Issue 46 / 2 December 2013

THE flood of one-off observational studies and “sensational” reporting that crosses a GP’s desk can be “cluttering” and “distracting” for the doctor and unhelpful for patients as a result, say Australian experts.

Dr Janette Randall, chair of NPS MedicineWise and a Brisbane-based GP, told MJA InSight that, while patients had much to gain from discussions of all options with their doctor, it was a constant battle to know what was relevant and what was not.

“It has become really tricky”, Dr Randall said. “There is a plethora of research and a lot of it is not relevant.”

She was responding to an observational study published in JAMA which, contrary to an earlier report, found no significant association between oral fluoroquinolone use and retinal detachment.  The 4-year Danish nationwide, register-based cohort study included 748 792 episodes of fluoroquinolone use and 5 520 446 control episodes of non-use, with 566 cases of retinal detachment. (1)

The findings were in contrast to a case–control study of patients visiting ophthalmologists in Canada, which matched 4384 cases of retinal detachment with 43 840 controls. That study found that patients taking oral fluoroquinolones were at a higher risk of developing a retinal detachment compared with non-users, although the absolute risk was small (number needed to harm = 2500) (2)

An accompanying editorial said it was not possible to be certain which of the two studies was more accurate but that doctors who were aware of both studies would likely have a “here we go again” reaction. (3)

“[This is an] all-too-common situation in which results of several observational studies conflict, leaving a quandary about implications for practice”, the editorial said.

“Influential medical journals are trusted by physicians and are ‘go-to’ sources for the public media. A trusted source of information has a corresponding responsibility to present the strengths and weaknesses of clinical research as clearly as possible, so that clinicians are not left hanging.”

Dr Randall said she would like to see academic journals become more discerning about the kind of research they published.

“This kind of research can be cluttering and distracting”, she said. “Journals need to be more selective about what they publish and we need to be more selective about what kinds of trials get funded.

“The question needs to be asked — is this really important?”

The JAMA editorialist agreed that the lack of time GPs had for in-depth research shifted the burden to journals.

“Journals should have a low threshold for publishing understandable and unbiased editorial commentary when observational studies generate unexpected provocative results or conflict with each other.”

Dr Randall was also critical of the media.

“This business of focusing on relative risk rather than absolute risk is a case in point”, she said. “Relative risk numbers can be scary, and the way this stuff is reported can be sensational.

“It is the job of journals and [organisations like] NPS MedicineWise to be balanced and to make sense of it.”

Professor Rob Moulds, medical adviser with Therapeutic Guidelines and one of the expert panellists who compiled the Antibiotic Guidelines, said the abundance of research “leaves GPs in a difficult situation”.

“The take-away message is conservatism”, he told MJA InSight.

“GPs must be very careful of being overly influenced by single studies. Data needs replication, and ancillary evidence to support.

“We need to be very careful of research that takes a leap.”

1. JAMA 2013; 310: 2184-2190
2. JAMA 2012; 307: 1414-1419
3. JAMA 2013; 310: 2151-2153

2 thoughts on “Quandary of conflicting studies

  1. Edward Brentnall says:

    My father who qualified in 1910 always warned me; “Never be the first or the last to use a new drug”.  Nothing much has changed in that wisdom.

     

  2. John Urlwin says:

    Don’t you love pharma’s dogged persistence in reporting benefits in relative terms but adverse reactions in absolute terms. The role of NPS deserves ongoing support to help us sieve through the forest of information.

Leave a Reply

Your email address will not be published.