9 Jul, 12 | by BMJ
Arch Intern Med 25 June 2012 Vol 172
909 The Archives are about to mutate into JAMA Internal Medicine, but I generally find them a better read than JAMA proper. One reason is the abundance of lively comment—and in the case of this paper on sex differences in the protective effect of statins, I find the comment more believable than the paper. This is a meta-analysis of eleven double-blinded RCTs of statin therapy for the prevention of recurrent cardiovascular events, and purports to show that although statins are as good at preventing recurrent cardiac events in women as in men, they do not prevent stroke or reduce all-cause mortality in women. Two British luminaries contest this, arguing that the literature search was incomplete and that the meta-analysis does not include a number of key studies which show that the protective effect of statins in women and men is remarkably similar in every category.
922 Even livelier is the comment which accompanies this Spanish study of an intervention to shorten hospital stays for community-acquired pneumonia. The intervention is simple and cost-free—all you need to do is encourage early mobilization and use a checklist to determine when to switch from intravenous to oral antibiotics and when to discharge the patient. This three-step programme certainly got patients home sooner with no effect on clinical outcomes, satisfaction or readmission, and so saved a lot of money. But there was no attempt to measure compliance with the protocol by clinicians, which in real life usually barely gets into double percentage figures with interventions of this sort. If this study managed to get everyone to comply, then it may not be a real life scenario. The commentator starts and finishes with the devil, and I’m afraid the title says “the devil is in the detail.”
938 Pay for performance diverts effort from high risk patients who need it towards low risk patients who don’t need it, and can even be harmed by it. This was demonstrated a couple of years ago with glycaemic control among British diabetics, and it’s now demonstrated again in Americans with diabetes—this time in respect of blood pressure control. Over 700,000 diabetic patients were subject to an “action measure” by the Veterans’ Administration that encouraged lowering of the systolic BP below 140 mm Hg. “While 94% of diabetic veterans met the action measure, rates of potential overtreatment are currently approaching the rate of undertreatment, and high rates of achieving current threshold measures are directly associated with overtreatment.” In other words, when “action measures” are imposed, low risk patients suffer the adverse effects of treatment to target. And these targets are generally derived from studies where the vast majority of the effect was derived from treating the patients at highest risk, as demonstrated by Timbie, Hayward and Vijan two years ago.
947 A group of notable US outcomes researchers derive a beautifully predictive prediction tool for initial survivors of in-hospital cardiac arrest, based on the outcomes of 43,000 resuscitated patients. Its eleven points lie on a straight downward line. This time the invited commentary is merely curmudgeonly, arguing that you’re either dead or you aren’t, so you mustn’t give up on any group, even the ones with a 2.8% chance of surviving without neurological damage.
955 Now here’s a lively, must-read paper that you probably wouldn’t find in any other main journal: not a review, not original research, not just an opinion piece—but an intelligent examination of what we mean by pulmonary embolism and how this sheds light on the phenomena of diagnostic drift, overtreatment, and futile therapeutic innovation. “Trials of newer anticoagulants and longer durations of anticoagulation have not yielded real improvements over heparin, inviting doubts regarding its efficacy. Thus, PE is the quintessential diagnosis of medicine not because it represents our greatest success, but because it captures all the complexity of medicine in the evidence-based era. It may serve as a metaphor for many other conditions in medicine, including coronary artery disease. New trials in the field continue to test trivialities, whereas fundamental questions are unanswered.” Do try and get hold of it.
Ann Intern Med 3 July 2012 Vol 157
1 A trial to make you pause. It was carried out in two leading academic hospitals in the USA using a highly labour-intensive intervention to prevent medication errors in the month following admission for an acute coronary event or acute heart failure. “Among 851 participants, 432 (50.8%) had 1 or more clinically important medication errors; 22.9% of such errors were judged to be serious and 1.8% life-threatening. Adverse drug events occurred in 258 patients (30.3%) and potential ADEs in 253 patients (29.7%).” But the point is that the intervention made no difference at all, even though it was intelligently thought out and well executed, as far as one can tell. Moral: half of patients who are of above average “health literacy” will suffer some kind of “preventable” adverse drug event after a cardiac admission, and there is nothing we know of that will make any difference.
11 So what does “preventable” really mean in contexts like this? Many rehospitalizations following myocardial infarction are also deemed “preventable” and preventing them is a quality of care indicator in several health systems. Here is an observational study from little Olmsted county which went through the charts of 3,010 patients who were discharged following MI between 1987 and 2010. Just over 40% of the 643 readmissions were definitely MI-related: the rest were either definitely not or uncertainly related. Co-morbidity played a large role, and complications of revascularization were relatively common too. How truly preventable these readmissions were is not at all clear.
29 “Despite increasing emphasis on the role of clinical decision-support systems (CDSSs) for improving care and reducing costs, evidence to support widespread use is lacking.” I’m certainly a non-user, not because I don’t think they’re useful, but because of old age and inertia. I probably disqualify myself from comment for those reasons, but the fact is that doing a great big systematic review of CDSSs and concluding that they could be useful is no way to change anyone’s behaviour. Striking people like me off the medical register and inspecting every consultation for the use of a decision aid may be the way forward, but please don’t bring it on for another couple of years.
JAMA 4 July 2012 Vol 308
43 In an editorial, Howard Bauchner et al try to explain why all the journals in the group are now going to be renamed with JAMA instead of Archives at the front. Perhaps to distinguish itself from all the rest, the parent journal should call itself JAMA Boring. There is really nothing of general interest to report on from this issue, but for those contemplating the use of herpes zoster vaccine in patients with immune-mediated diseases who are taking immune-modulating therapy, here is a study which shows that live attenuated vaccine does not cause disease and was fully protective over the two years of the study. That’s about it for now. Maybe by next week it will have progressed to JAMA Slightly Interesting.