One of the purposes of this blog is to invite colleagues to add to the knowledge base on drug groups. To submit a paper or to provide your comments, please do so on the form on the Share Your Story page.
I’ll start the ball rolling with the following draft Data Based Medicine (DBM) papers:
We have draft papers in preparation on: Mood Stabilizers, Antipsychotics, Stimulants, Dopamine Agonists, Statins, and Hypoglycemics.
We need papers on: Treatments for Asthma, Hypertension, Osteoporosis, Antibiotics, Anti-Ulcer Drugs, Contraceptives, Analgesics, Anti-inflammatories, Drugs for Sexual Functioning, and others.
Principles underpinning DBM papers
The emphasis is on highlighting missing data. This may be:
- Hidden data: data from trials hidden through miscoding, ghostwriting, etc. (see recent posts)
- Missing data because of trial design: e.g., flawed design, too short, or surrogate outcomes
- Missing data because the right studies have not been done: as in adverse events
- Data that is hidden by statistical sleight of hand
The questions used in the current drafts are almost more important than the answers. We want questions that reveal our uncertainties and lacunae in our knowledge rather than questions that falsely reassure. The emphasis is on highlighting the unanswered questions where doctors and patients have poor information on which to make therapeutic judgments.
We encourage anyone writing, editing, or contributing to a paper to resist the temptation to advocate for a treatment, in particular a non-drug treatment, and to resist the temptation to denigrate the use of drugs for what they see as lifestyle or trivial purposes.
The aim is to appraise doctors and patients of the state of the data — in contrast to the state of what is called evidence — and to encourage them to observe previously unreported or poorly documented effects of drugs and to contribute these observations to the pool of data that helps guide decisions.
The aim is to make doctors and patients aware that the current state of the data at best permits guidance to supplement therapeutic judgments and does not mandate guidelines to replace clinical judgment or patient values.
Data Based Medicine believes that controlled trials are extremely important but have come close to being made into a fetish. They are, moreover, probably not the best method to reveal treatment related adverse events (see posts).
Authors are encouraged to return to the original definition of evidence based medicine which was an “integration of the best research evidence with clinical expertise and patient values” (Sackett & Rosenberg 1995).
This will result in papers that will have a judicious mix of best evidence and consensus view that may in some instances have a “return to the 1990s” quality to them. The best example may be pregnancy and antidepressants: it was so much received wisdom in the 90s that women should avoid drugs in pregnancy that there is almost nothing saying this. This has left a gap through which companies and other interest groups have been able to march – claiming that all the evidence points to the need to use antidepressants in pregnancy. Going back to the 90s will mean apparently going against what purports to be evidence.
Justifying such papers will be an interesting exercise that may need to tap all the resources this approach can mobilize.