Hacker News Reader

Comment by ska | Hacker News Reader

The cost of components is often not a big part of the cost of producing an imaging device like this. Building hardware and software in a regulated market where human safety is involved requires levels of validation, verification, process control, risk analysis and documentation that can easily dwarf component cost at the sort of run sizes we are taking about.

This isn't the only thing going on, but it is naive at best to think of the price of these units as a parts list.

View on HN

I agree with you but it can still be frustrating and there's a lot of entrenched interests involved.

My story is that I broke my foot about 10 years ago (in the Grand Canyon no less). Long story short, when I eventually got the doctor, she told me that this kind of fracture always took a long time to heal and the literature suggested that electromagnetic stimulation had some benefit. Of course, the insurance company wouldn't pay until it was deemed a "problem fracture" in a few months (as my doctor thought it would be) but I did eventually get my $1000 stimulator at about 50% off.

It had the most horrible interface one could imagine. If it slipped off the foot, it apparently simply shut down for that day's stimulation. In any case, my foot eventually healed.

A few years later, a friend had a similar injury. I gave her my stimulator. Apparently the allowable stimulations had expired by then. She took it into a medical supply place and they were horrified by the idea that they might simply reset the device.

I'm not a big pharma hater but this sort of story is why people get really mad.

Those ridiculous pricing policies are necessary to recover the cost of regulatory compliance which is seriously huge component of anything medical in USA.

As a side note, I know a company in India that manufactured a state of art blood test device. They tried to market it in USA and eventually gave up. Currently the company has tie up with Singapore hospitals which send blood to India for testing via flight and reports are mailed electronically automatically. There is a rig of blood testing machines in Powai-Mumbai which tests thousands of blood samples each day. This machine clearly can not be sold in USA (last time I checked was 6 years ago) because the cost of regulatory compliance is prohibitive.

If the medical device manufacturers indeed made windfall profits I would love to invest in their stock but they don't seem to be doing as well as the hatred directed towards them.

My portfolio is about 50% medical supply and machinery manufacturers, and they're incredibly steady investments which pay strong dividends even and especially when the stock price is down. Furthermore, they're often able to make biotech plays and their guaranteed revenue means they're strong candidates for acquisition: http://www.bloomberg.com/quote/HSP:US

Do you have any sources for your claim that regulatory compliance is hurting profits beyond a one-off anecdote? Because my experience has been that it's an industry rife with windfall profits.

Also, what's the name of this Indian company, and what's the brand name for the device? If it does what you say it does, you would be pretty foolish not to invest.

"More than three-quarters of the cost to bring a medical device from concept to the U.S. market is spent clearing regulatory hurdles, according to a Stanford University report."[1]

I am not saying the companies aren't profitable. They aren't profitable the way people make them out to be.

And if there are indeed windfall profits to be made what is preventing smart-ass entrepreneurs from jumping in and making those profits ?

Tim Cook says that the Apple Watch won't become a regulated medical device, but Apple might make another product that is. Cook made the comments earlier today in an interview with The Telegraph, stating that putting the Apple Watch through Food and Drug Administration testing would slow its release cycle down too much and "hold [Apple] back from innovating."

[1] http://www.massdevice.com/study-fda-related-activities-cost-...

"And if there are indeed windfall profits to be made what is preventing smart-ass entrepreneurs from jumping in and making those profits?"

Incredibly high barriers to entry and very large established players who have already cornered the market? Following your logic, oil companies must not be very profitable since there aren't very many entrepreneurs buying their own offshore drilling platforms.

"They aren't profitable the way people make them out to be."

Who are these people? What is this way?

What concrete assertion are you refuting? How do you explain the P/E ratios of the biggest medical device companies?

If the cost of regulation is such a drag on profits and innovation, why are medical supply companies that primarily serve the US market so much more profitable in the long term than their counterparts in other countries?

Are you alleging that there's a bubble, or that they're cooking their books, or that they're not one of the most consistently profitable industries?

Wish there was something for medical manufacturing like pricepain.com - including a reduction in unecessary red tape. Sad, how it stifles innovation and actually harms more than it helps.

We could never have an 'open' or cheap medical device because of the extreme level of regulations and the tendency to pretend that the laws of economics shall not apply when it comes to human lives and medical devices.

i'm in the medical industry and love FOSS and I've contemplated this a lot. nearly every day.

I believe the way forward is with open, LGPL-style, reference designs, almost like the Red Hat enterprise model. individual companies must still be accountable and liable for their devices, but their real value-add is in the verifications and validations.

The auto industry doesn't share software, but the liability model is similar in that: how to build any car is mostly known, there are few important secrets. You choose your car based on features and quality and the companies provide those validated features and quality to customer expectations.

Reference designs in medical devices would especially help with the security crises we're undergoing.

So you could buy, say, an insulin pump based on a reference design from a dozen different manufacturers, but your feature set and quality of materials would be up to you and your insurance company. This is the best compromise I can find which allows competition to improve quality and spur development (competition is weak in med. devices) yet allows thousands/millions of eyeballs to improve software - and therefore life - quality.

The holy grail is a medical-grade embedded Linux. If any hardcore entrepreneurs want to talk with me about that, I have a lot to share on it.

That device practically screams, "hack me".

Are you just trolling? This is a device that isn't connected to any network. Indeed, it barely has a user interface.

That word doesn't think what you think it means.

> That device practically screams, "hack me".

As in, "That device practically screams, 'please find a way to tweak me to overcome my restrictive and unreasonable limits (embedded in what is otherwise a functional device)'"

Or in other words, "hack me".

Fair enough although "hack" in the context of medical devices is most commonly used these days to mean something like "maliciously access."

Hack in most contexts these days means maliciously access. Yet, you are not currently browsing a community of malicious accessors, but hackers.

User also didn't get the jones joke. Lost cause.

I assume they meant hack in the friendly, Hacker News-type sense instead of the malicious, my-grandpa-thinks-someone-hacked-his-Google sense.

fractured a metatarsal huh?

Jones fracture, specifically.

I also fractured my fifth metatarsal. It was a Pseudo-Jones fracture so I was up and walking in about 4 weeks(with lots of pain from the toe region and heel for some reason). I was just thankful it wasn't a full jones fracture.

J-what? That's a whole other "leg" that doesn't even have metatarsals.

https://en.wikipedia.org/wiki/Jones_fracture

I worked in R&D at Philips making ultrasounds. From my experience this is definitely what it is.

The amount of people that were employed to simply test, do documentation, and verify safety alone was staggering, let alone all the smart people it took to develop custom ASICs, FPGAs, mechanical housings, image processing etc. There are physicists who do research, there are sonographers who help achieve the best possible images for clinical scenarios.

The amount of effort it takes to make a high quality, safe, ultrasound is immense. When you add in the fact that you also need to follow a myriad of FDA regulations, even small code changes often require a huge amount of overhead in documentation and review before they get checked in, and then there will be whole teams of people to manage source control.

Ultrasound companies do an immense amount of R&D, but also put out a safe, well-regulated product, and the prices reflect that. I'm surprised the the author in the article found even a low quality ultrasound as cheap as he did.

Sometimes that's just a hangover from "the way it was always done.' Telephone switches used to be insanely expensive, and developed with fleets of coders, QA, documentation people etc. in projects costing hundreds of millions. Now any shady outfit with an Asterisk server is a phone company. Five nines? Is that a burger chain?

Not saying that's how FDA approved devices should be developed, but don't be surprised if "prosumer" grade equipment starts penetrating the medical field especially if anyone gets the cost incentives right.

Having taken courses in biomedical engineering, and done some research for a company looking to do what you describe, I can say with some certainty that it is very difficult to start a medical hardware company. You need to outsource production to an expensive certified CM with large expensive MOQs, unless you want to go through time consuming and expensive manufacturing approvals yourself (before being allowed to make or sell anything). You must also develop (easy) and have your device approved (hard), as well as not being allowed to change anything without additional time consuming and expensive approvals. Add in the fact that most diagnostic electronics require the highest level of scrutiny, paperwork, bureaucracy, and delays, and you quickly figure out that you're better off doing something else.

For what it is worth: It can be done though, completely in house, in a couple of years by a smallish group of people who know what they are doing.

I wouldn't be surprised if we see a number of modernized devices in spaces where there are established consensus standards.

Are the regulations very different in different countries? I can think of a few places where getting live-saving machines an order of magnitude cheaper with a small percentage increase in risk of causing harm might be an acceptable trade-off in some cases.

I've been to hospital in one of the places you're thinking of, and I noticed that phrases like "approved for use in the EU" were used as code for "this is not a scam".

I fear the market niche between "scam" and "not a scam" might be very narrow.

I mean, I was thinking less private-clinic of dubious standards in a city, and more "what can the national health services of developing countries reasonably deploy to most rural communities?".

For example, I am from Mexico, and cysticercosis (a brain parasite infection) is not uncommon in rural areas in Mexico. MRI scanners, which can be used to detect it, on the other hand, are quite uncommon outside of major cities. This might or might not be a good example, since I don't know the cost/safety trade-offs for MRI or all the alternative solutions, but if one could get, say, $20,000 USD MRI scanners that are close to in practice to $1 million ones, it might be worth a small safety delta. Public health in Mexico at the national level also tends to fall under "generally trustworthy enough, but critically under-resourced".

Edit: Again, not sure how realistic this particular scenario is, but I think in general getting cheaper equipment for national health services might be worth it in cases where having a lower-standard device can save an order of magnitude more lives than it harms.

This is not my area of expertise, but in speaking to Regulatory Affairs people when I was working in the field, most of the EU and some of Asia base their regulations on the US FDA, or at least have very similar regulations. In addition to FDA audits, we were also periodically audited by the regulatory bodies of other countries.

However, in many countries, enforcement is lax, so you get lots of devices that don't come close to meeting whatever regulation is in place.

They appear to take some very specialized training to operate them well also. There is all the overhead of training the operators and attempting to build something that is reasonable for them to use

I can understand the need for quality testing, but what's the medical threat they're trying to guard against from ultrasounds?

Equipment and other supplies are expensive in medicine, in part, because a lot of it is bought by insurance and by government funds. These sources of cash are milked for all they're worth by the vendors. Then the fact that the market is difficult to get into (big hoops creating an entry barrier) creates a ring of price protection.

But this article isn't proposing to make a low-cost ultrasound for regulated life safety applications. It is proposing to make a cheap ultrasound imaging device for fun or other general purposes.

So why aren't we creating some high level framework, for development of a large class of devices, validate it a single time - and enable such devices be designed with much less effort ?

Because it doesn't work that way. You can't validate the framework, you have to validate the product

Actually you have to validate the iteration of the product. So if your software changed significantly enough from version 1.0 to 2.0, enough that the results of processing data may have changed, now you have the pleasure of going back to validate the entire product to ensure it still works as expected instead of just doing a software verification.

Sucks, but it's the price of safety.

One of the stories I love to tell from my medical device days was spending about 6 weeks getting the validation procedure document for a software build machine approved and then spending all of 5 minutes actually building the product (which had to be separately validated), never to use the build machine again. Now, that's crossing the line into stupidity!

Considering the tragedies that could be avoided by having access to cheaper ultrasound equipment, the price of safety is not easily quantified here.

We're talking about two different things. Safety is the lack of Hazard that a Device presents to the Patient (or Operator). If the device doesn't exist, then there's no Safety issue.

What you are talking about is access to medical care, which is an entirely different topic

But why can't you validate the framework ?

It's not that you can't it's that it makes no difference since it's the final program that has to be validated, not just its components.

There are cases where individual elements of a system must be validated separately, but even then, the system as a whole still has to be validated.

Reply on news.ycombinator.com