Bryan Caplan  

Optometry Challenge

PRINT
Value of Self-Rated Health Ble... Adam Smith Bio Corrected...
Give me one good reason why basic eye exams can't already be done by a robot. 

Written while waiting for a human with an M.D. to repeatedly ask me "Better like this... or like that?"


Comments and Sharing





COMMENTS (29 to date)
Ian Brown writes:

First Online Eye Exam that writes you a prescription:

http://techcrunch.com/2014/02/25/do-i-need-glasses/

Hugh writes:

1) You can't discuss the results afterwards with a robot;

2) Robots make poor salesmen for high-end frames for glasses.

Richard Manns writes:

Well, depends on what you mean by 'basic'. By the bye, you're likely to be talking to an OD, rather than an MD, in the US.

In terms of refraction, an auto-refractor will do that, but it is not as accurate as humans are, and you'll get a dose of cycloplegic in your eye, that will stop you from drive for a few hours at best. They are often used as a starting point for optometrists more-accurate subjective testing, in people who cannot communicate, and in mass screening programmes. But even the best auto-refractors are not as good as humans using the relatively rough-and-ready retinoscopy (before you get to the - https://www.ncbi.nlm.nih.gov/pubmed/15630406

In terms of eye disease screening (at least, that's what British optometrists do), no robot is close.

In summary, the auto-refractors of today are not very good at what they do, and don't do the rest. But would it be good enough for you?

nl7 writes:

Customer service suggests it might be more comfortable for some people to do it with a person, particularly technophobes, seniors, or those with disabilities. But other people would probably prefer to do it without a person, which may produce some marginal embarrassment.

I know that it causes me to feel a little pressure to get it right, since the other person is 100% waiting for me. A robot doesn't care how long I take, so I can be neurotic and take a while or go back if I want to do it again.

No particular reason for a specialist to be involved at this stage. Just like you don't technically need a nurse to take your BP or temperature or height, but they do. For the time being it may be simpler, since they have certain nurse staffing needs anyway. As costs of robots continue to fall, it's almost inevitable that they will be involved in the near future, even if optometrists and nurses are notionally operating them.

MG writes:

I am not even sure this is the "hurdle" here.

For me the pertinent question is: "Give me one good reason why basic eye exams can't already be done by a smart phone app? There are already prototypes of this, and we will see how well they work. This is more of an "end game", since the biggest cost to routine eye eyams is wasting 1.5-2.0 hrs getting to/back and sitting through a site visit - with or without a human doing the test.

Wojtek Grabski writes:

You can. In Ontario anyway, we have the choice of either seeing an optometrist or an opthalmologist - a well functioning and non-covered market. optometrists use these large instruments that spit out the prescription in under a minute - it's nearly free. Ophthalmologists, generally seen by people with disease or with ritzy insurance, take forever and still do it all by hand. When it comes to simple vision tests, they're claims generation robots. The contrast is in service is astounding.

Despite the speculation above, the optometrist is perfectly happy to spend time going over frames after the incredibly short test -- which they hope to subsidize with the sale. Ophthalmologists often don't sell frames, that time can be spent racking up billings.

Thankfully the market did open when the province stopped making it "free", so now we can get an exam by walk-in at a moment's notice at what seems like every major intersection. This should be a lesson for the world, but most people are too anti capitalist to acknowlede it.

The only remaining problem, and it's a big one, is that a vast majority of people are entitled, through work insurance, to a new set of 500 dollar frames every two years. The impact this has on prices is obvious.

William Newman writes:

It takes a certified medical school to do it, and we seem to have largely lost the political technology to create certified medical schools even for humans, so unless you can find an existing certified medical school for robots that was grandfathered in, perhaps it is simply impossible.

RPLong writes:

Because the concept hinges on the word "basic." Caplan may be right or wrong, depending on what he means. It has nothing to do with what the robot can or can't do.

silly sailor writes:

Rent seeking by opticians

ColoComment writes:

I'd be more interested in why so many humans need vision correction. I'd guess that 50% of the population (if not more), exclusive of age-related presbyopia, wear some kind of corrective lenses or have had corrective surgery.

Have humans always been this vision deficient? Is this just a faulty design or is it primarily our modern, industrialized environment that has generated this need?

If we can 'Rolls alone' on the ocean blue, we can do just about anything with robots (or automation).

The ships, envisioned above, work much like self-driving cars or drones, along with remote human intervention, by relying on cameras and sensors to autonomously pilot their course. The Rolls team has constructed a virtual prototype of the autonomous ship bridge in its offices in Alesund, Norway, where it is running simulations of the operating system.

The London-based company claims the ships would be faster, cheaper to operate, and have a smaller carbon footprint.
Curtis L. writes:

Malpractice suits for the robot missing something it wasn't programmed to find.

Jon writes:

Because optometrists also check eye health to ensure you don't have problems like macular degeneration, etc... I'm not sure robots can be trusted with this yet.

One step in this direction would be to have some of the routine operations covered by robotics while the optometrist assesses the results afterwards.

Jess Riedel writes:

For those who mention screening for eye diseases: I think most (all?) of the screening methods use some sort of device that you look into, and the doctor looking into the other end. You could insert a high resolution video camera and have all of this done remotely (and therefore more efficiently and cheaper). This kind of remote assessment of imaging already occurs with radiology (e.g. x-ray reading), apparently called "teleradiology".

http://www.advisory.com/research/imaging-performance-partnership/the-reading-room/2012/11/the-ins-and-outs-of-todays-remote-radiology-market

If you combine this with a robot that determines your prescription, you really should be able to eliminate the doctor from nearly everyone's visit. The only people who will need to see a live human are those who are flagged as possibly having a disease.

On the other hand, I've often gotten my eye check ups at Walmart. It's not as cheap as a robot, but it's darn close.

Arthur_500 writes:

Monopoly

Doctors fought, but lost, to the idea of Nurse Practitioners doing many of the basic doctor visits and writing prescriptions. However, now they demand a doctorate in order to qualify as a nurse practitioner so that will drive up the cost, reduce the value of a nurse practitioner to the patient and keep the value of the medical doctor high.

Most of what an optometrist does can be done by someone well-trained but not educated as a doctor. Some things currently are being done by office personnel. However, if you were to open the floodgates of non-certified persons doing such tasks then you would de-value the optometrist.

The cost to become an optometrist doesn't go down but the revenue would. Therefore you need a monopolistic barrier to entry to keep the revenue stream adequate to support the profession.

Les Cargill writes:

http://en.wikipedia.org/wiki/Therac-25

Brett Champion writes:

Because robots have yet to master the very subtle skill of telling corny jokes.

Wojtek writes:

Les, what does this have to do with anything? Vision tests are optical.

This is like that cliché argument doctors always make: "What if someone comes into a pharmacy with a stomach ache and it's actually a burst appendix!" I think WebMD has proved anyone can be just as much a hypochondriac as the next person.

Anyway:
http://www.eyeglassguide.ca/my-visit/vision-testing/autorefractor.aspx

Rob writes:

You're probably right, there probably isn't a technical constraint to robots administering eye tests.

But show me a robot that says "like this and like that and like this and uh" more entertainingly than these humans.

David C writes:

As Wojtek mentioned, in Canada, you have other options. And if you have no reason to suspect your vision has changed, you can also order online from Canada without a prescription. And that practice, to my knowledge, is decriminalized though not perfectly legal.

Jeff writes:

To follow up on the rent seeking: why do you need a prescription (in most states) to order eye glasses?

P.S. Also there are already machines that tell you your vision. I believe Zeiss has a machine that maps your eye and can provide more precise correction than the usual half-diopter steps. (So yes a robot can already do it better).

Hazel Meade writes:

One of the sick things about the US optometry industry in the US is that optometrists will generally refuse to tell you your pupil distance or provide it as part of your prescription. This effectively forces you to purchase a pair of eyeglasses from a brick-and-mortar store. Which usually biases people towards buying from the often-adjacent (though legally separate) frame store.

In order to buy online, you have to know yur pupil distance, or your glasses won't work right. But good luck getting the optomotrist at Lenscrafters to tell you.

MingoV writes:

Ophthalmologists usually leave routine eye exams to optometrists.

Ophthalmologists do more advanced exams that include dilation of the pupils and fundoscopic exams of the retinas. The list of conditions that cause abnormal fundoscopic exams is long. Today's robots are not good enough to identify AND interpret retinal abnormalities.

Silas Barta writes:

"Robot" might be pushing it. An easier threshold to discuss might be "lightly-trained worker".

RPLong writes:

24 comments to date, and not a single "optimal-ogist" joke... tsk, tsk...

John T. Kennedy writes:

[Comment removed pending confirmation of email address. Email the webmaster@econlib.org to request restoring your comment privileges. We have attempted to contact you previously. We'd be happy to publish your comment. A valid email address is nevertheless required to post comments on EconLog and EconTalk.--Econlib Ed.]

Finch writes:

There are only around 25,000 ophthalmologists in the US. The market is not that big relative to the technological level and quality control required, so it will generally be slow to attract R&D dollars. It will progress, but not at a really high rate.

It's not like inventing the next top-selling drug where the money will poor in. It's like inventing the next large marine diesel fuel injection system. There's a market, and it will get done, but there's not much reason to spend what would be necessary to hurry it.

geoih writes:

Because it's more profitable to invest in the labor of a human versus investing in the capital to create and maintain robots, based on the projected earnings.

Duh. I thought this was an economics blog.

patri friedman writes:

Because optometrists are an entrenched, concentrated interest group. To compete with them, robot manufacturers would need to somehow exert greater political pressure, despite not yet being allowed to sell a product. as with most disruptive technologies, the key is primarily regulatory hacking and clever stealthing from the targeted industry and associated interest group.

As an aside, I find it somewhat frustrating when professional economists ignore public choice. It is extremely good at answering most questions about bad regulation that I see econbloggers post.

Comments for this entry have been closed
Return to top