JAMA  23-30 Jan 2013  Vol 309
355    From time to time I like to shock my GP colleagues by saying that the most important part of any health system is the hospital and that secondary care takes rightful priority in any health budget. I’m really sorry to see our college president Clare Gerada, whom I admire unreservedly, putting her signature to support the Chief Medical Officer’s arrogant plans to close down local hospitals and devolve more care into “the community,” whatever that is. Public health “experts” and specialist bodies with vested interests have been suggesting this for the last 25 years: all that happens is that local hospital beds disappear while hospital admissions rise, and patients (especially outside London) travel long distances and complain of inhumane, rushed care. The gap between primary and secondary care, which should not exist, becomes ever wider as hospital doctors cluster into large remote units and interact ever less with their GP colleagues.

The American system is learning to do without general practitioners altogether, which is quite mad: and Britain’s problems are gradually moving closer to America’s as our NHS is being dismantled. All this is by way of preamble to two studies of US hospital usage following discharge conducted by two groups of friends and colleagues from Yale. In the first, a brilliant cardiology fellow, Kumar Dharmarajan, together with the forces of the CORE team led by Harlan Krumholz, look at the reasons for hospital readmission after hospitalisation for heart failure, myocardial infarction, or pneumonia. These Medicare patients are all age 65 or over: their mean age is 80. They come out of hospital weak and disoriented, and about 20% overall end up back inside within a month, in most cases with a different primary diagnosis. I don’t know if there is a British literature on this or what it shows. I certainly agree with Harlan that there must be a “post-hospital syndrome,” which probably affects all ages. And in both our health systems, “communities and care systems need to adopt a population-based measurement system that brings accountability to the triple aim of improving patient experience, enhancing population health, and reducing costs.” (see Viewpoint).

364    Many discharged hospital patients end up in the hospital emergency department even if they are not readmitted. This study, carried out by Robert Wood Johnson Scholars from Yale, examines what happened following over 5 million hospital admissions for a wide range of conditions in three US states. About 18% of patients ended up coming back to the emergency department, representing 40% of acute post-hospital care. Looking briefly through the data on their multiplicity of diagnoses, I think the vast majority could have been sorted out by experienced primary care doctors with access to basic diagnostic tests.

372   The next paper looks at readmission rates for American paediatric hospitals. These are typically low (6.5% within 30 days) and although there is some variation between centres, it is by no means as huge as you will find in most sections of the Dartmouth Atlas of Variation in Healthcare.

381    Readmission is now being looked at routinely as a quality marker for providers in the USA: but does this make any sense? Do we know of any effective means to reduce it? This study of a quality improvement initiative for care transitions by healthcare and social services personnel and Medicare Quality Improvement Organization staff in defined geographic areas was essentially negative. A Viewpoint piece looks at the logic of using 30 day readmission as a quality marker for heart failure care and decides there isn’t any. “Each quality metric should be data driven and patient focused. The 30 day readmission measure does not appear to meet either of these criteria. The important remaining question is whether patients will benefit from this measure. Thus far, it appears that this metric has the potential to result in more harm than benefit for patients with heart failure.”

NEJM  24 Jan 2013  Vol 368
307   I very much hope that every reader of this column has already signed the http://www.alltrials.net/ petition calling for all human trials from 1992 to be registered and published. It’s amazing—scandalous—bizarre—that this should still be needed. The Association of the British Pharmaceutical Industry has now produced a weak response which questions whether it is really such a good idea: “the debate around clinical trial transparency is important, but it is essential that patients have confidence that the medicines their doctors prescribe for them are appropriately safe and effective. However much data is published—or not—the regulatory authorities have access to all the relevant data as part of the approval process for new medicines.”

Hmm. Why then did we have the withdrawal of Vioxx and Avandia; why do we still not know if Tamiflu has any useful effect; and why was it so long before there were safety concerns over epoetin? It’s partly because of hidden data, and partly because of inadequate surrogate endpoints. Epoetin is very good at raising haemoglobin levels in renal dialysis patients: the only problem is that those with a “normal” Hb die sooner. Now there is a new kid on the block, called peginesatide, which works as well as epoetin but only needs to be injected monthly. Here is a study (EMERALD 1) which shows that the stuff not only raises Hb but is as safe as epoetin using a composite end-point of all-cause death, myocardial infarction, stroke, or serious cardiac events in dialysis patients.

320   However, when you compare peginesatide with darbepoetin in patients with advanced renal failure who are not on dialysis, the picture is less favourable. Cardiovascular end-points are commoner with the new agent. That’s a high price to pay for greater convenience of use, and I’m not convinced we yet know enough about this new agent to license it for general use.

333    When the H1N1 influenza pandemic first struck, one of the scariest features was its ability to kill pregnant women. This Norwegian study shows that it could harm the fetus as well. Women who developed clinical influenza during pregnancy ran about twice the chance of fetal death compared with those who did not, or who were vaccinated.

341   It is evil to grow tobacco because smoking kills people, having first rendered them addicts. It is evil to sell tobacco for the same reason. It is evil to use tobacco to raise tax revenue, again for the same reason. The conclusion of this survey from the tobacco-growing, -selling, and -taxing USA is exactly that from the Lancet a couple of weeks ago: “Smokers lose at least one decade of life expectancy, as compared with those who have never smoked. Cessation before the age of 40 years reduces the risk of death associated with continued smoking by about 90%.”

351   And among women, the evils of smoking have grown: “The risk of death from cigarette smoking continues to increase among women and the increased risks are now nearly identical for men and women, as compared with persons who have never smoked.” Come on Obama, when you have finished with the gun lobby, take on the tobacco industry. Watch Lincoln at least once a month. Not worth a civil war? A quarter of your countrymen are already dying because of tobacco.

Lancet  26 Jan 2013  Vol 381
303    Friends assure me that the Lancet has many virtues—good book reviews, for instance, and a consistent interest in global health—but whenever I see it publish an interventional trial, I groan inwardly. The usual rules apply to the two trials of regorafenib reported in this week’s issue. “The study sponsor (Bayer) provided the study drug and collaborated with the investigators on protocol design, data collection and interpretation, and preparation of this report.” “The sponsors funded writing assistance.” And perhaps they have ordered plenty of reprints, because the paper states that “regorafenib could be a new standard of care in late-stage metastatic colorectal cancer.” But is that really true? Regorafenib (already licensed) is a last-ditch drug with a range of actions to shame a blunderbuss: it has potent inhibitory activity against vascular endothelial growth factor receptors 1—3 (VEGFR1 [also known as FLT1], VEGFR2 [KDR], and VEGFR3 [FLT4]), PDGFRB, FGFR1, RAF, TIE2, and the mutant oncogenic kinases KIT, RET, and BRAF. And it has a range of adverse effects to match: hand-foot skin reaction, fatigue, diarrhoea, hypertension and rash, or desquamation. In this trial it produced a survival benefit of 1.4 months at a cost of about $20K, and 93% of patients suffered treatment-related adverse effects. But the authors (who include ghosts) are allowed to get away a positive conclusion, while the journal clears its conscience in an editorial stating, “we find it difficult to agree with the authors’ conclusion that the drug ‘could be a new standard of care in late-stage metastatic colorectal cancer’ in view of the small incremental survival benefit, reported toxic effects, and low likelihood of future analyses finding regorafenib to be a cost-effective option.” A classic case of having it both ways, and no way for a serious medical journal to behave.

313   Like most jobbing doctors, I stand well back when approached by anything resembling serious immunology. However, the question of donor kidney rejection is undoubtedly important and this paper seems to show that until now it has not been very well understood. A Parisian immunohistological study shows that there are four basic subtypes of allograft rejection, of which by far the most serious is a type not previously described, in which antibodies rather than T-cells produce an acute vasculitis.

320   Sometimes the Lancet features clinical reviews which I think every doctor ought to read, but more often it features accounts which go out of their way to be impenetrably learned. This week we have the pemphigoid diseases—gosh, there are more of them than you ever thought, and they are so, so complicated; and Noonan syndrome, which “was characterised by Jacqueline Noonan, who reported nine patients with pulmonary valve stenosis, small stature, hypertelorism, mild intellectual disability, ptosis, undescended testes, and skeletal malformations.” To my mind it should therefore be called Noonan’s syndrome in any respectable British journal, but they have different rules for the eponymous genitive in America where this was written. To save you looking up “hypertelorism,” it means widely spaced thingies:—eyes in this case. When you can’t even understand the opening description, abandon hope for the rest.

BMJ   26 Jan 2013  Vol 346
A systematic review and meta-analysis of early interventions to prevent psychosis takes as its starting point the idea that schizophrenia is usually preceded by a prodromal period, and “it has been shown that 22-44% of people at ultra high risk will transition to develop schizophrenia.” For me, that was a learning point and led me to read the whole paper, which is well written and clearly presented despite the difficulty of the topic. It’s interesting to note that the lead author describes herself simply as “systematic reviewer.” There are no trials which give definitive or lasting results, but there are signs of promise from cognitive behavioural therapy and, more surprisingly, from omega-three fatty acids.

Mammographic screening for breast cancer was introduced with a blaze of positive publicity and information leaflets that grossly overstated the benefits and failed to mention the problems of overdiagnosis. As we repent at leisure, the thing that screening proponents most fear is that women may decide not to undergo screening because they are scared by the facts. This is rather similar to the ABPI’s defence of data hiding: if people know all the harms as well as the benefits, then they may no longer have faith in medicines. Don’t scare the punters. This interesting study shows that most women are perfectly able to understand the problem of overdiagnosis from routine mammography when it is clearly explained. The tipping point for most women in the focus groups came when more than 50% of “abnormals” turn out to be false positives; but there was a wide range of responses. If I were of the opposite gender contemplating mammography, I would most of all want to know the all-cause mortality difference, which is discussed elsewhere by that eminent sceptic, Michael Baum.

While in the USA, we lived close to the Yale Peabody Museum, and so had several chances to admire the special exhibition of human bugs which was held there, with enthusiastic advice for onlookers to admire and cherish these resourceful parasites and ignore the minor discomfort they occasionally cause. I am pleased to see that the BMJ takes a more robust stance on bed bugs: kill the lot of them, say I, in chorus with the French authors. Look under hotel mattresses, in cracks in the walls, in furniture upholstery, and then get in a professional bug-killer, because most insecticides are neurotoxic, and you don’t want to go to bed in their company either. Sleep tight.

Plant of the Week: Cimicifuga foetida

Cimicifuga means bug-dispelling, and this genus of plants has been used to kill bedbugs for centuries, especially in Siberia, where I guess it is particularly advisable to look under mattresses and in cracks in the hotel walls. I had not realized how many members of the bugbane family have been brought into cultivation as late-flowering garden perennials. The one that you generally see in garden centres is the purplish-leaved form of C simplex, which has white flowers in September.

These are generally tall but self-supporting plants which disappear without trace in the winter but get up to 2 metres high in late summer. I don’t think I have ever seen C foetida, and we haven’t got space to grow it, but it sounds quite a temptation to the collector:

“N.E. Asia 1777. A very rare plant in cultivation, and the only one that can be called yellow, on account of the yellow sepals around each white tassel of stamens. The stems branch and the centre spike arches gracefully, like a shepherd’s crook. The leaves are dark green, deeply veined and much divided. The flowers have the same fragrance as the yellow water-lily or Brandy Bottle. The arching sprays of green pods are extremely handsome.”

Graham Stuart Thomas Perennial Garden Plants, or The Modern Florilegium. 3rd ed (revised) 1990