Hacker News Reader

Comment by rnavara | Hacker News Reader

124 pointsrnavara68 comments

I’m in my 20s and have been through the 3 day hospitalisation process NINE times so far - I cannot stress how huge a deal what you’ve built is for me

The amount of hassle and anxiety this will save me is impressive on its own, but so to is the ability to have a normal lifestyle (such as going out for a drink) without having to worry about whether or not I’m always in range of a hospital that I want to spend 3 days in if I’m unlucky

Thank you and good luck!

My 72 year old mother was just in the hospital for 5 days with AFib (not her first). We had an escalating series of interventions and procedures, none of which proved effective. By the end, I suspected she had been substantially over treated, at high cost to medicare and with high risk of complications.

So, I think this is a great idea.

She is currently waiting for an ablation, which was a close call vs leaving her in it, but she wanted to try it.

Thank you so much for sharing your personal experience. Best wishes for a speedy recovery for your mother - this is unfortunately a common theme among our patients and we're excited to make a big change here. Let your mother know that we're pulling for her!

Fascinating. The anti-arrhythmic load was one of the less interesting reasons for hospitalization from my days in residency, and more important than my own intellectual curiosity is the ability to keep people out of the hospital.

Typically at my institution we load first and cardiovert later if needed. How is that handled in your system?

Also, what is the clinicaltrials.gov registration for your completed Phase I?

Yes, you get it - it's a boring hospitalization all around! For patients getting a cardioversion after their drug-load, they can fortunately come get one as an outpatient (most cardioversions are for elective outpatients, and in fact the inpatients often get bumped for the scheduled outpatients). Hospitals are particularly happy about our workflow because they don't get dinged for readmissions - whether for cardioversions or for drug dose adjustments, which also require readmission under the current standard.

Phase I Software validation doesn't require clinicaltrials.gov registration, but our Phase II trial will be per FDA!

Thanks! Glad you've already considered the cardioversion workflow. If your system is good enough to do the load+monitoring safely at home, this sounds like a win-win.

Also, I didn't realize that Phase I trials didn't have to be registered on clinicaltrials.gov! But you're right, they're specifically exempted[1].

1 = https://clinicaltrials.gov/ct2/manage-recs/fdaaa#WhichTrials...

Amazing, seeing how you as a doctor could think of processes that you do, and conceptualize it in software!

Congratulations on the launch, and I wish you well, and hope it is of great use to thousands!

> existing FDA-cleared hardware Can you please link to this?

Thanks so much! Medicine is often a mix of extremely high tech and very low tech, so we're in a great position to fix the latter :) We can actually pair with multiple different types of hardware (our software is hardware-agnostic) but we have a couple preferred hardware devices. One of which is https://www.kardia.com/kardiamobile6l/

So you recommend their hardware, and compete with them on software? (Assuming I understood that correctly -) what stops them catching up/replicating your (presumably) better software?

Or perhaps more to the point - the problem you mention in OP seems, as described, like it's already solved by Kardia, what was missing that you had to build?

Great point, indeed the problem of automated ECG analysis has been around for the last several decades and there's still no automated solution accurate for AFib patients (one of the the hardest categories for making ECG measurements). Our software is specifically trained on AFib ECGs and our specific focus of this market and guiding drug dosing, as well as our clinical expertise in the field gives us our edge. It's a similar reason why startups can do things that Google and Amazon can't :)

Sold!

Now, how do I also get your software that will work with this hardware?

Software still needs to pass its FDA clearance, but great to hear your interest!

They are referring to 510(k) Clearance [0]

> The 510(k) Clearance process is one of two ways by which a medical device manufacturer can obtain approval from the Food and Drug Administration, although legally, a device clearance under this section is not a true “approval.” 510(k) Clearance simply gives the company permission to market its product, based on comparison to a “predicate” [previously approved] device.[1]

Basically, If a somewhat similar device is already on the market, the FDA will clear the new device for the market.

[0] https://www.fda.gov/medical-devices/device-approvals-denials...

[1] https://www.levinlaw.com/fda-510k

Excellent, yes - thank you for listing these. EKG devices have been around for a long time, so plenty of predicates for 510(k) pathway.

Lots of potential questions here, but let's start with three:

- What makes you "better" than your competitors, perhaps especially those of your competitors with substantially deeper pockets than you ? (e.g. Apple Watch which also has "ECG" functionality).

- Building on the above, how's the prior art looking on the patents ?

- Finally, general cynicism on "home ECG".... clinical evaluation is 12 lead ECG. Can you really get away with substantially less input ? (Garbage In - Garbage Out).

Great questions. The Apple Watch is a single lead, so doesn't have the resolution to do the computations. Also, the algorithms aren't trivial for automation this, but the data moat is extremely important for capturing and appropriately testing this population. There is prior art, which our own filing took into accout. Lastly, the hardware devices we use are all FDA approved to measure the appropriate interval, one of which is FDA approved 12-lead ECG patch approved for patient self-administration.

That and the Apple Watch is on the wrist, which could introduce artifact to the ECG.

There are single lead chest ECG patches used for specific diagnostic reasons. Zio patch is one common device used. [0]

[0] https://www.irhythmtech.com/patients/how-it-works

Excellent points. Artifact is indeed a tricky problem for EKGs, and the ziopatch is a great device. Generally, multiple leads are needed (ideally 6 or 12 leads) to accurately measure repolarization of the heart.

> That and the Apple Watch is on the wrist, which could introduce artifact to the ECG.

If patent filings are a guide[1], Apple are aware and have been investigating options for user-friendly additional sensors elsewhere on the body.

Whether it comes to market is another matter, of course.

[1] http://patft1.uspto.gov/netacgi/nph-Parser?patentnumber=1108...

Good points!

Someone I know had their open heart surgery yesterday(doing well), Although about 50 years old he had very active, healthy lifestyle. He missed the master health checkup for a year due to covid lockdown and the latest one showed narrowing of arteries from ECG. It came as a surprise.

I hope your service makes ECG more accessible to identify heart diseases earlier.

Glad to hear he is doing well. Yes, there is a tremendous need to be able to push healthcare to the home setting, and of course early detection makes a huge difference. Thanks for your comments, and we're glad to be part of the solution to address this need, our focus being more on heart rhythm disorders.

> We’ve done a 103 patient pilot study, which showed that our software successfully interpreted 100% of ECGs and could even replicate doctors’ clinical decisions.

Do you have a link to the study? Isn’t that a relatively low number of patients to validate against?

Great question, our training data involves thousands of ECGs to validate the software performance of the algorithm itself. As far as in-human data, 100 patients is well-powered for Phase I software validation, in preparation for our upcoming prospective clinical trial (Phase II). These results are hot off the press so we don't have a public link yet - more to come!

Somewhat related: William Osman saves $70K by building his own home X-ray: https://youtu.be/IiJAq53knwc

That's a cool hack. Xray machines are quite costly for clinics to purchase, and clinics have to perform hundreds of xrays before they recoup their cost.

Ha awesome link, thank you!

Interesting concept. Basically the main use case in the demo video seems to be assessing the QTc in patients on sotalol and potentially uptitrating the dose? I’m surprised people are kept in hospital for three days to just check qtc for sotalol. I wonder if this is a US only thing because it doesn’t happen where I work. Also the video doesn’t make it clear which QT correction formula you use. If you’re using your own correction formula for AF is it published and validated?

Great pickup - yes indeed, inpatient admission for drug initiation has been the standard for too long (30 years to be precise). We use the standard Fridericia formula preferred for assessing drug risk, but we also use Bazett correction for algorithm validation as it is used more often clinically

This is a great team - addressing a real unmet need. I have family members that can benefit from this. Keep up the great work !

Thank you so much for the support! We aim to make a dramatic improvement in AFib care not just for patients but their families as well

> Our take-home kit combines FDA-approved wearable hardware devices with our ECG software that automates measurements, recommends drug dosing, and catches changes before they become heart attacks.

It seems fair to summarize this as a "smart(er) Holter monitor".

Yes - that's a pretty good way of putting it. Holter monitors are only single lead, but otherwise yes!

There are Holter monitors that feature numerous electrodes.

The other thing is that typically Holter monitors do not provide real time analysis. Patients usually put them in the mail, at which point the 14 day data is reviewed. If you know of some recent advancements (real time, multiple leads, FDA approved for interval measurements) would love to get a link. Thanks!

Yes; that is what I mean by "smarter Holter monitor". A regular Holter monitor contains some firmware that logs to some flash storage or whatever. And there may be no display or anything.

I wonder whether there exist some Holter monitors that could be made more functional with just a firmware reprogramming job. You really need interfacing though: BT, Wi-Fi, ...

Agreed!

How does a 3 day hospital stay cost $17k?

I am genuinely curious to see the breakdown of where that money goes

That's not expensive if they are doing tests and stuff. I got an MRI of my foot and did it at a hospital after getting rescheduled three times at an outpatient facility. It lasted an hour and was $2700.

On the outpatient facility they would have charged $1100.

So true, and a revealing insight into US healthcare! The vast majority actually goes toward hospital room and board, with a fraction of the cost for the ECG monitoring. This is why insurers are keen to reimburse an out-of-hospital alternative

This sounds great.

On a related note, I've been thinking about building something similar for personal use using one of those audrino based ECGs (just recording and then finding anomalies). Never got the time though.

Interesting, that sounds like a useful tool - we all could benefit from a better understanding of our heart health at any given time!

It's mostly because my young kid has SVT and it would be useful to stick on a monitor occasionally to see if there are brief breakthrough episodes evading our periodic pulse checks. Of course they have monitors to do this already, but they are very expensive even for short uses and this use case wouldn't need as much detail or sensitivity.

Sorry to hear about your kid, SVTs can be tough. May be worth asking your pediatrician about other options for monitoring that are suited for pediatric use!

I actually have another idea for a medical device (different specialty). Do you have any advice or suggestions for someone like me (non-medical) to get started?

Cool. I would recommend emailing a doctor in that specialty and asking them if seems like a good idea. Or alternatively, you could ask a bunch of patients to determine need for it based on their experiences. If it is something in pain management, I may be able to help.

Thanks, I meant from a paperwork perspective, trials, etc. I know about the FDA website for medical devices. Not sure if there's anything specific to what is essentially lab equiptment interfaced with a cell phone (not patient worn or anything). So I'm just guessing, but I would think approval and testing would be easier for say a home pregnancy test vs a ECG.

I've already validated that the idea would be useful. There's a potential annual patient population of around 7 million people in the US. Even at 10% adoption, that could be revenue of $100-200M with a conservative guesstimate profit margin of 10-20% at that level (variable costs are low, so that should be higher as it scales). Current price of testing is high enough that I think this device could fit into that range or under cut it, with a multifaceted improvement to patient experience. A brief search turned up no competitors.

Now I'm trying to decide if I should spend the money to build a prototype and file a provisional patent or not. It depends on exactly how expensive and lengthy the approval process is, how HIPPA might or might not apply, etc. The type of software it will require will also be a challenge for me since I've never made anything like this.

How will your product prevent deaths? I understand it allows for convenient ECG monitoring at home. But if there is a problem observed in the ECG, is the patient diverted back to the hospital?

Great question, by detecting and alerting the earliest ECG changes that are well-validated risk factors for sudden cardiac death, we can prevent these deaths and get the patient the care they need

Is this available now? I have advanced heart failure and would be interested in being a guinea pig

I also have an lvad and spend a lot of time in the hospital taking quite a few medications

Thanks for your message. Unfortunately, it is not available at the moment, still needs FDA clearance. Wish you well.

The dashboard looks really simple with great features! Awesome

So glad to hear your feedback! Our aim is simple and safe :)

Great idea and super exciting results. The more we can keep patients out of the hospital (even outside of Covid times) the better for everyone.

Thanks so much, yes we agree - patients have hated being hospitalized to start pills long before COVID! But now, as you point out, there is even more urgency to treat people safely at home.

that's really cool. how do you classify a patient as "low risk"? can you base this off of history and lab work, or will they need TTE and a baseline EKG first?

Love these cardiology-specific questions! Low risk would indeed be based on the history, labwork (e.g. renal function), baseline ECG including HR, and EF

This is awesome guys! Congrats on the launch!

Thanks so much for your support!

Wow, I see so many great use cases

Thank you! Yes, feels like new applications seem to arise faster than we can consider the previous ones.

Awesome, yeah we mentioned it briefly but heart rhythm side effects are a problem for so 300+ drugs - antibiotics, pain meds, psychiatric meds. The list goes on!

great idea and traction! do patients know how to take their own heart tracing?

Thank you so much! Our patient interface guides them through how to hold the device, it's a very simple device so even our 90 year old patients love using them :)

This is awesome!!

Thanks, glad you share our excitement.

Side note. The site does not look like a "medical" site, too much black. Some green (#57b792) and white will look more "medical" to me.

Ah interesting note, we chose that color scheme for our software interface including that green - you're right, that does convey medical!

Reply on news.ycombinator.com