Late last year, news stories trumpeted the findings: Older women who received surgery to fix their cataracts were less likely to die over the course of a study period.
Now, the same group of authors is saying the exact opposite may be true.
Last week, the researchers retracted their 2017 paper in JAMA Ophthalmology and replaced it with an updated version that reports that women who received the surgery actually had a higher risk of dying.
The original paper was covered by many news outlets — including Reuters and the New York Times — some of which suggested the procedure may help with more than just vision (even though the study, by its nature, couldn’t determine whether or not surgery caused women to live longer). Annette Flanagin, the Executive Managing Editor for The JAMA Network, told us the publisher tried to get the word out about the significant change to the findings:
The Notice of Retraction and Replacement and a PDF of it was available to news media on the JAMA Network For the Media site 2 days before the embargo. Anyone who sees the article from an old news story and selects the link to the article will see the corrected version.
Nancy Lapid, the editor in charge of Reuters Health, told us they were going to update their 2017 news story, and hadn’t seen the retraction before we contacted them. Lapid said that the retraction was listed under the “Comment & Response” section of the JAMA Network, news release materials — a section that rarely contains anything of interest to the news service — instead of in the “Original Investigation” section.
“Association of Cataract Surgery With Mortality in Older Women” has been cited twice since it appeared online in November 2017, according to Clarivate Analytics’ Web of Science.
Retract and Replace is a tool many journals use — and two years ago, Flanagin spoke to us about how it might appear more in JAMA journals, in cases where the findings were affected by honest error, not misconduct. We asked Flanagin if the journal considered retracting the paper outright, given that the findings — and their potential implications — had changed so dramatically. She told us:
Our standard process for assessment of authors’ inadvertent pervasive error that affected study results was followed. The study was not invalid and was not a result of misconduct.
We contacted last author Anne Coleman at the University of California, Los Angeles, but haven’t heard back.
According to the detailed retraction notice, the authors were contacted by a reader with questions about the results. This prompted the authors to take a second look, when they realized:
…the time-to-event variable for the exposed group (women with cataract surgery) was mistakenly defined as time since the date of surgery rather than the date of diagnosis for this model.
The authors have now replaced the original version of the paper with a corrected one:
These results now indicate that cataract surgery was associated with a higher risk for all-cause mortality and mortality attributed to multiple causes of death, except for neurologic-related causes of death….We regret the errors in our definition of the time-to-event for the women with cataract surgery and in our incorrect analyses and any confusion this has caused the readers of our article…The abstract, text, and Tables 2, 3, and 5 in the original article have been corrected and replaced online.
Like Retraction Watch? You can make a tax-deductible contribution to support our growth, follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up for an email every time there’s a new post (look for the “follow” button at the lower right part of your screen), or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at [email protected].
make ’em give back all grant money with interest
Ralph, Ralph, your persistence is worthy a better cause.
Breaking news in science these days is all too often based on low level, initial, noisy and unreplicated findings.
Right there in the Reuters article
“The study wasn’t a controlled experiment designed to prove whether or how cataract surgery might help people live longer. It’s possible that other factors influence both risk of death and likelihood of getting cataract surgery, for example.”
So despite the huge error in time to event definition, this study should not have garnered so much attention.
News agencies need to update their science reporting standards. Fresh studies hot off the press should be clearly labeled as initial, exploratory, low level evidence findings. Only after independent high level replication should findings be given more weight than initial, exploratory, low level evidence findings.
Making a statement in the news article such as
“This research argues for ready access to cataract surgery for older women,” Smith said by email. “In a similar vein, this research argues for regular eye checks for older people.”
based on such low level evidence is inappropriate.
“However, several previous studies have linked cataract surgery to lower odds of premature death for both men and women, Coleman said.” Based on that statement, the authors saw what they expected to see. Now with the re-published paper, it appears those several previous studies should be given closer scrutiny.
Did the Reuters team track down these papers and assess their level of evidence?
I just did. The retracted paper lists three previous studies. The first is from this same team (a red flag as far as establishing high level independent replication)
Tseng VL, Yu F, Lum F, Coleman AL. “Cataract surgery and mortality in the United States Medicare population.” Ophthalmology. 2016;123(5):1019-1026.
which states
“Previous studies have examined the association between cataract surgery and mortality with conflicting findings.” not exactly the statement Coleman provided to Reuters. Reuters reporters should have noticed this.
Further,
“Only 2 existing studies have examined the association between cataract surgery and mortality in a cohort of patients entirely with cataract, and both of these studies found that cataract surgery was associated with decreased mortality. However, both of these studies were conducted in isolated voluntary cohorts within Western Sydney, Australia, and it is unknown whether their findings are generalizable to the United States population.”
so the other two are from one team in Australia, not using high level assessment strategies (the highest being randomized clinical trials).
In the retraction notice the authors describe a time length bias – the people who end up receiving surgery had to survive the period from cataract diagnosis up to surgery date. Any patients who died before their surgery date would automatically end up in the other group, making that group look worse off.
This is why randomized clinical trials are so important. A clear evaluation of this issue would require randomizing patients to cataract surgery versus no cataract surgery and the survival analysis would compare those two groups. The patients randomized to the surgery arm would include some patients who died before receiving their surgery, which would balance out similar cases in the no surgery group.
This kind of flip-flopping in science news, especially health care news, is really not what we need right now, and not what we need in general.