This forum post is hidden because you have chosen to ignore leveller7. Show Details
This forum post is hidden because you have submitted an abuse report against it. Show Details
The question above is based on a debate with the same title at the European Congress of Epileptology in London. Class I evidence includes systematic reviews and randomised controlled trials (RCTs).
We covered the first-ever clinical trial: by James Lind for scurvy. There were 12 seamen with the disease after 2 months at sea. He divided them into pairs and each pair received a different treatment in addition to their usual diet: barley water, vinegar, cider, seawater, sulphuric acid, or oranges & lemons. In this way he found that oranges & lemons led to near-recovery of both subjects. However, he then recommended boiled limes to sailors - which of course destroyed the vitamin C that cured scurvy!
In a way, the above tale illustrates the nature of RCTs. They are an excellent way of assessing efficacy. However one must ensure the results are applicable - in terms of population, intervention, etc. Lind did not later use the same treatment as in the trial. This is akin to trials using a different drug dose or formulation to that used in regular clinical practice.
From the 'yes' camp, Dr Tony Marson informed us of the increasing international consensus on evidence-based medicine. Trial evidence is preferable to anecdote, expert opinion or marketing from drug companies. Being an epilepsy conference, the major RCTs in anti-epileptic drugs were covered - and how they underpinned NICE guidance even now. Not following conclusive evidence could result in unnecessary seizures, toxicity, and a worse quality of life for patients.
Dr Jacqueline French, the speaker against the proposition pointed out that RCTs only look at efficacy. Side-effects of drugs are usually only picked up from observational studies: examples are topiramate and renal stones, or vigabatrin and visual field defects, or felbamate and aplastic anaemia. Felbamate may indeed be superior to other drugs in suppressing seizures, but no controlled trial showed that it may have a rare but possibly fatal side-effect. Even for efficacy, RCTs apply a narrow definition. They select people in restricted age groups with no other diseases, possibly with very frequent seizures, and then like to compare the intervention with placebo. Then there is the whole question of industry funding of trials, suppression of 'bad' results, and statistical manipulation of data.
Using an example of a young woman who presented after two generalised seizures, we were asked which drug she should be offered first. Dr French suggested that valproate has the best evidence for stopping further seizures, but almost none of us would offer that to a young woman given its cosmetic effects, weight gain, and teratogenicity. Dr Marson argued that we should still tell the young woman the facts about side-effects and drug efficacy, derived from RCTs, and then let her make up her own mind.
What do you think?