Hacker News Reader

Comment by ijustlovemath | Hacker News Reader

We're building a closed loop artificial pancreas (think dialysis but for blood sugar) for hospital use -- the first of its kind in the US. There's a massive unmet need; all critical care patients, and all people with diabetes in the hospital could benefit. Studies have shown you can achieve a 30% reduction in mortality, and 25% reduction in length of stay, in addition to the hours per day you save nurses from managing blood sugar. It's a win/win/win on the lives saved/cost savings/nursing time saves, so we think it'll be pretty important when we hit the market!

Sad to see how few other hard healthtech people there are here, they seem to be few and far between.

View on HN

Amazing! I'm a doctor and founded a software company (https://www.piahealth.co) - even for software as a medical device (SaMD), the regulatory hurdles are tricky and time-consuming, I imagine it's at least 10x for hardware. I have huge respect for what you're doing and hope it makes a big impact.

Can you tell me the difference between your product on the types of prescription cgm + insulin pump combos like dexcom/tandem which offer some level insulin control?

I'm just curious. I run an xDrip set up and I've played around with a couple of the "DIY" closed loop setups.

Our device is for inpatient use, so it's a little more complicated than your typical DIY APS you might be used to as a T1/T2. You have to account for all kinds of different drug interactions, perfusion issues, undergoing surgical procedures, etc etc. The biggest single difference is that we use dextrose as a way to quickly recover from lows (like an automated orange juice dispenser).

Because we're in the hospital and can access IV lines, we also have rapid access to data, and the drugs we infuse get taken up much quicker (5-10 minutes for insulin, 3 minutes for dextrose).

The terminology is overlapping but the space is very different than outpatient glucose control.

Why use dextrose instead of glucagon for lows? B/C patient liver function may be compromised more often in the in patient setting?

In addition to liver function, you can't always rely on patients to have glycogen stores to draw from.

Additionally, dextrose is inexpensive, more available (rural county hospitals don't stock glucagon), easier to mix and store, has a longer shelf life, and most importantly, has a far quicker response time.

Glucagon has promise for outpatient work, where the volume of fluid is much more of a factor, though the stability and cost are still unsolved problems. The patients we treat are in a bed, monitored periodically by trained healthcare providers, with routine access to a pharmacy.

TL:DR Hospital control is just a different beast!

Cool - check out my company: https://replica.health - We automatically log a bunch of data for diabetics based on things like activity and location, and provide an LLM powered search engine for that data. You can ask our system questions w/ natural language like "show me data about whenever I get low post-exercise". If theres any crossover, I'd love to chat about it.

I'm mostly curious about the modelling that goes on under the hood; are you just using deep learning, or also integrating something like UVA-Padova to fit your absorption curves?

Hey, my wife has diabetes, and she's had really awful luck with automated blood sugar monitors. Somehow their readings are always off by insane amounts vs a finger poke

Have you done much research into that area? Do you know if there's a brand we should check out or any common gotchas? (I can't find much reasonable info on this online due to my poor Google skills and all the bad info out there..)

Consumer finger pricks are actually less reliable than CGMs these days (relative difference of up to 25%, vs 10% for something like a Dexcom G6). That being said, a few things you can try:

1. Wait 24 hours, the CGM needs time to adjust to your body

2. Don't overcalibrate in the first 24 hours, or when sugars are in flux. You'll mess up the factory calibration which can lead to worse accuracy over the session.

3. Try a different insertion site. Behind the arm and on the abdomen are the two most common ones.

4. Talk to your endo

5. Call your CGM maker, they will almost always overnight you a replacement if it's demonstrably failing.

We had great luck with the latest generation of Medtronic pumps (the 700 range).

It used to be that they (older models) were off by huge margins, and the auto modes had to be constantly calibrated against as it was often not only off, but suggesting/dosing dangerous amount in either direction.

Nowadays, SO calibrates a few times a week and the pump is extremely reliable. Longer periods of highs are usually due to infusion set leaking (eg. line snagged and pulled the needle).

Like the other reply, it takes a day or two until the system is reliable.

Sensor on back of the arm, infusion on stomach is the combo we've found to be the most safe and reliable sites. Make sure you rotate as often as possible and avoid putting them in the exact same spot.

My wife has Diabetes and recently switched to a CGM. She also saw the insanely inaccurate results.... BUT it improves. After a few hours the numbers get more accurate as it adapts to you. Now she does a thing where she overlaps the old sensor with the new sensor for a couple of hours every couple of weeks. She also still occasionally uses the finger prick method because its results are more indicative of where you are going in the short term.

Wait until your wife's blood sugar is stable to apply a new sensor.

Calibrate it several times when her blood sugar is stable. (Stable means not changing. I want to see a line on the CGM like this "-------------" not one that is going up or down.)

Make sure your test strips and CGM sensors are in date and have been stored at reasonably indoors temperatures.

Source: My wife and I both use CGMs. Our common medical issues were something we had in common.

I was working in medical devices for sometime. I remember Medtronic doing artificial Pancreas. How is yours different?

Theirs is ambulatory, i.e. for people to wear in their daily lives (mostly T1 and insulin-dependent T2). We're targeting a market that currently doesn't exist -- hospitalized patients, specifically people in critical care and people with diabetes on the general care floors. Similar terminology but very different markets!

Their latest AID (automated insulin delivery) system is killer, btw, some of the best results I've ever seen for ambulatory!

Ah ok. I have worked on medical devices in hospital environments i.e the recent Roche Glucometers similar to Cobas.

>Their latest AID (automated insulin delivery) system is killer, btw, some of the best results I've ever seen for ambulatory!

Yes, their medical devices are really good and also very interesting especially their pacemaker implants and also DBS implants for Parkinsons disease.

There was another fantastic device called VAD pumps by Thoratec for which I worked on a PoC. What a crazy tech that is. Thoratec got acquired by St. Jude and finally by Abbott.

This is fantastic, good luck to you. We need more of people like you.

That's awesome! At least in my limited experience developing health care hardware is much more challenging when there isn't a clear "regulatory path" that has been done before. Which makes it harder for completely novel devices (eg versus making an improved pacemaker which already has been approved)

The good thing for us is that artificial pancreases are regulated not as one device, but as three separate (interoperable) devices: the pumps, the sensors, and the control software. Only our control software is under a de-novo pathway ("totally new thing" pathway), everything else is 510k ("we know what this thing is" pathway). We also have Breakthrough Device designation, which really accelerates the regulatory timeline

Through your work have you heard about anyone working on something of a more permanent replacement?

I hear a lot—relatively speaking—about insulin regulation, but in my case the issues are enzymatic. Susceptible to pancreatitis (hopefully all it is).

Any word on the street?

At any rate, keep it up!

Something like an implant? Unfortunately I think we're probably (and this is a wild guess) 50+ years away from being able to fully replace all the hepatic & metabolic functions of a human pancreas in some kind of implantable. They're incredibly efficient, and also horrifically complex, and if you mess up the balance they keep, it's pretty much a death sentence (see: pancreatic cancer).

Controlling glucose is far easier, by comparison; just one input, and two outputs.

Sorry to hear about your pancreatitis, hope you can find a treatment plan that works for you!

I'm doing well, so no worries!

I had no idea those two functions were so different in scope. Nature is a never-ending source of amazement for me. Thanks for the info!

Sounds super cool, are you guys hiring?

Not atm, but Soon (TM)

Reply on news.ycombinator.com