I’m feeling a little testy on ophthalmic testing this month. It’s the application of science and statistical validity in our field, an area that seems to get very little attention. The key word is “outcomes,” and is the common thread that we may be missing. Does your test or procedure truly have a benefit in terms of aiding quality of life, or preventing vision loss? Is it repeatable, and does the result truly distinguish abnormal patients from normal ones? Now, I know you’ll argue that “we” don’t contribute to this issue as much as do other specialties, but that holds no water. We are as responsible as any other profession to provide benefit and value in quality of care and cost.
Let’s look at some examples.
Visual field testing for headaches is a cornerstone. Yes, we all know that it’s subjective and difficult for the patient, especially the full-threshold version. The basic principle is that if a test has a higher chance of FALSE POSITIVES than the chance of the patient actually HAVING the disease in question, we are in deep fertilizer for running the test. Our office experience has been that visual fields have a 10-20% rate of false positives. We’ve all seen visual field cuts that are pretty convincing, yet non-pathologic. Have you considered the false positive rate for your own practice? What are the chances that a headache patient without neurologic symptoms will have a brain tumor? It’s not zero, but it’s not 10-20%, either. The residue is a visual field that has eyelid, nose, and poor attention defects that require some explaining.
Screening with our high tech scanning instruments is in the same category.
Dr. Jim Fanelli and I wrote about the folly of relying on normative databases in our testing instruments. (Footnote 1) Yet, we hear of doctors making referrals or initiating treatment for things like “nerve fiber layer defects” and “thin ganglion cell complexes” based on single tests. These aren’t false positives per se, but comparing the patient to a normative database is a weak technique, as compared to the proper comparison using repeated testing. One of our early optic nerve scanning devices had 300 Englishmen and women as its normative database. Imagine the results when we scanned patients who were black or Hispanic! ALL were shown to be “outside normal limits.”
Another problem with scans is the multiple findings that have nothing to do with the patient’s symptomatology or clinical findings.
Wouldn’t it be a GREAT public health benefit to do MRI on every patient at every physical exam? Don’t laugh, as this was an actual “business” here in New Jersey. Overdone MRI scans produce findings that have nothing to do with anything, and would unfairly require the family doctor and radiologist to interpret artifacts and other findings that are not relevant to the patient’s well-being. (The “business” was ultimately shut down by the state medical board.) Our “screening” with high-tech digital instrumentation dances dangerously close to this line.
Genetic testing…..oh, boy…….
The future for this is amazing, as evidenced by the breast cancer community’s results in personalizing treatment based on individual genetic features. On the other hand, look at our own efforts in treating macular degeneration and understanding its risks of vision loss. We’re told by some (Awh) that we have to do genetic testing in order to decide on appropriate use of zinc supplementation. The opposing view, held by Chew et al. at the National Eye Instutute, is that there isn’t evidence to support this. (Footnote 2)
For what it’s worth, our large OD/MD group has found that the current, too-primitive level of genetic testing for AMD was upsetting to patients on one hand, or falsely reassuring to others. The average patient lacks the capacity to understand what these risk analyses mean, and the analyses themselves did not lead to better outcomes………for now.
This is why:
It’s interesting to look at the work of medical ethicist Arthur Caplan, PhD of New York University. Caplan summarized his work in the area of genetic testing by saying, “We haven’t really figured out the significances of all the genetic mutations that are found, and what they mean in different environments, meaning the same mutation in a different person may not mean the same disease.” (Footnote 3) In other words, we aren’t quite there yet.
The best practice is to have an acute understanding of what our instruments DO and not just what they measure. Grasp the accuracy, flaws, and published science for each clinical test that is run. Go case by case and ask, “Will this information lead to a better quality of life for this patient?” Let’s be part of the solution, and not part of the problem.
Liu Y, Seddon J, and Sobrin L, “Genetic Testing for age related macular degeneration,” Retina Specialist, June 2015
CNN online “Misdiagnoses: A hidden risk of genetic testing” by Jacqueline Howard, October 31, 2016