"Prediction is difficult, especially about the future,” physicist Niels Bohr said, sounding quite a bit like Yogi Berra. Every doctor knows that all too well. There’s no guarantee that the clinical decisions you make will yield the results you expect.
When selecting a medication, how sure can you be that the outcome you anticipate will actually come to pass? Maybe the patient is a nonresponder. Or maybe your diagnostic tests are vulnerable to false positives and negatives.
Even well-funded, high-profile health issues can get waylaid by incorrect or incomplete data. A sobering article in the April 25th New York Times Magazine describes a recent reconsideration of the clinical value of routine breast cancer screening. After reviewing three decades of outcomes data, cancer experts reached an uncomfortable conclusion: mammography is more likely to yield false positives that lead to unnecessary surgery than it is to find true early-stage cancers that warrant intervention. Should doctors and advocacy groups now back-pedal from the accepted narrative that early screening is essential? Could they, even if they wanted to, now that the practice is so well entrenched?
With these thorny questions in mind, we turn our attention to the world of ocular research, front and center in this issue as we again review some of the most novel, thought-provoking studies presented earlier this month at ARVO.
The breadth of the subjects on display every year is truly amazing. It’s a testament to the tenacity of the research community that so many lines of inquiry are pursued. But how can a busy clinician assimilate so much new information? Which studies are conclusive and which are more speculative?
The biggest news to come out of ARVO was the release of the long-awaited AREDS2 data. The main result may have disappointed—no increased risk reduction by adding omega-3 fatty acids or carotenoids, although safety improved—but the ambitious study showcased can-do science at its best.
The AREDS2 dataset included a whopping 4,203 subjects, from 82 sites, studied for five years. Meanwhile, some of the most interesting ARVO studies reported on work done with just a handful of patients. And of course case reports (where n=1, the loneliest number) have a long tradition in medicine. Always consider sample size when reviewing data, but keep an open mind. If nothing else, small studies prompt reconsideration of popular axioms.
Big Data is Watching You
A new book called Big Data explains how “data mining” of enormously large samples, such as Google’s five billion searches per day, is transforming many aspects of everyday life. Google can now predict flu outbreaks in real time, faster than the CDC, and translate languages purely based on prediction rather than content.
Can medicine still find room for small studies in a big-data world? Outcomes data will continue to be a driver of health care policy, insurance coverage and medical reimbursements. The clinical decisions you make will increasingly need to be backed by evidence-based medicine, and your performance in the exam room will be factored in to your compensation. Big Data, not Big Brother, is watching you—those PQRS scores are being collected for a reason, you know. Diagnostic tests that don’t yield relevant info won’t be reimbursed.
But data has its limits. What if new research conflicts with time-honored remedies that have proven their merit in the exam room? “We must not slaughter judgment on the altar of data,” Big Data’s coauthor Kenneth Cukier eloquently put it on the BBC4 podcast Start the Week (which I highly recommend if you like to hear smart British people discuss important stuff). There will always be a need for clinical acumen and intuition.
For a great example of how to walk that line, look no further than this month’s special annual supplement,
The Clinical Guide to Ophthalmic Drugs. Authors Ron Melton and Randall Thomas are renown for their ability to balance scientific rigor with keen clinical instincts honed through decades of practical experience. Thanks to leaders like them, our ARVO report authors plus many others, optometry may be awash in data but has a steady hand on the tiller.