The Evolution of Clinical Trials
Thursday, 08 January 2015
Medico-Chirurgical Hall
Professor Shaun Treweek. Professor of Health Services Research, University of Aberdeen
Notes
Minute of meeting held in the Society's Hall, Foresterhill on Thursday 8th January 2015
The President, Professor Mike Greaves, presided.
The President welcomed members and intimated the deaths of three former members. He gave brief eulogies for Dr David Millar, an ex President; Dr Margaret Brown and Professor Arnold Klopper. Full obituaries are filed in the Society's archive.
The President then presented the Strachan Bursary (a certificate and cheque) to Miss Kirsten Cumming.
He then introduced the evening's speaker - Professor Shaun Treweek, Professor of Health Services Research, University of Aberdeen. His topic was The Evolution of Clinical Trials.
Prof Treweek started by talking about the first recorded randomised clinical trial - that conducted by James Lund on Scurvy in sailors. In 1793, Britain had been at war for 20 years and of those sailors who had died, 6% had been killed in action compared with 80% who had died of disease especially scurvy. Various 'treatments' were in use, none of which seemed to work. Although citrus fruit was available, it was not widely used and had not been evaluated.
Lund evaluated 6 treatments on board HMS Salisbury, allocating sufferers of scurvy to one of the treatments and he also included a control group. After 6 days, those taking citrus fruit had recovered. Cider was the next best but the others were no better than the control group.
Turning to the present day, Prof Treweek said that the Cochrane Collaboration listed over 800,000 studies with about 25,000 trials being published each year and although randomised trials are at the hub of evidence based medicine, many are not well done and how to do them is almost totally evidence free. The main problem is with recruitment and half of all trials fail to reach the recruitment target.
Cochrane has a review on recruitment and there are 1.600 papers on the subject which do suggest some methods of improving recruitment such as telephone reminders, open recruitment and making them opt out rather than opt in. Other studies have looked at retention.
Data collection is often difficult and imputing and collating it is time consuming. O'Leary et all found that 82% of data collected did not get used.
Trials should be designed to be relevant to decision makers but participants are often not typical of those for whom the intervention will ultimately be intended. Prof Treweek also pointed out that many trials are actually targeted at answers which are already known.
Therefore trials should try to improve all aspects even if each improvement is relatively minor. He talked about Trial Forge which aims to improve the design conduct and analysis of trials and to evaluate the trial designs - a 'trial within a trial'. The first thing to improve is recruitment then data collection remembering that quite often too much data is collected.
He spoke about site selection and then mentioned a trial in which he had been involved 'Football Fans in Training' This trial had been so well designed that the interventions, which had been shown to be effective, simply carried on after the trial had finished.
At the end of his enthusiastic and fascinating talk, Prof Treweek answered several questions before being thanked by The President.