Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Booking the Cooks
- L Sam Lewis (4 July 2009)
-
The wrong Ingredients ?
- Des Spence (5 July 2009)
-
The flip-flop rhythm of medical science
- Mark Struthers (6 July 2009)
-
Right Ingredients, Wrong Recipe ??
- L Sam Lewis (20 July 2009)
-
Occam’s razors are simply too blunt for banana splitting
- Mark Struthers (21 July 2009)
-
Re: Occam’s razors are simply too blunt for banana splitting
- L Sam Lewis (22 July 2009)
-
The Question of Our Time
- Des Spence (28 July 2009)
|
|
|||
|
L Sam Lewis, GP Surgery, Newport, Pembrokeshire, SA42 0TJ
Send response to journal:
|
Now I like a good story as much as the next man, and I'm partial to an Aussie Chardonnay.. so why let the truth get in the way of some long- held habits of mind ? As I read Des' claim that the decline in cardiovascular mortality was largely unexplained , I found my inner angel chafing at the leash. Unsurprising to me that the 80s decline be explained by BP-lowering, the 90s by Statins, and the whole period by general decrease in Smoking, and at least some attributable to better health rescue services. So I quickly checked using the BMJ's useful Search facility and came up with Unal et al's demonstration that two-thirds of the decline was explainable by concomitant BP, Smoking , and cholesterol reduction. To be fair to Des - that still leaves one-third to explain. I suspect it's not a magical effect, nor the ineffable mystery of general practice, but rather the combined effect of splashing Aspirin, Warfarin, Thrombolysis, and CABG all around the place.. And I'm not averse to the notion that QOF might have rewarded some of that effort, after the fact. Sam Lewis Modelling the decline in coronary heart disease deaths in England and Wales, 1981-2000: comparing contributions from primary prevention and secondary prevention Belgin Unal, Julia Alison Critchley, and Simon Capewell BMJ 2005 331: 614. [Abstract] [Full Text] [PDF] Competing interests: None declared |
|||
|
|
|||
|
Des Spence, GP Glasgow G20 9DR
Send response to journal:
|
This paper [1] is one of many that seek to explain the decline in cardiovascular deaths over the last 30 years. It is worthwhile reviewing the data in this particular study as it has been raised. Firstly this is modelling and therefore subject to bias and the confidence range seems to vary as much as threefold . But even if we accept the conclusion that 50% of the decline in deaths is due to the effect of “primary prevention”. This still leaves a large proportion unexplained by conventional risk factors. And looking more closely at the notion of “primary prevention”. Lets accept that smoking cessation is highly effective. Then we need to consider the decline in blood pressure since 1980 which was 7.7% ,however, this was due to vague “secular trends” not treatment with antihypertensives. The question is what caused the decline in blood pressure – an unknown factor? Likewise the decline in cholesterol was 4.2 % across the population – this decline again not due to Statin medication (which were only widely use in the later part of the 90s) – but a vague unknown dietary effect. So in this modelling the notion that medical intervention has resulted in major improvement cardiovascular death is high debatable and there are obviously other unknown processes are effecting this decline. You will also note that this reflects cardiovascular death and does not address the even greater decline in stroke deaths, for which there remains a deafening unaccountable silence. I maintain that the background incidence of vascular disease is declining largely independent of medical intervention and that “bananasplit effect” of decline in background incidence accounts for the inability of mass prescription policy of the QOF to show an effect after 5 years . Belief systems operate in medicine – we do as we are told and taught and doctors believe in conventional risk factors . But treatments that were widely used say in the 1930s would be ridiculous to modern medicine, yet no doubt were held with strength and conviction at the time . [1] Modelling the decline in coronary heart disease deaths in England and Wales, 1981-2000: comparing contributions from primary prevention and secondary prevention Belgin Unal, Julia Alison Critchley, and Simon Capewell BMJ 2005 331: 614 Competing interests: None declared |
|||
|
|
|||
|
Mark Struthers, GP Bedfordshire
Send response to journal:
|
Thank you, Des Spence. Thank you for another excellent article - for this other tasty piece on cookery books, the medical experts who cook them, and for expanding on the unwholesome expanse of current medical unknowns. It is rare for a doctor to admit that he hasn’t got the answers to the complexities of life – and to admit the truth that neither has medical science. Thank God too for Gary Taubes. Gary Taubes is neither a doctor, deluded, nor a scientist, but as the author of the ‘Diet Delusion’ he is an accomplished expert on cookery and how epidemiologists have cooked the books over what recipes make people healthy and what prescriptions might make them ill - or ultimately kill them. The book is an excellent all round critique of less than flavoursome epidemiological expertise gathered from around the world. I also recommend a review of epidemiological cookery published in the New York Times, entitled ‘Do we really know what makes us healthy?’ [1] In his September 2007 review, Taubes expands on the ‘post hoc, ergo propter hoc’ maxim - the confusion of sequence and consequence; the confusion of correlation and cause. Hypertension and the cholesterol hypothesis are simple recipes for lots of blood pressure pills and oodles of money-making statin drugs. And the simple statin is the corporate epidemiologist’s elixir of life for men, much as HRT was once the elixir of cardiovascular life for women. Like a fine wine, the truth about the elixir, the recipe, the prescription for a longer life is a so much more complex item on the menu - for the delectation of both men and women. But give me a cold beer: I suggest a glass of fizzy Fosters lager (for its antipodean origins) in reproach for the deeply topsy-turvy world we live in, man or woman. [1] Do We Really Know What Makes Us Healthy? Gary Taubes. Published: September 16, 2007 in the New York Times. http://www.nytimes.com/2007/09/16/magazine/16epidemiology- t.html?_r=2&pagewanted=all&oref=slogin Competing interests: None declared |
|||
|
|
|||
|
L Sam Lewis, GP Surgery, Newport, Pembrokeshire, SA42 0TJ
Send response to journal:
|
Dear Des, When competing hypotheses are equal in other respects, Occam's Razor recommends selection of the hypothesis that introduces the fewest assumptions and postulates the fewest entities while still sufficiently answering the question ! in http://www.heartstats.org/datapage.asp?id=722
The actual declines, and specifics you mention, are largely explainable by the well-researched Risk Factors which we already know, so why postulate the unknown ? Firstly, the rise to a peak in the 70s, then decline, can be largely explained by healthier lifestyles ( more exercise), less smoking, lowering of BP, then lowering of cholesterol, in serial over-laying effect. The relatively greater effect on Stroke over Heart deaths is what we expected from BP lowering. Your quoted reductions of 7.7% in population BP, and 4.4% in Cholesterol would be expected to produce exactly the substantial benefits ( circa 25% and 10% - see Figure from MRFIT group ) which you strive to attribute to 'unknowns'. Such reductions may be a composite of diet, exercise, or drugs.. but DRUG intervention is evidentially by far the most potent, but not the most pervasive. What you call the 'banana-split' effect - a law of diminishing returns - may account for the flattening of the decline in latter years. Essentially an expected 20% reduction of deaths by Statin is clearly smaller in absolute terms in populations already not smoking, exercising more, and taking BP-lowering agents. But it still adds up. The cumulative effect is just as we would expect from gradually widening the implementation of evidence-based interventions. As these effective interventions saturate all parts of the population then the decline would be expected to cease, and the curve become flat. Your 'unknown' factor that appeared in the 70s is not ineffable. I suggest it is nothing more than our combined medical Antithrombotic Attack ( BP lowering, betablockers, then Aspirin, Thrombolysis, Statin and finally CABG ). Each element has a significant mortality benefit proven in controlled trials, and has demonstrably been 'rolled-out' over the UK..
A more interesting 'unknown' is the nature of the "Framingham Gap" between predicted and observed outcomes in different populations. The interesting paper by Collins and Altman shows that QRISK corrects for missing factors, and closes this gap by 'calibrating to a UK population' ( with large numbers of substituted parameters). This does not help us to identify the extent of our ignorance, as measured by the Framingham gap. Smoking cessation, BP and cholesterol treatment are proven modifiers of initial risk, as are Aspirin and thrombolysis (at least in reducing death at a secondary prevention level). But does 'deprivation' work in ways other than those explained by smoking, diet, BP and cholesterol ? But how? The Regional variations map conceals deprivation and ethnicity correlates. Interestingly, Scotland much like the rest of the UK is closing the deprivation gap, seemingly through risk-factor management. The decline is flattening in the most deprived groups, because of rises in known risk factors, say the authors. This fits with the UK trends , whereby the flattening in recent years is largest amongst the under 45's. Come on Des ! The statistics are just possibly testimony to the proposition that we doctors are getting something right ! We need to sharpen these tools and use them. Sam Lewis GP, Surgery , Newport, Pembrokeshire REFS
http://www.heartstats.org/datapage.asp?id=722
Coronary heart disease mortality among young adults in Scotland in relation to social inequalities O'Flaherty et al. Published 14 July 2009,
doi:10.1136/bmj.b2613
Competing interests: None declared |
|||
|
|
|||
|
Mark Struthers, GP Bedfordshire mark.struthers@which.net
Send response to journal:
|
Dear Sam Smoking, drinking, eating, blood pressure and cholesterol are all simplistic hypotheses, razor sharp in their inability to cut through the whole scientific complexity of the longer life. But deprivation! How does that not work? How could deprivation be that known unknown responsible for the banana split effect that Des describes? What is stress? How do we recognise it? How do we measure it? Is it an illness? Is it a symptom of a greater unease? Is it a disease of deprivation? Is it a sign of the deprived? I don’t want to be deprived of another slice of my banoffee pie, before I die? Would you like the recipe, Sam? Competing interests: None declared |
|||
|
|
|||
|
L Sam Lewis, GP Surgery, Newport, Pembrokeshire, SA42 0TJ
Send response to journal:
|
Immortal, invisible, God only wise.. ? The ineffable mysteries of Life ??? No, no, no, that aint me babe ! I'm looking for the best-evidenced claims.. using the sharpest tools in the box. So much to do, and so little time. sam Competing interests: None declared |
|||
|
|
|||
|
Des Spence, Gp G20 9DR
Send response to journal:
|
I am very grateful to have this discussion. The truth , however , is that medication intervention in any real scale was not till the mid 1990s. Prior to this , for example the phrase was the “rules of halves” applied to hypertension – half known , half on treatment, half on treatment at target BP. So I struggle to accept that medicine and medication have made much impact ( other than importantly showing that smoking kills us ). The identifiable risk factors date back to what was measurable in the 1950s. I wonder if these days wrapping a rubber balloon around an arm and inflating would be accepted as a valid and reproducible cardiac investigation? The body is not simple and we understand only a fraction of the processes. Current risk factors are but the crudest and rudimentary measures of risk. The banana split effect is in essence that just as medical treatment was tacking off the background incidence was declining. This divergence I believe explains why now we see only modest effects of widespread medication. I assert that vascular disease is in decline largely independent of medical intervention. I lay down the gauntlet to public health and university epidemiologist to address this fundamental issue with some new and open skepticism .If not , we will see increasing medication and medilisation with next to no effect but with plenty of harm to society’s wellbeing. Competing interests: None declared |
|||