The Problem of Evidence

Barbara Stanley FRCA

With thanks to Dr Jan Friedrich

Aims

• To categorise and illustrate some of the published arguments against EBM

• Demonstrate why published trials don’t necessarily result in reliable evidence

• To try and examine the impact EBM outside of everyday practice

• (almost) Everything I am going to say has been published somewhere in a journal!

Definitions

• Evidence ‘ that which furnishes or tends to furnish proof’

Merriam-Websters Dictionary of Law

• Evidence based medicine ‘The judicious use of the best current evidence in making decisions about the care of individual patients’

• Evidence-based practice ‘ integration of best research evidence with clinical expertise and patient values’

Strengths of EBM

• Prevent unnecessary wide variations in clinical practice

• Reduce the use of unproven interventions

• Produce consistent practice guidelines

• Rid medicine of dogmatic practice and concepts that are incorrect or outdated

Origin of EBM

• Epidemioligists of McMaster University in 1993

• Archibald Cochrane 1909-1988 – Professor of Chest Disease and Honorary Director of MRC’s Epidemiology Unit, Cardiff

• Poor use of scientific method in Medicine

Origins of EBM

• “It is surely a great criticism of our profession that we have not organised a critical summary...of all relavant randomised controlled trials”

Effectiveness and Efficiency: Random Reflections on Health Services 1972

• 10% therapies proven – rest may even harm• Promoted RCT as best way to prove

effectiveness• Promoted use healthcare resources to

‘maximise the delivery of effective intervention’A categorisation and analysis of the criticisms of EBM. Cohen et al. Int J Med Inform 2004

Use of EBM

• ‘Regulatory’- rationalise practice at level of health service delivery and control expenditure

• ‘Bedside’- treatments with the best evidence offered to individual patients

Categorisation of criticisms

1. Aplicability

2. Effects on proffession

3. Workload for clinicians

4. Interpretation

Common Criticisms 1: Applicability

• Medicine based on evidence, rather than Evidence Based Medicine

• ‘Basic error’ – biological variablity hinders the extrapolation of data obtained from basic lab or clinical research to the treatment of individual patients

• Applies only to that which can be objectively measured

• ‘Orphan specialties’ • Does lack of evidence mean an intervention

doesn’t work?• EBM itself has never been validated

Category 2: Effects on the Profession

• Promotes ‘doing’ rather than ‘thinking’ and promotes ‘mindless algorithms’ and ‘protocols which act as straightjackets to care’

Ethics and EBM. Loewy. 2007

• EBM protocols tend to be seen as ‘the right way’ instead of the ‘current’ way – ‘insists rather than assists’

• Skew toward pharmacological interventions and treatments rather than preventative measures or complex social interventions leading to the expectation of a ‘pill for every ill’

Category 3: Workload

• The sheer volume of it! Two MILLION articles published annually in 23,000 biomedical journals

Evidence-Based Medicine for the New Millenium:Critical Appraisal and Pragmatic Approach.Ibrahim and Stuart. Ann Saudi Med.1999

• Trials beget more trials to confirm or refute results

• Sorting ‘good’ and ‘robust’ from...’poor’ and ‘weak’?

Workload

• Increased expectation of clinicians and trainees contribution to evidence base – increasing numbers of poorly designed trials unlikely to make useful contribution – yet appear in peer reviewed journals

• ‘...busy clinicians who devote their scarce

reading time to selective, efficient, patient driven searching, appraisal, and incorporation of the best available evidence can practice evidence based medicine.’

D. Sackett. EBM:what it is and what it isn’t. BMJ 1996

Category 4:Interpretation

• Publication bias – ‘Studies with positive findings are more likely to be published than studies with negative or null results’ AMA Trustee Joseph M. Heyman 2004

• ‘Any tendency to put negative results in a drawer ..can bias reviews of treatments reported in the medical literature, making them look more effective than they really are.’ AMA Trustee Joseph M. Heyman 2004

Interpretation

• Metanalysis may amplify errors and bias of the trials examined

• Similarly designed RCT’s researching same question often disagree

Cohen et al. 2004 A categorization and analysis of criticisms EBM

• ‘Subgroup analysis... Influence conclusions much more than is justified’Subroup analysis and other (mis)use of baseline data in clinical trials. Aassmann et al.Lancet 2000;355:1064-69

• "If you find that [a] study was not randomized we'd suggest that you stop reading it and go on to the next article EBM turns patients into pets. L Gorman. Independence Institute. 2004

Industry and EBM

• Declaration Helsinki – ‘...investigators are obliged to preserve the accuracy of results.Negative as well as positive results should be published... Any possible conflict of interest should be declared in the publication..

• ‘Studies show that research funded by pharmeceutical and food manufacturers are likely to report results beneficial to the industry......’

‘Does Industry support bias resaerch? Chimonas. Chen et al American Sociological Assoc 2007..

• Negative findings were found in 13% of industry-funded trials vs 35% of non-industry funded

Yaphe, Edman et al Family Practice 2001

Conflict of interest in EBM: Vioxx

• VIGOR study 2000 NEJM showed 4 x increase risk MI in Vioxx vs. Naproxen

• Merck scientists attributed this to protective effects Naproxen (not validated)

• Scandal centres around communication between researchers and Merck executives regarding increased CV risk and failure of Merck to disclose this to FDA as early as 1996 – uncovered by The Wall Street Journal

• APPROVe study 2004 Demonstrated unequivocal increase in risk MI – Merck withdrew Vioxx

The price of Vioxx

• 27,000 law suits• $4.85 billion Vioxx legal expenses• Merck’s ultimate liability estimated to be $10 -$25 billion

Merck agrees to settle Vioxx suits for $4.5 billion. Berenson. NY Times Review 2007

• Study quoted ‘ghostwriting’ (failure acknowledge as author a contributor to a manuscript) in 13% research articles, 10% review articles, 6% editorials and 11% Cochrane reviews: Guest authorship (designation of individual as author who does not meet criteria) identified in 16% research articles, 26% review articles, 21% editorials and 41% Cochrane reviews

Guest Authorship and Ghostwriting in publications related to Rofecoxib. JAMA 2008

• "Between 1997 and 2004," the pharmaceutical industry critic Sheldon Krimsky noted in a 2005 open piece in the Newark Star-Ledger, "12 major prescription drugs, with a market value of billions of dollars, were recalled by the FDA." Krimsky claims such dangerous drugs have been allowed to reach the market because "conflicts of interest have become endemic in the system of drug evaluation," a trend that "has been exacerbated by the rise of for-profit clinical trials, fast-tracking drug approvals, government-industry partnerships, direct consumer advertising and industry-funded salaries for FDA regulators" since the mid-1990S.

Evidence and the Media

A History of MMR

• Introduced in 1988

• Fall in mumps by 79% over subsequent 2 years and further – to a 92% total reduction

Monovalent mumps vaccine or combined MMR vs. placebo or no vaccine. Ellman et al. BMJ Feb 2007

• 1992- withdrawal of early version – mumps component associated with meningitis

MMR and Autism

• Media and public furore• Wakefield paid by solicitor (£435,643) of parents

litigating for alleged injury as a result of MMR• All 12 children specifically referred to

substantiate lawsuit• Ethics committee led to believe tests were

routine and would be performed in absence of study

• Wakefield lied about behavioural symptom onset and vaccination timing

• One child suffered 12 bowel perforationsThe MMR-autism scare – our story so far. Brian Deer

MMR and Autism

• Found to have constipation• “increased vascularity and irregularities entirely

subjective”• Ileo-lymphoid hyperplasia is common, benign

and not linked to inflammatory bowel• Wakefields own tests did not find MMR virus in

childrens bowel• Wakefield patented a monovalent measles

vaccine as competitor to MMR – as well as ‘cure’ for IBD and autism

The MMR-autism scare – our story so far. Brian Deer

MMR and Autism

• Lancet published apology and Wakefield’s co-authors a retraction

• In my view, if we had known the conflict of interest Dr Wakefield had in this work I think that would have strongly affected the peer reviewers about the credibility of this work and in my judgement it would have been rejected."

Richard Horton, Editor The Lancet Feb 2004

• 23 subsequent studies that refute linkMMR vaccine does not cause autism. Examine the

Evidence! www.immunize.org/catg.d/p4026.pdf

Subsequent research

• Finland – 3 million doses, no reports autism• UK- Autism incidence rising but no ‘step-up’ since MMR

Bandolier database: ‘No Measles in Finland’ Lancet 1997

• Denmark- no increase in autism with MMR in 537,303 children studied, 440,665 vaccinated

A population based study of MMR and autism. Madsen et al. NEJM 2002 ,7;347 ,1477-82

• Japan – Incidence autism continued to rise after MMR was discontinued

MMR-vaccine and regression in autism:negative results presented from Japan. Uchiayma et al. J Aut Devel Disord. 2007.37:210-17

Measles resurgence

• Incidence of measles risen. 1,000 cases in UK 2007 vs. 449 in 2006 vs. 77 in 2005

• 200 cases SE London Jan-May 2008• Including 1 fatality• Herd immunity requires 95% uptake

BBC News: Emergency Measles steps ordered: May 2008• Average uptake 88.6% for MMR and 5.2% for at least 1

single antigen vaccineMillennium Cohort Study Child Health Grp Factors

associated with uptake of MMR in a contemporary UK cohort: prospective cohort study. Pearce A et al. BMJ 2008. 336, 754-7

• Uptake rates only 71% first dose and 50% secondHealth Protection Agency

When Evidence divides us-the case of Activated Protein C

• PROWESS – phase III multicentre DBPCRCT of 1,690 patients with sepsis.

• Stopped early by independent board as objectives met

• Absolute and relative risk-reduction for death were 6.1% and 19.4%. Mortality rate 30.8% placebo vs 24.7% APC

• Multiple subgroup analyses performed- for which PROWESS was not poweredRecombinant APC:the key is clinical risk of death not subset analysis. Dellinger. Crit Care 2006

When Evidence divides us-the case of Activated Protein C

• Criticisms of the PROWESS trial include:– A change in protocol during the trial – pts with

metastatic ca or pancreatitis or organ failure >24 hrs excluded–

– The test drug was manufactured from a different cell line

– A change in placebo drug to albumin 0.1% from saline Risks and benefits Activated Protein C for Severe Sepsis. Warren et al. NEJM 2002

Controversy

– FDA analysis indicated an improvement of efficacy pre protocol amendment mortality was 28% APC vs 30% placebo, after 31% placebo vs 22% APC

– Conflict over the change in protocol and the different source split FDA Advisory Committee 10 to 10 as to the safety and efficacy of the drug Risks and benefits Activated Protein C for Severe Sepsis. Warren et al. NEJM 2002

Suspicious facts brought to light

• Adequacy of Blinding – obviously different fluids were ‘wrapped’ in plastic and some sites couldn’t use albumin as placebo

• Further sites added after protocol ammendment – supposedly to incraese pt recruitmet but were noted to;– Show a strong effect for the drug– Had performed well in a previous Lilley –funded drug

study– Weren’t recruiting any more than previous sites that

had been droppedConflict of opinion-is PROWESS real progress? AF Mackenzie.ICM 2006,32;610-12

Not as robust as we thought?

• DNR order rates were lower in APC group after protocol change 9% vs 17% in placebo group– end point was 28 day mortality

• A change in outcome of only 14 patients would remove the statistical significance of PROWESS mortality improvement

• Home discharge rates were similar in placebo and APC groups

• Patients lost to follow-up at 3 months, making survival stats at 1 and 2.5 yrs questionable

Conflict of opinion-is PROWESS real progress? AF Mackenzie.ICM 2006,32;610-12

And....

• Calculation of APACHE II score varies between clinicians up to 20% - PROWESS demonstrated no benefit in patients with scores <20 and increased risk of bleedingRisks and benefits Activated Protein C for Severe Sepsis. Warren et al. NEJM 2002

• PROWESS was stopped early – Disproportional treatment effect?

But the supporters rallied;

• Subsequent arguments claimed that;– APC from new cell line had no difference in

vitro– If indeed it was more effective than the

previous batch then a better product is available

– Protocol ammendments actually ended up excluding subsets of patients more likely to benefitAssessing the use of APC in the treatment of severe sepsis.JP Seigel. NEJM 2002

More Confusion

– Confirmatoy studies ‘ ‘ENHANCE’– ENHANCE – single arm unblinded study to

gather more safety and efficacy data

– mortality reduction rates similar but bleeding rates higher 3.6% vs 2.4% POWESS

– Showed better mortality reduction if given <24hours of organ dysfunctionClinical trials in severe sepsis with drotrecogen alpha (activated). P-F Laterre: Crit Care 2007:S5

– ADDRESS looked at single organ failure or APACHE<25 but was

• Stopped early in accordance with the futility guidelines-chance of meeting the defined objective was <5%

• Mortality rates in APACHE II scores of >25 higher in drotaa group 29.5%compared to placebo 24.7%(subgroup analysis)

• AND mortality rates in APACHE II >25 placebo group were 24.7% vs 43.7% in PROWESS

• Conclusion – sample size too small to detect difference – only 2,640 enrolled

Clinical trials in severe sepsis with drotrecogen alpha (activated). P-F Laterre: Crit Care 2007:S5

• Commentators argue over definition of high risk of death. This not clearly defined in PROWESS except via APACHE II score, included multi and single organ failure

• ‘Seasoned clinician’s clinical assessment of high risk of death’Recombinant APC:the key is clinical risk of death not subset analysis. Dellinger. Crit Care 2006

• Reviewers of evidence have financial links to Eli Lilley (authors of above cited papers)

Post Marketing Surveys

• Higher mortality rates than PROWESS or ENHANCE overall – 45%

• Attributed to a sicker poulation and later use of APC

• Higher bleeding events 7.3%, again attributed to sicker population

• ‘study underpowered to adequately assess the risk-benefit ratio for DAA...’Evaluating the use of Drotrecogen Alpha in adult severe sepsis:a Canadian multicentre observational study. Kanji et al Intensive Care Med 2007, 33:517-523

Cochrane Review ab 004388

• ‘There is insufficient evidence to support the use of human recombinant activated protein C for adults or children with severe sepsis; moreover there is an increased risk of internal bleding associated with it’s use.’

Cochrane Reviews. November 2007

Other ICU Evidence controversies-Tight Glycaemic Control

• Tight glycaemic control in cardiac surgery ICU patients significantly reduced mortality, length stay and infection.

• Mortality benefits may persist up to 4 years later with tight glycaemic control for 3 daysStrict Blood glucose control during ICU after cardiac surgery: Impact on 4 yr survival. Ingels et al.Eur Heart J 2006;27,2716-24

• Morbidity benefit seen in medical ICU poulation with stay >3 days,but no mortality improvement.

Other ICU Evidence controversies-Tight Glycaemic Control

• Outcome worse in patients in ICU <3 days who could not be identified before therapy.

• Logistic regression analysis identified hypoglycaemia as an independent risk factor for death.

Intensive Insulin Therapy in the Medical ICU. Van de Berghe et al. NEJM 2006;354;449-61

VISEP study

• Multicentre trial severe sepsis for intensive insulin therapy(4.6-6.1mmol/l) vs conventional (glucose 11.1mmol/l)

• Intensive insulin therapy stopped early for safety as hypoglycaemia occured 12.1% intensive insulin grp vs 2.1%

• Mortality rates no different at 28 daysIntensive Insulin Therapy and Pentastarch Resus in Severe Sepsis. Brnkhorst et al. NEJM. 2008;

358;125-39

• Intensive insulin not indepedant risk factor for death BUT hypoglycaemia was

• Intensive insulin group showed tendancy to stay longer in ICU

• No morbidity reduction identified• Assumption that second intervention (HES

vs Ringers) would not influence resultsIntensive Insulin Therapy and Pentastarch Resus in Severe Sepsis. Burnkhorst et al. NEJM. 2008;

358;125-39

Glucontrol Trial

• 1100 pts – 4.4 – 6.1mmol/l vs 7.8 – 10mmol/l. • No survival benefit;• Episodes hypoglycaemia more frequent• Mortality rates higher in those with at least 1

hypoglycaemic episode• Tight control = four – six times increase

incidence hypoglycaemiaControversies about tight glucose control. Devos and Presier. Expert Rev. Endocrinol. Metab;3;295-97 (2008)

Summary

• Control of blood sugar desirable and confers a survival advantage- absolute mortality reduction of 4.83%

• Severe Hypoglycaemia increases mortality risk – even a single episode

• Tight glycaemic control associated with up to six times increased risk severe hypoglycaemia

• What range to aim for?Severe hypoglycaemia in critically ill patients:Risk factors and outcomes.Krinsley. Cri Care Med.2007;35,2262-68

Controversies in Evidence in ICU: Albumin

• Cochrane collaboration – albumin increases mortality in trauma patients

• Criticised as it omitted several studies• SAFE study – no difference in mortality overall

A comparison of Albumin and Saline for fluid resusin the ICU. SAFE study. NEJM 2004;350,2247-56

• Subsequent subgroup analysis in traumatic brain injury = more deaths in albumin group 25% vs 15%

Controversies in Evidence in ICU –Albumin

• Once head injured excluded, overall mortality in trauma group with albumin was still higher (14% vs 10% p-0.06).(Authors of SAFE conclude by chance)

• Excluded burns, hypoalbuminaemia, liver failure and cardiac surgery.

Fluid resus among the critically ill: more water under the bridge. M. Jaka. CJA 2006,53:1258-59

.

Controversies in Evidence in ICU –Albumin

• SOAP study: observational of use albumin accross European ICU’s. Those receiving albumin had cancer or cirrhosis, were surgical admissions and had higher SAPS II and SOFA scores with a longer ICU stay

• When these factors were corrected for using both the Cox proportional hazard model and a propensity analysis, albumin use was still an independent predictor of lower 30 day survival.

SOAP Study.Crit Care 2005

Summary

• ICU and hospital mortality rates higher in pts who received albumin

Is albumin administration in the acutely ill associated with increased mortality? Results of the SOAP study. Vincent et al. Crit Care. 2005,8;R745-754

• Conclusion: not for head injured patients, or trauma....or ICU patients

Trials stopped early

• Tend to show large treatment benefit – Bisoprolol trial showed treatment effect of 0.09 (prevention of cardiac death/non fatal MI) in patients with positive Dobutamine echo undergoing elective vascular surgery and is inconsistent with the effect of beta blockers in thousands of post MI patients – 0.65 – 0.85

• Trial short vs longterm Warfarin on thromboembolic complications post hip arthroplasty showed large treatment benefit. Once result adjusted for full sample size no benefit seen

Randomized Trials Stopped Early for Benefit: A systematic Review. Montori et al. JAMA 2005;294(17)2203-09

Trials stopped early

• Are increasing in frequency – 0.008% in 1980 – 84 vs 0.1% in 2000-2004

• Tend to be reported in high impact factor journals - 0.1% in 1980 – 84 vs 1.2% in 2000-2004

Randomized Trials Stopped Early for Benefit: A systematic Review. Montori et al. JAMA 2005;294(17)2203-09

The Law and EBM

• Perhaps the real fear behind critcisms of EBM• Evidence-based medicine (EBM) is the

integration of best research evidence with clinical expertise and patient values

• Medical malpractice currently centres around the ‘Bolam’ principle – a doctor is not liable if his practice follows a responsible body of medical opinion

Medical Law and Ethics. SE Salako. BMJ 2003

A move away from Bolam

• Bolitho vs City and Hackney Health Authority (1998) child with repeated episodes respiratory distress eventually died after failure to be attended by duty doctor. Issue became should she have intubated during earlier episodes and would the child have survived. Expert opinion was divided but majority would have intubated earlier than the arrest. Judge ruled in favor minority opinion because it was reasonable – ie against majority of expert testimony

Bye-Bye Bolam. Mbrazier. Med Law Review 2000

The Law and EBM

• Expert testimony under increased scrutiny –’Where clinical guidelines have been developed....the Judge will be enabled to judge the individual expert testimony in context of the professions’ considered judgement’

Bye-Bye Bolam. Mbrazier. Med Law Review 2000

• Commentators fear that doctors, to avoid litigation, will cease to treat the individual and concentrate on ticking the relevent boxes

Guidelines and negligence

• Proportion guidelines fall short of quality markers – courts do not call experts in guideline methodology to assess robustness and quality

How does evidence based guidance influence determinations of medical negligence? Hurwitz. BMJ 2004

• GMC – doctors should “normally follow guidelines”

• Guidelines may not protect doctors. Helling vs Carey(1974): occular pressure not measured in pt <40 for glaucoma as not standard practice. “ opthalmologists prosecuted

EBM and the Law. C. Williams 2004

Practice Outside Guidelines and keeping up to date

• Practice outside guidelines lawful: Cranley vs Medical Board West Australia (1990)GP gave i.v Diazepam to heroin users against Australia’s Methodone guidelines, but small body opinion concluded this was acceptable practice

How does evidence based guidance influence determinations of medical negligence? Hurwitz. BMJ 2004

• Burton vs Brooklyn doctors hospital. Doctors found liable for causing blindness in prem baby by liberal 02 administration when evidence suggested it was harmful

EBM and the Law. C. Williams 2004

Litigation for Practice within guidelines

• Merenstein case(2003) PSA not checked in patient <50yrs in accordance with guidelines and agreement with patient, subsequently found to have Gleason 8 tumour. $1 million damages from Family Practice Residency Programme despite informed patient consent and adherence to guidelines

Rethinking informed consent: the case for shared medical decision making. King, Moulton. Am J Law Med. 2006

Summary

• Precedent setting cases which– Rule in favour of minority opinion

– Rule against doctors who practice within guidelines

– Rule against doctors who don’t keep up to date

– Beginning to question ‘expert testimony’ and Bolam principle

EBM and Public Opinion

• There is an over reliance on RCT’s that retards innovation

• In 2002 $125 million ALLHAT trial (antihypertensive trial sponsored by National Heart,Lung and Blood Institute) concluded diuretics should be 1st line despite subsequent criticism of failure to achieve adequte control and bias of the trial in diuretics favour , use of primitive drug therapies and the observation that diuretics increase all-cause mortality in Caucasian men by 11%

EBM turns patients into pets. L Gorman. Independence Institute. 2004

• The problem with evidence-based medicine is that its proponents want everyone to pretend that high-priced literature searches can substitute for the informed judgment that is the foundation of science-based medicine

• Had law given evidence-based masters dominion over specialists in treating high blood pressure, the millions of Americans who don't fit the ALLHAT profile would have been subjected to cheap treatments known to produce poorer outcomes. In short, when the law finally requires evidence-based medicine, you'll know that medicine is going to the dogs. EBM turns patients into pets. L Gorman. Independence Institute. 2004

On prosecution despite adherence to guidelines…..

• ..He defined EBM as a cost saving method and stated his belief that the few lives saved were not worth the money. He urged the jury to return the verdict to teach residencies not to send any more residents …believing in EBM

• The jury sent a message to the residency programme that they didn’t believe in Evidence Based Medicine. They also sent a message saying they didn’t believe in national guidelines and they didn’t trust the shared decision making model

Winners and Losers. Merenstein. JAMA 2004

What can we conclude?

• Look for transparency of method and measurements in publications

• Be aware of conflicts of interest• Don’t change your practice immediately• No-one can have an exhaustive

knowledge of evidence• The law may be changing• Therapies based on evidence and sound

scientific and physiological principles

And Finally...

• ‘There’s no Evidence for that!’

• “The art of medicine consists in amusing the patient while nature cures the disease”

Francois Marie Arouet (Voltaire)

References

• ‘Adrenal Insufficiency and Steroids’: Andy Walden PhD, MRCP: AnaesthesiaUK

• ‘SAFE,VASST,LIPOS Trial 3, CORTICUS and more; Implications for the Surviving Sepsis Campaign Guidelines’ Gregory S. Martin MD, MSc: Medscape Critical Care: Highlights of the Society of Critical Care Medicine 36th Critical Care Congress, Feb 2007

• ‘Fluid resuscitation among the critically ill: more water under the bridge’ M J Jacka MD MSc FRCPC, D Alberton BSc (PHARM), R T Noel Gibney, MB FRCPC: Can J Anesth 2006/53:12/1258-9

References

• ‘Conflicting clinical trial data: a lesson from albumin’ Greg Martin: Critical Care 2005, 9:649-650

• Intensive Insulin Therapy in the Medical ICU. Greet Van den Berghe, M.D, PhD., et al NEJM 2006;354:449-61

• Severe hypoglycaemia in critically ill patients: Risk factors and outcomes. Krinsley, J. F et al: Critical Care Medicine, 2007

• Warning signs:E-mails suggest Merck knew Vioxx’s Dangers at Early Stage. Nov 1 2004 The Wall Street Journal

• The consequences and Lessons of the Vioxx Recall. M Etzler. 2007

References

• Clinical trials in severe sepsis with drotrecogin alpha (activated). Pierre-francoise laterre. Critical care 2007, 11(suppl 5):s5

• Recombinant Activated Protein C: the key is clinical assessment of risk of death, not subset analysis. Phillip Dellinger. Critical Care 2006, 10:114

• What do evidence based secondary journals tell us about the publication of clinically important articles in primary healthcare journals? K A McKibbon et al. BMC Medicine 2004,2:33

• Guest Authorship and Ghost writing in Publications related to Rofecoxib. Ross, Hill et al JAMA 2008;299(15) 1800-1812

References

• Journal reading habits of Internists Sanjay Saint et al. J Gen Intern Med. 200 December, 15(12):881-884

• Evidence-Based Medicine in the Law Beyond Clinical Practice Guidelines: What Effect Will EBM Have on the Standard of Care? Carter L. Williams Washington and Lee Law Review Winter 2004

• Ethics and Evidence-Based Medicine: Is There a Conflict? Erich H Loewy, MD Medscape Bioethics 2007

• Ethics and Evidence-Based Medicine Simon R Tomlinson MJA Feb 2002 vol 176:137

• Winners and Losers. D Merenstein: JAMA 2004,291;15-16`

References

• Risks And Benefits of Activated Protein C Treatment for Severe Sepsis. H. Shaw Warren, MD, Anthony F. Suffrendi MD, N Engl J Med, vol 347, no. 13, September 2002

• Randomised Trials Stopped Early For Benefit: A Systematic Review. Victor M. Montori et al. JAMA 2005:294(17):2203-2209

• Is Statistical Significance Always Significant? Ronald L. Koretz, MD Nutrition in Clinical Practice 20:393-307, June 2005

• Parachute use to prevent death and major trauma related to gravitational challenge:a systematic review of randomised controlled trials. Gordan C. Smith and Jill Pell. BMJ 2003;327;1459-1461

• False dichotomies:EBM, clinical freedom and tha art of medicine. M. Parker. J Med Ethics; Medical Humanities 2005;31:23-30

• A categorisation and analysis of the criticisms of Evidence-Based Medicine. Aaron Michael Cohen et al. Intl J Medical Informatics (2004) 73, 35-43

References

• Subgroup analysis and other (mis) uses of baseline data in clinical trials. Susan F. Assamann et al.Lancet 2000:355:1064-69

• The association between funding by commercial interests and study outcome in randomised controlled trials. John Yaphe et al. Family Practice:vol 18 no 6:2001

• Seven alternatives to evidence based medicine. David Isaacs,Dominic Fitzgerald. BMJ 1999;319:1618

• Evidence based medicine: what it is and what it isn’t.D. Sackett, W Rosenberg, J A Muir-Gray, R B Haynes, W S Richardson. BMJ 1996;312:71-72

References

• Issues in Comparisons between Meta-analyses and Large Trials. J.P.A Ioannidis, J.C.Cappellerri. J.Lau. JAMA. 1998;279(14):1089-1093

• What kind of evidence is it that Evidence based medicine advocates want health care providers and consumers to pay attention to? RB Haynes. BMC Health Services Review 2002

• Rethinking informed consent:The case for shared medical decision making. JS King,B Moulton. Am J Law Medicine. 32(2006):429-501

• Point/Counterpoint editorials: EBM lacks a sound scientific base.M. J Tobin. Chest 2008:133 1067 - 1080