgeons from 8 hospitals across the UKConsequent potentially eligible for randomisation clinical cases identified via the UK HeFT published via secure online forum after consent being obtained in 6 weeks follow up clinic or laterSurgeons alerted about new cases via email and SMS (optional) and express their opinion onlineLevel of uncertainty assessed by application of 80:20 ethical uncertainty distribution limit (Johnson et al, 1991), by accounting all votes in favour of operative treatment (a bit better + significantly better + much better)“strong votes” (significantly better + much better) were accounted separately to demonstrate support or otherwise for recommendation
yuri.kulikov@warwickorthopaedics.org
Results 30 eligible cases, of those 17 (56.7%) not randomised for HeFT4 bilateral injuries11 declined participation
4 did not want to have surgery5 wanted to have surgery1 wanted to be treated privately1 did not want to be randomised
1 randomised in error1 had previously infected tibial plate same side (removed)
Of 13 randomised 2 (15.4%) declined intended treatment (surgery)
On average 5 surgeons voted per case (min 3, max 8)
26 cases incl. all bilateral injuries could be recommended for randomisation
LegendchXXX – case number
Grey bars – votes by individual surgeons
Burgundy bars – cumulative average votes
yuri.kulikov@warwickorthopaedics.org
Results
3 cases the panel recommended for non-operative treatment
CH007 – 8.4% for surgery (1.2% strong votes)
CH027 – 13.7% for surgery(1.7% strong votes)
CH014 – 15% for surgery (4% strong votes)
1 case the panel recommended for operative treatment
CH015 – 87% for surgery (72% strong votes)
UK HeFT: CH007 – randomised to non-operative treatment; CH015 – randomised to operative treatment;
CH014 – declined to take part (did not want surgery);CH027 – declined randomisation (treated non-operatively)
yuri.kulikov@warwickorthopaedics.org
Discussion
PitfallsSurgeons’ reluctance to vote (maybe overcome if votes will be more relevant)
Technical (PACS required in hospitals involved; very few glitches so far, overall simple cheap and stable system)
StrengthsEase of use (3-5 min to vote per case)
No geographical boundaries
Instant application in real time (48 hours required to obtain votes)
Ethical value (randomisation only when the panel feels appropriate; individual, personal approach )
Measuring Collective Uncertainty in our study showed potential to DOUBLE (from 43.3% to 86.7%) patient recruitment for the UK HeFT
At the same time patients would not have been offered randomisation where current specialist opinion (prior probability) is strongly in favour of one or another treatment
Broader inclusion criteria possible, because every patient is assessed for randomisation individually
Both surgeon and patient are supported in their decision by an expert panel
The Uncertainty Measurement is an opinion (prior probability); the final decision remains between a treating surgeon and a patient
yuri.kulikov@warwickorthopaedics.org
Conclusion
We propose Measurement of Collective Uncertainty to be introduced into Surgical RCTs where decision about randomisation is especially challenging (operative vs non-operative; standard against new but popular well-marketed treatments etc)
It is possible to set up international expert panels to suite international studies
“Empowering choice will be given precedence by those who, like me [us], think the obligation to respect individual autonomy outweighs the common good in all but the most extreme cases…” (Lilford, 2003)
References• Freedman B. Equipoise and the ethics of clinical research. N Engl J Med 1987;317:141-5• Johnson N, Lilford RJ, Brazier W. At what level of collective equipoise does a clinical trial become
ethical. Journal of Medical Ethics 1991;17:30-34• Lilford RJ. Ethics of clinical trials from a bayesian and decision analytic perspective: whose equipoise
is it anyway? BMJ 2003;326:980-1• McCulloch P, Taylor I, Sasako M, Lovett B, Griffin D. Randomised trials in surgery: problems and
possible solutions. BMJ 2002;324:1448-51