Have you ever stopped yourself in the middle of a patient encounter to think, have I been influenced by a recent adverse outcome, or a favorable interaction with a company representative, or persuasive prose in a journal report?
While clinicians use evidence to care for patients, human nature plays a big role, too.
Off the main stage at this year's American Heart Association (AHA) 2018 Scientific Sessions, two research groups presented important observations on the role of bias and spin in the practice of medicine. Until robots steal our jobs, studies that expose the brain's susceptibility to be influenced deserve attention.
Dollars Make a Difference
Amarnath Annapureddy, a postdoctoral fellow from Yale University in New Haven, Connecticut, and his team studied the effects of industry payments to cardiologists. Electrophysiologists are among the most highly compensated.
Their poster chronicled the association of industry payments and choice of implantable defibrillator.[1] Using the ICD Registry from the National Cardiovascular Data Registry , they included all patients who had an implantable cardioverter defibrillator (ICD) or cardiac resynchronization therapy-defibrillator (CRT-D) in a 24-month period (2014 to 2015) from any of the four major device manufacturers. Then they merged these data with the Open Payments program, which mandates public disclosure of payments from industry to doctors.
They identified almost 3500 physicians who received nearly $23 million dollars from the four companies (deidentified as companies A, B, C, and D). The median payment per physician was $1758 (mean, $6562). Physicians who received the highest payments from a particular manufacturer were more likely to implant devices made by that company.
The odds ratios told a clear story: Doctors in the group who took the most money from a particular manufacturer were 3 to 11 times more likely to use that company's device over another brand.
A dose-response relationship adds confidence in the findings: the higher the payment received from a given manufacturer, the more likely that doctor was to implant a device from that company. Some companies seemingly got more bang for their buck. For instance, doctors who received more than $10,000, from company A were 6 times more likely to use that device than if they received less than $100. Company D had an even stronger pull: doctors who received more than $10,000 from them were 37 times more likely to use their device.
Tricking Our Brains Through Prose
How the findings of a clinical trial are framed can also be subject to bias. Muhammad Shahzeb Khan, MD (John H. Stroger Jr Hospital of Cook County, Chicago, Illinois) and a team of researchers from other institutions studied[2] the prevalence of positive spin in cardiovascular research publications. They were inspired by similar analyses from oncology[3] and obstetrics/gynecology,[4] which have found high rates of spin in studies in those fields.
In short, spin is defined by manipulation of language to potentially mislead readers.[5] For example, when a primary endpoint is neutral, authors may emphasize significant results in secondary endpoints or subgroups or claim the treatment is beneficial without acknowledging that the trial was neutral.
My all-time favorite example of spin comes from the Lancet publication of the RITA-2 trial,[6] which compared percutaneous transluminal coronary angioplasty (PTCA) vs medical therapy for angina. The primary endpoint of death and myocardial infarction occurred in 6.3% of patients in the PTCA arm and 3.3% in the medical arm (P = .02). But instead of saying angioplasty was twice as bad as medical therapy, the abstract conclusion begins: "In patients with coronary artery disease considered suitable for either PTCA or medical care, early intervention with PTCA was associated with greater symptomatic improvement, especially in patients with more severe angina."
Using MEDLINE, Khan and his team searched for studies with neutral topline results from six major journals between 2015 and 2017. They identified 93 such randomized controlled trials. In these 93 papers, they found spin in 53 (57%) full texts and 62 (66.7%) abstracts. Each part of a paper could contain spin: They saw it in the results section in 38% of papers and the conclusions of 50%. Almost a third of papers had spin in all sections of the abstracts.
Comments
Don't look away from these findings.
Senior author of the open payments and cardiac device implant study, Jeptha Curtis, MD, from Yale University, wrote via email that he was "surprised by both the strength of the association and the observed variation across manufacturers. Clearly what may otherwise appear to be innocuous interactions with industry may have an outsized impact on practice."
It's hardly news to say money buys influence, but I've often heard clinicians say dollars don't affect them. "I choose devices or drugs based on what is best for my patients," goes the common proclamation. Well, evidence says otherwise, and we'd be wise as a profession to recognize this bias. What's more, do you think device companies would pay physicians if doing so had no influence?
Spin in medical research papers shows well the role of cognitive bias in both generating and consuming evidence.
Positive studies more often get published in bigger journals and bring more grant money. Incentives favor spin, but perhaps spin stems less from material gains and more from optimism. Scientists want their discoveries to work. We, the readers of the literature, also want our drugs and devices to work.
But medical breakthroughs come slowly. Reversals are common.[7] On Twitter, I've started the hashtag #EyesOpen to show the importance of reading medical evidence with an eye for spin. Khan and his team's work show that there's a lot of it out there.
Eyes open.
© 2018 WebMD, LLC
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this: John M. Mandrola. Think You're Immune From Bias? Think Again - Medscape - Nov 16, 2018.
Comments