NEJM  13 Mar 2014  Vol 370
1029 Doctors, by and large, make bad scientists. We train our minds for years in some of the hardest intellectual disciplines, and then make do with the sloppiest excuse for thought when it comes to believing what we wish to. All of us learned, at some time between the ages of five and fifteen, that just because something happens after something else, it doesn’t mean that the first thing caused the second. The whole endeavour of medical research is to get beyond post hoc, propter hoc. And yet we go back to square one with embarrassing regularity. We deploy a device for frying renal nerves and are awe-struck at the 30+mmHg drop in BP that follows. Then we do the same thing by random allocation and blinded follow-up, and the drop turns out to be less than 10mm. Five years ago, a special report appeared in the New England Journal describing a 36% drop in post-surgical morbidity and mortality in a “global population” following the introduction of surgical checklists. In three of eight sites, the intervention made little difference, but in others there were reductions in adverse outcomes of 30-50%. Soon surgical checklists became mandatory in whole countries and regions. Now, guess what? The introduction of mandatory checklists turns out to have made no difference in some places, as shown by this special report from Ontario. Whereas in another before-and-after report, from a neurosurgical unit in Finland, it made an impressive difference. That’s life, guys: science begins where “look at these numbers, you gotta do this” leaves off. In fact the science which is most likely to tell us about how checklists work or don’t work is not numerical at all, but descriptive and ethnographical.

OL  Obinutuzumab. What would our forebears have made of a word like that? They might have guessed it was the compound name of some ancient king or priest meaning “victorious under the protection of Obinu.” And that’s not so wide of the mark. It is a made-up compound name, which follows the conventions of the Expert Advisory Panel on the International Pharmacopoeia and Pharmaceutical Preparations. Working backwards through the syllables, “mab” means monoclonal antibody, “zu” means humanized, “tu” means a miscellaneous tumour, and Obinu is indeed the name of the great god who will bring victory to Obinutuzumab. In this trial, the mighty warrior was pitted against his rival rituxumab, in patients with chronic lymphatic leukemia. The median age of the patients was 73, so comorbidity was common. In past trials rituxumab has helped to prolong life in CLL patients who were fit, but not those with significant other conditions. The Hoffman-Roche trial was a massive logistical effort, randomizing 781 patients in 189 centres across 26 countries to three arms: chlorambucil alone, chlorambucil-rituxumab, or chlorambucil-obinutuzumab. It provided a fairly clear win for the last group. Kar-ra-lit-um-ob-inu-tu-zu-mab! That’s Sumerian for “hurrah for the victory of the humanized monoclonal anti-tumour antibody of Obinu!” Somewhere in the deserts of Iraq, you may find an ancient clay tablet bearing these cuneiform syllables. The Sumerians generally got to things first.

JAMA Intern Med  Mar 2014
OL  In the last few years, attention has gradually swung from the alleged benefits of tight glycaemic control in diabetes to the real hazards of hypoglycaemia. This report looks at the scale of the problem of insulin induced hypoglycaemia in US emergency departments. Because there is no denominator, it’s hard to work out the meaning of the figures, except to show that people over 80 taking insulin are at very much higher risk than younger diabetics. “The risks of hypoglycemic sequelae in this age group should be considered in decisions to prescribe and intensify insulin. Meal-planning misadventures and insulin product mix-ups are important targets for hypoglycemia prevention efforts.” I suspect, though, that meal-planning misadventures and other mix-ups are an intrinsic part of the being-over-80 experience. I can feel it coming on already.

JAMA  12 Mar 2014  Vol 311
1023  It’s taken a long time to arrive, but it’s good to see this randomized comparison of two of the commonest gynaecological repair procedures. For good measure, the investigators in 9 US centres threw in behavioural therapy with pelvic floor exercises for half the women in each surgical group. This made no difference. Nor did the type of surgery—sacrospinous ligament fixation and uterosacral ligament suspension—carried out for apical vaginal prolapse and stress urinary incontinence. Time for some gynaecologists to come up with an Option Grid.

1035  When it all began, it seemed so promising. Our fate, we were told, is written in our genes, so if we could read our whole genome we would be able to predict illness with astonishing accuracy. This delusion continues to sustain a research effort consuming billions of dollars and billions of hours of scientific time. Here is a summary of progress to date: “In this exploratory study of 12 volunteer adults, the use of whole genome sequencing (WGS) was associated with incomplete coverage of inherited disease genes, low reproducibility of detection of genetic variation with the highest potential clinical effects, and uncertainty about clinically reportable findings. In certain cases, WGS will identify clinically actionable genetic variants warranting early medical intervention. These issues should be considered when determining the role of WGS in clinical medicine.”

1045  Getting to the starting block of a research project is hard work. Reaching the finishing line is even harder. There’s many a slip, and a survey of randomized controlled trials based on archived protocols approved by six research ethics committees in Switzerland, Germany, and Canada between 2000 and 2003 shows that a quarter never reached completion. Poor recruitment was the commonest reason. And yet the effort that these trials represented is doubly wasted if nobody can learn lessons from their difficulties. “Greater efforts are needed to ensure the reporting of trial discontinuation to research ethics committees and the publication of results of discontinued trials.”

1063  Another week, another great medical student paper from Yale. Jessica Becker looked at 96 clinical trials for which the primary results were published between 1 July 2010 and 30 June 2011 in high-impact journals, and that were registered on ClinicalTrials.gov and reported results there too. Overseeing her efforts (you guessed it) were Joe Ross and Harlan Krumholz. She discovered that 93 of the 96 trials had at least one discordance among reported trial information or reported results. And these were the best of the trials, published in journals like NEJM and The Lancet. This is wonderful police work, and it points yet again to fundamental flaws in the quality control of trials conducted on human subjects and their reporting to clinical decision-makers.

Lancet  15 Mar 2014 Vol 383
955  “And God saw every thing that he had made, and, behold, it was very good.” Gen 1.31. This meta analysis of new oral anticoagulants for atrial fibrillation breathes a similar air of contentment. They are, quite simply, better than warfarin, especially when INR control is suboptimal. In an overall comparison with warfarin, they decrease all-cause mortality by about 10%, and reduce stroke and other embolic events by about 19%. What’s not to like? Well, a bit more gastrointestinal bleeding and a lot more cost. So if you have AF and feel strongly, you should follow the advice of the director of NICE and march along to your GP and demand some dabigatran. Or apixaban. Or rivaroxaban. Or wait a bit, until edoxaban has been approved. Why should you worry about cost to the NHS? When did anyone ever ask you about how much the nation should spend on health, or fulfil an electoral promise about the NHS?

963  It’s blood-in-the-boots time again folks. In the UK, more than 90% of ruptured abdominal aortic aneurysms are repaired by open surgery, whereas in the USA, the figure is 79%. Over there, they offer surgery to a lot more people too: 80% of those with ruptured AAA as opposed to 58% here. Yet their in-hospital mortality is somewhat better than ours, at 53 versus 66%. So we must learn lessons from America, but it isn’t quite clear what these are. The main one seems to be to rush off people with rAAA to high volume centres.

970  As an ageing man who likes food and doesn’t get much time away from a desk, I puzzle about the importance of overweight. It seems to me that the main problem is the tightness of one’s clothes. This is somewhat confirmed by this Lancet analysis, which concludes that “Interventions that reduce high blood pressure, cholesterol, and glucose might address about half of excess risk of coronary heart disease and three-quarters of excess risk of stroke associated with high BMI. Maintenance of optimum bodyweight is needed for the full benefits.” The last has to be true. Taking losartan and a statin has made no difference whatever to the tightness of my clothes.

999  In a few weeks’ time, all who can should visit Framingham, Mass. The Woodland Garden there will be full of the most wonderful erythroniums, trilliums, and sanguinarias. Others may wish to make the pilgrimage because it is 65 years since the first subjects were recruited to the Framingham Heart Study, marking the beginning of a new era in epidemiology. Here is a fascinating history of the project, which was stimulated in part by F.D. Roosevelt’s death from galloping hypertension. It’s a wonderful reminder of how much good work could be done by means of messages written on rickety typewriters, simple paper charts, and card indexes. No internet for the first 40 years.

BMJ  15 Mar 2014  Vol 348
I’ve recently been fortunate enough to be offered a small role at the UK Cochrane Centre, and a couple of weeks ago I became aware of my ignorance of the methods and reporting of network meta-analyses. Do you share this ignorance, dear reader? If not, you are truly remarkable, because this article about the reporting of network analyses reveals what can only be described as confusion on a grand scale. “In 52 reports that ranked interventions, 43 did not report the uncertainty in ranking. Overall, 119 (98%) reports of network meta-analyses did not give a description of the network or effect sizes from direct evidence, indirect evidence, and the network meta-analysis.” So next time somebody starts talking to your group about a network meta-analysis, just put up your hand and say “Please Miss, I don’t understand.”

When faced with a possible deep vein thrombosis, I try to apply the Wells rules and do a d-dimer test, but I’m constantly frustrated at how often they simply don’t apply to the clinical situation. This individual patient data meta-analysis, whose authors include the great Professor Wells himself, concludes that “Combined with a negative D-dimer test result (both quantitative and qualitative), deep vein thrombosis can be excluded in patients with an unlikely score on the Wells rule. This finding is true for both sexes, as well as for patients presenting in primary and hospital care. In patients with cancer, the combination is neither safe nor efficient. For patients with suspected recurrent disease, one extra point should be added to the rule to enable a safe exclusion.” Add to this pregnant women and elderly people with fragility bruises, and you find yourself giving a lot of LMW heparin until you can get a scan, just in case.

Later today, I’ll go off and see seven hours’ worth of poorly people, half of whom will tell me that their symptoms are still bad although they have dosed themselves up with ibuprofen, paracetamol, and steam inhalations. I blame 111 for instilling the belief that these things are of benefit. Here is a corrective study from UK primary care: “Overall advice to use steam inhalation, or ibuprofen rather than paracetamol, does not help control symptoms in patients with acute respiratory tract infections and must be balanced against the possible progression of symptoms during the next month for a minority of patients.” People who took ibuprofen were more likely to return with symptoms, though few of these were serious: in particular, very few people got a wheeze.

So it’ll soon be time to prescribe quite a lot of amoxicillin, some of it with advice about delaying use. This next UK trial shows that, “Strategies of no prescription or delayed antibiotic prescription result in fewer than 40% of patients using antibiotics, and are associated with less strong beliefs in antibiotics, and similar symptomatic outcomes to immediate prescription. If clear advice is given to patients, there is probably little to choose between the different strategies of delayed prescription.” Nice to read such a clear message.

Imagine being able to analyze all clinical encounters. How much shared decision making was really done? What was the connection between the history, the findings, the decisions made, and the evidence used? How much assessment could be achieved by speech analysis and natural language processing? Although this might seem unrealistic, the research to achieve this goal has already been initiated. Actually, these are not my words but Glyn Elwyn’s. Read more in his terrific piece on the routine digital recording of clinical encounters. This really will change the practice of medicine in the future.

Plant of the Week: Magnolia denudata

Still the most perfect of all the magnolias, ours has just opened its pure white chalices against a blue sky, a month early. A moment of garden bliss.