News

Authors’ industry ties may bias neuraminidase inhibitors reviews

View on the News

Efficacy vs. effectiveness

Data and opinions are always colored by a variety of external factors, including financial relationships. Specialists who care for a specific condition are also at risk of wanting to believe a new therapy will make a difference in all patients and not parsing the data for limitations.

Good stewardship and astute clinical care require understanding what is effective. Unfortunately, published data usually reflect efficacy (the drug works in highly specified conditions) and not effectiveness (its impact when in generalized use). As a Medicaid medical director, I am engaged in a continuous search for effectiveness assessments and have to sift through a lot of assessments of uncertain validity.

Dr. William Golden is professor of medicine and public health at University of Arkansas, Little Rock, and medical director of Arkansas DHS/Medicaid.


 

References

In systematic reviews of research on neuraminidase inhibitors for either prevention or treatment of influenza, authors who had ties to drug manufacturers were much more likely to present the evidence as favorable and to recommend the drugs than were authors who had no such conflicts of interest, according to a report published online Oct. 6 in Annals of Internal Medicine.

“Reviewers with such conflicts of interest were also less likely to address issues with the underlying primary clinical evidence, such as publication bias and the lack of access to comprehensive study data,” said Adam G. Dunn, Ph.D., of the Centre for Health Informatics, University of New South Wales, Sydney, and his associates (Ann. Intern. Med. 2014 Oct.6 [doi:10.7326/M14-0933]).

The findings suggest that “further measures may be necessary to ensure that industry collaborations do not compromise the scientific evidence,” the investigators noted.

“If the benefits of neuraminidase inhibitors are eventually found to have been inflated, millions of patients will have been unnecessarily exposed to drugs that may be of little or no benefit,” the researchers added.

Dr. Dunn and his colleagues observed that reviews of neuraminidase inhibitors had widely divergent conclusions, with some strongly endorsing the agents as influenza prophylaxis or treatment, while others questioned the drugs’ safety and efficacy.

To examine whether industry influence may have contributed to these inconsistent conclusions, the investigators analyzed 26 systematic reviews. Thirteen assessed prophylaxis, and 24 assessed treatment; so, there were 37 distinct assessments. Seven of those were Cochrane analyses.

Two of the investigators read redacted copies of the reviews and rated them as favorable or unfavorable toward the use of neuraminidase inhibitors. Those copies showed no identifying information for the authors and no journal name or formatting clues. The agreement between the two investigators was “strong,” at 86%.

At the same time, Dr. Dunn and his associates searched for all the financial conflicts of interest for all the authors of these reviews, paying special attention to ties to the pharmaceutical companies that made the neuraminidase inhibitors under review. To do so, they went beyond the affiliations and funding listed in the systematic reviews, examining the authors’ personal and institutional websites, as well as disclosure lists provided by GlaxoSmithKline and Roche. They also performed Web searches combining the authors’ names, drug names, and manufacturers’ names.

Reviewers who had ties to the pharmaceutical industry rated neuraminidase inhibitors favorably in 88% (seven of eight) systematic reviews. In contrast, reviewers who had no such ties rated the drugs favorably in only 17% (5 of 29) of systematic reviews. That pattern persisted when prophylaxis and treatment were assessed separately.

Regarding prophylaxis, 100% (2 of 2) of the reviews with financial conflicts of interest were favorable, compared with only 9% (1 of 11) of those without conflicts of interest. Regarding treatment, 83% (5 of 6) of the reviews with financial conflicts of interest were favorable, compared with only 22% (4 of 18) of those without conflicts of interest.

In addition, reviewers who had ties to the pharmaceutical industry were less likely to include information about publication bias in their reports (14%, or only 1 of 7), compared with reviewers who had no such ties (79%, or 15 of 19), Dr. Dunn and his associates said.

The study findings indicate that financial conflicts of interest “are associated with product assessments favorable to the sponsors involved,” the authors asserted.

Factors that may contribute to biased conclusions include “the design of the review, the patient populations and outcomes assessed, the selective inclusion of primary evidence, [and] the critical appraisal of evidence quality and provenance,” the investigators noted. In addition, “the tone, emphasis, and interpretation provided by the authors may also influence the message that is conveyed,” they said.

The Australian National Health and Medical Research Council funded the study. Dr. Dunn and his associates reported having no financial conflicts of interest.

Recommended Reading

Fluoroquinolone bested other antibiotics in Haemophilus influenzae-related pneumonia
MDedge Family Medicine
Bill requiring childproofing of e-cig refills advances
MDedge Family Medicine
VIDEO: Resistant infection risk grows by 1% for each day of hospitalization
MDedge Family Medicine
VIDEO: Experts offer top tips for flu season 2014-2015
MDedge Family Medicine
E-cigarettes don’t help cancer patients quit smoking
MDedge Family Medicine
Canadian hospital’s Ebola scare exposes lack of readiness
MDedge Family Medicine
ACIP recommends PCV13 for all adults 65 and up
MDedge Family Medicine
AUDIO: CDC expert – What physicians need to know about enterovirus D68
MDedge Family Medicine
Top health officials say isolation is key Ebola strategy
MDedge Family Medicine
Prenatal BPA exposure linked to lung effects in 4-year-olds
MDedge Family Medicine