 |
 |
Hi PSA screening is the obvious example. I would put frequent self monitoring of glucose in type 2 DM in the same category, especially if linked to a "tight control" regime (especially in older people at risk of chronic hypoglycaemia). In my field (stroke), there's no doubt that scanning for carotid stenosis (ultrasound, CT, MRI or catheter angios) has done more harm than good over the years. Although trials have now shown which patients ought to benefit from surgery, the benefit>harm window is quite narrow and so many ops are done outside the limits of evidence (eg for not-recently symptomatic stenosis, especially in women), that some surgical centres probably still do more harm than good. The same definitely still applies to newer obsessions like scanning to identify cases for intracranial artery stenting, ec-ic bypass and device closure of PFOs, though it's possible that technical improvements and better case selection might alter the ben/harm ratio in future... Incidentally I'm disappointed to hear that anal leeching doesnt work for stroke (it's a much more attractive idea than some of those now being proposed), but not sure if this is evidence of lack of benefit or just lack of evidence. Despite huge pharma investment in trials of dozens of drugs for acute stroke - nearly all abandoned - we can only rule out a modest but extremely useful benefit in v few of them. Of the 2 "effective" drugs, tPA has only been shown to be safe in a tiny minority of pts - though as usual its use is pushed way beyond the evidence base - and the benefits of aspirin are so small (NNT ~ 100) that 2 n=20K megatrials had to be combined to get a significant p-value. Getting to the bottom line ito benefits/harms of clinical test needs even bigger numbers. Thus despite all the big talk about EBH, I think most clinicians have barely started to get to grips with the true limitations of reliable clinical evidence. Cheers David Barer Date: Tue, 3 Apr 2012 11:17:46 +0000 From: Jeremy Howick <[log in to unmask]> Subject: Diagnosis Hi, There are well-known examples of therapies that were not evidence-based and were shown to be harmful (treating stroke with leeches applied to the anus, treating ventricular arrhythmias after myocardial infarction with antiarrhythmic drugs, etc.) I'm looking for parallel examples of diagnostic tests or strategies that have been introduced without sufficient evidence and subsequently were revealed to be useless or harmful when the tests were rigorously evaluated in unbiased tests. Jeremy