The Power of More Data in Managing AMD
Digital health technology could change how we follow disease progression in AMD.
MARK S. BLUMENKRANZ, MD
Years ago, physicians directed their diabetic patients to take a standard dose of insulin and then adjusted the dose as needed every few months. Today, routine home monitoring of blood glucose (along with better forms of insulin) means that patients can much better titrate insulin delivery to their actual needs throughout the day, adjusting for diet, exercise, and other factors.
While not perfect, home monitoring has made it easier to keep patients at consistent blood glucose levels, rather than adjusting medication based on damage that has already occurred. I believe we are in the midst of a similar trajectory in the treatment of age-related macular degeneration.
We know that monthly injections of anti-VEGF therapy are effective in treating AMD, but they are also very difficult to sustain, from both a patient compliance and a health system perspective.
Many doctors are therefore moving to treat-and-extend regimens, under which patients are gradually moved from treatment every four weeks to every six, eight, or 12 weeks, based on the stability of their VA and retinal imaging. But the timing of these intervals is largely guesswork. If we guess wrong, there is a risk of significant deterioration between visits.
Dr. Blumenkranz is professor of ophthalmology at Stanford University. He is cofounder and chairman of the board of DigiSight Technologies, a partner in Lagunita Biosciences, and a director and chairman of the Board of Avalanche Biotechnologies. He can be contacted at firstname.lastname@example.org.
Smartphones and other smart appliances (eg, tablets and wearable health technology) can help make more frequent monitoring of vision — and therefore better control of the disease — a reality.
In 2015, approximately one-quarter of people older than 65 in the United States had a smartphone, and the penetration rate among people only slightly younger than that is much larger and growing rapidly.
As these people enter the age bracket at risk for AMD, they will be increasingly connected via devices that are Internet-enabled and that can process large amounts of data very quickly — presenting a great opportunity for us to improve their medical care.
I am using a mobile technology platform, Paxos (DigiSight Technologies, San Francisco, CA), which I codeveloped, to monitor AMD patients between appointments as an adjunct to my established management paradigm for these patients.
Using a simple app called Paxos Checkup (Figure), patients can take a VA test — or one of a number of other vision tests — at home on their smartphone or tablet. They can test weekly, daily, or as frequently as I decide is appropriate, and the results are automatically sent to me.
Figure. The Paxos app as it appears on a smartphone.
I can look at all the data for a given patient, but I can also have a technician review a daily dashboard for all the patients doing home monitoring. We receive an alert when any patient meets a pre-established “trigger,” such as three consecutive measurements showing a two-line loss of vision.
The VA and other tests for which there are predicates have been clinically validated. The VA test, for example, meets high standards for test-retest reproducibility and demonstrates good concordance with in-office ETDRS distance VA and standard near vision testing.1 All of the tests are HIPAA-compliant.
Of course, determining trigger thresholds and figuring out the best algorithms for using these data constitute an evolving science. Initially, the home testing data are to be used as an adjunct to the established management paradigm for a particular patient, but I can certainly envision a day when its full potential is realized, and it becomes much more central to patient care.
MONITORING AT-RISK DRY AMD PATIENTS
Amsler grids have been used as a form of “home monitoring” for AMD for the past 50-60 years, but technology offers ways to improve upon the paper and pencil version. Home monitoring of preferential hyperacuity perimetry for patients with AMD is also available. Patients use a device called ForeseeHome (Notal Vision, Chantilly, VA) to test themselves daily, and the results are sent to a data monitoring center, which notifies the physician of any statistically significant change in the results.
We have very good ways to characterize AMD risk. Patients older than 80 years old with drusen in both eyes are known to have certain genetic polymorphisms or with active disease in one eye are at significantly higher risk of developing wet AMD. Testing at home could be an accurate and sensitive way to detect when meaningful changes occur, even if they are not apparent to the patient.
The reality is that it can be very difficult for patients with good vision to recognize a decrease in vision in one eye due to the conversion from dry to wet AMD, and it is very important to catch it quickly, because we know unequivocally that early detection and treatment of neovascularization lead to better outcomes.
Depending on my level of concern, I ask at-risk patients to perform the Paxos Checkup VA test (Figure) one to three times per week at home. Imagine a typical mild dry AMD patient: she might have an office visit once per year, but if she tests at home even once per week, I obtain 52 times more information about her vision compared to an annual exam. The more frequently she does the test, the smoother the curve becomes, and the easier it is to note a change.
If I receive an alert because her vision has decreased, my office can schedule an appointment immediately, instead of waiting for her next scheduled appointment to perform a full ophthalmic assessment. For many patients, this expedited care is a great relief. They feel like they are taking active steps to manage a disease process over which they otherwise have little control.
At-risk patients can also be screened by an ophthalmologist in a nursing home or other settings where a fundus camera is not available using Paxos Scope, a combined mobile app and hardware add-on that allow the doctor to capture retinal images with a smartphone.
Mobile tools can also be useful in monitoring the course of the disease and treatment efficacy in patients with active neovascular disease. I encourage these patients to undergo home testing every other day or no less than three times per week. Gathering more data is already beginning to help us understand who responds well to a given therapy.
In addition to VA testing, I also like for my AMD monitoring patients to perform Amsler grid testing within the same Paxos Checkup app if they are able. This test is a measure of metamorphopsia, which we know is relatively specific for macular disease.
My goal with these tests is for them to serve as a safety net, so that I do not see a patient at eight weeks who had a hemorrhage at five. Even when a treatment regimen has been working very well, there is no guarantee that it will continue to.
A recent case serves as a good example. The patient was a younger woman (early 60s) with atypical exudative AMD who was being treated with anti-VEGF therapy. Her findings on optical coherence tomography were inconclusive. There was a little thickening but no large pockets of fluid that would have raised a red flag. Her VA was 20/70, compared to 20/80 at the previous visit — again, not a remarkable change. She received her scheduled injection and was sent home.
Shortly thereafter, however, the data from her daily home monitoring showed a clear decline in vision. Based on this, I brought her back in for additional testing and found that her vision had declined to 20/100. I decided to switch her to a different anti-VEGF agent. Within days, her VA started improving. Neither the decline nor the improvement with the change in therapy would have been as obvious without daily measurement.
Home monitoring, of course, is not for everyone. It requires a partnership between the physician and the patient. There are patients who do not want the bother of home monitoring or do not like anything “techie.” For the rest, I believe that digital health tools can become a very important part of their care.
NEW PREDICTIVE MEASURES
Home monitoring and the data it generates also provide new opportunities for refining our risk calculators. We are just beginning to see, for example, that day-to-day and intraday variability can be highly predictive of disease. It is impossible to even detect this kind of variability when you only measure a few times per year.
Low-illumination acuity, which is related to contrast sensitivity and may simulate reading in low-light situations, may also be quite predictive of retinal abnormalities. It is very exciting to contemplate that, with more data, we might be able to identify measures that are more sensitive to early changes.
At-home monitoring of AMD patients in no way replaces care by physicians. In fact, it increases the engagement between the ophthalmologist and the patient. With more information, we can take better care of our patients.
A FINAL THOUGHT
One of the things I love about digital health tools is how they contribute to the democratization of health care. One can envision these tools being used for highly personalized, concierge medicine, for example, for people who can afford to have a physician come to them. But home monitoring is equally useful in underserved areas where telemedicine and modern mobile devices can fill the gaps in care that would otherwise not be provided at all.