MedXT (YC W13) Announces FDA-Cleared Medical Image Platform | Hacker News Reader

MedXT (YC W13) Announces FDA-Cleared Medical Image Platform

techcrunch.com

84 pointsbrown21 comments

Any suggestion about medical technology startups on HN inevitably brings a rain of comments (often including one from me) about how the regulatory burden is too high, getting approval is a job for lawyers not coders, etc. Kudos to MedXT for doing it anyway.

Sadly most people who complain about regulatory burden know very little about actual regulations. I can't speak to hospitals (they do a LOT) or medical records but if you're doing something like running a clinical diagnostics lab, the "regulatory burden" can be as little as hiring a part time (FDA certified) lab manager who signs off on internal self-regulatory protocols and HIPAA can largely be distilled to "KEEP PATIENT DATA SECRET" with sane retention and security policies.

In my experience, healthcare and medical science are so difficult that the cost of the regulatory burden is a rounding error when you consider the amount of due diligence you have to do to make sure you don't badly mess up someone's life/health.

You're confusing the burden of running a practice, with the regulatory burden of developing a medical device.

The two are completely different.

FDA 510(K) approval for medical devices takes 3-7 years and tens of millions of dollars for all but the extreme cases (i.e. artificial heart or the new bionic eye).

If you're making stuff like pacemakers that keep people alive while implanted in their body, the vast majority of that cost will be unavoidable if you don't want to risk killing a bunch of people (which is, you know, a thing worth paying for).

And the approval process for changing one line in the polling interval in version 19.64.101 of a back-end server that converts from one file format to another file format takes WEEKS.

I'm just pointing out that all of your anecdotal stories were related to practice, not device development.

Please link me to the regulations that require you to wait weeks for approval to change a line of code during device development (BEFORE deployment). After deployment, you need to do a lot of paperwork and testing to change anything, whether it is a drug, medical device, or clinical test and for good reason. You can't just do a git push when you're building a complex system where a misplaced bit can easily kill someone.

As for my anecdotes, I worked on bringing online robotics and brand new diagnostic tests (with and without golden standards) which are under entirely different regulations than record keeping or the day to day operations of a lab.

I applaud their effort, and I sincerely wish them success...

But this is the medical equivalent of Notepad. It is lacking a ton of critical features, which makes it inadequate for almost any real use cases.

The regulatory burden is substantially lower when you don't have much functionality.

Also, $2 per case is way too expensive for what they're offering.

Source: I worked in competing products for a decade.

EDIT: Looks like they're moving in a decent direction, though:

https://news.ycombinator.com/item?id=8038386

EDIT EDIT: I'm not as impressed with the "volume rendering" as I wanted to be. So, time will tell. I wish them luck.

The silver lining of regulatory burden is that there is less competition and higher willingness to pay. It is still annoying though.

Jonathan Bush, CEO of athenahealth, recently ranted at congress about excessive healthcare regulation. His frustration was palpable: https://www.youtube.com/watch?v=CekfvGDiab8

>The silver lining of regulatory burden is that there is less competition and higher willingness to pay.

From the perspective of consumers I'd say this is not exactly a silver lining. (Though obviously I don't think you were suggesting that the excessive regulation is a good situation overall, just that at least there's some incentive for startups to power through.)

The FDA recently announced that they do not plan to regulate medical device data systems.

http://www.fda.gov/MedicalDevices/DeviceRegulationandGuidanc...

That's a draft. I hope it passes, but right now it's just a draft.

From my quick glance,

This ISN'T: novel drug/algorithm/X --causes-better--> patientoutcome, which would acquire trial/approval.

This IS: things we do now, except in the cloud, so you don't have to keep two IT guys on hand even in the graveyard shift - "virtualization".

Firstly, Great job on launching! Secondly, get ready for the most difficult, sloth-slow, gut wrenching sales cycles of your life. Any serious ($$) players looking for software like this take absolutely forever to make decisions on this. And, if you're workflow-process is even slower by a few seconds compared to what rad's are currently using, they will shut you down. So please please, make sure you are absolutely faster, smoother and better then _anything_ else on the market.

Disclosure: I've worked with some serious venture backed (50MM+) companies doing almost exactly this. I would love to give you a full heads up on what to expect. But making excellent scheduling alone is an incredibly difficult feat. Our team spent considerable time working with the absolutely top imaging facilities in the US just to get scheduling down to a science.

Go destroy those big guys! (AGFA/Fuji/GE etc)

I think the major gain here really is getting expert opinion over the internet. I almost worked on a project in grad school trying to compress 3D medical datasets compiled from CT scans using marching cubes, so you can stream these to doctors. I know there is definitely a need, and this solution looks far superior. Are there any plans for 3D datasets on either the server or client side? Radiologists can discover plaque in your heart much faster without sorting through each slide.

(MedXT CEO here)

Yes, we're currently integrating "The X Toolkit" from Boston Children's: http://www.goxtk.com/ It should be ready in October/November.

How is this different from existing PACS applications such as what eHealth Technologies offers?

http://www.ehealthtechnologies.com/solutions/health-informat...

It's not... everyone already has this. It's integrated in most PACS... but it's a WHOLE lot more expensive from the big guys. That said... yeah... radiologists don't really care about cost... so why you would not just go with one of the big guys from RSNA... I'm not sure.

We'd need to know a bit more about what this company offers to answer that question. On its surface... it doesn't look like much more than what other companies already put out.

MedXT also has a reddit integration. Read about it here: https://medium.com/@reshmakhilnani/marketing-your-boring-ent...

3$/study with unlimited HL7 integrations? I'm really curious how that makes financial sense unless HL7 integrations and their required VPNs are billed separately or strictly a "on your own" proposition.

HL7 integrations that are for EHRs that we already support are included at no additional charge. For example, you can export to PracticeFusion, Athena Health and DrChrono automatically as part of the base platform and more are coming. You can also customize your HL7. Some institutions require custom medical/legal documents or complex integration with multiple on-premise EHRs, VPN connections and advanced automation etc, and those are billed separately.

Is this a plan to bring the health information closer to the patient or is it just for the hospital to have the data in the cloud?

Reply on news.ycombinator.com