Slashdot Mirror
Embedded Linux On a Digital Stethoscope
An anonymous reader writes "A team of electrical and computer engineering students at Calvin College is designing a digital electronic stethoscope running uClinux as its operating system. While there are many embedded devices built on Linux operating systems, medical devices running open-source software are extremely rare because of the perceived difficulty in obtaining FDA validation. The device is in its early stages of development, but major hardware choices have been made, and the team has recently released a Project Proposal and Feasibility Study."
OK I'll bite. Why does a stethoscope need an OS in the first place? How much file I/O, multitasking, networking etc. does a single-purpose device like this really do?
...installing Windows Vista on the aforementioned stethescope. Would there be room enough for McAffee?
No, no sig. Really.
ThePromenader
If you had, then you would have read their site and had your questions answered.
I can't see this going very far. What doctor will wrap that thing around his neck? Unless, of course, it also plays MP3 files.
There may be, however, a solution looking for a problem in the fact that the audio can be recorded for playback later. However, I can see a situation where a doctor is sued for malpractice, and the audio is used against him rather than for him in the case. I'd be rather wary of using this machine, based on those grounds alone.
It is dangerous to be right when the government is wrong.
does it run linux? I am gonna plant linux into my brain! does my spoon run linux? does my dog run linux? does my gf run linux? is the water linux friendly? will consuming that extra donut will have marginal effect on penetration of linux? omfg i need stethoscope now!
sorry...i had to do it. hail Free speech!
I would date a girl who runs linux. Sadly, I think that women will remain closed-source at least for the rest of my lifetime.
I think if I was doing this, I would have started with rockbox - it already has all the audio playback and recording, runs on cheap, easily portable hardware (MP3 players). I'm assuming you only need 2 channels of audio input - otherwise you might have some extra hardware hacking to do. The only tricky bit would be the noise cancelling I don't know how much processing that takes.
ccalam - acoustic versions of new songs.
littmann make a variety of stethoscopes including digital ones, and have done for some time.
However, dont believe the hype. They are of very limited clinical value. No hospital cardiology diagnosis these days would be based on auscultation. (echocardiography would be used). For taking BP the traditional tube plus ear-pieces 'scope more than suffices.
When the seagulls follow the trawler, it's because they think sardines will be thrown in to the sea
If it has an OS can it run other applications? like for playing mp3s?
Doctor: Oh my, you seem to have arrhythmia!
Patient:?!?!
Doctor: Oh sorry i forgot to turn off the mp3 player, that was just a Murcof track, my bad.
"While there are many embedded devices built on Linux operating systems, medical devices running open-source software are extremely rare because of the perceived difficulty in obtaining FDA validation."
Perceived? You mean that there's no actual difficulty?
has been done, although without Linux, but thats made up for with some SW patents.
http://www.medicom.bang-olufsen.com/sw435.asp
The problem is that the sound quality is too good for conservative doctors, so they had to make a function that imitates the sound from an old tube stethoscope.
the look on someone's face when the thing locks up while in use. Either this man is dead or my linux driven digital stethoscope needs a reboot.
OK, I realise that most of the other replies are critical, too
Really this is just using tech. for it's own sake - and introducing a whole bunch of unnecessary problems into what is really a very simple procedure. Instead of a simple piece of rubber tube, with a "thingy" at each end you are now reliant on an embedded device with it's own power requirements, a link to a host PC with software compatability and yet more power requirements and finally the doctor or nurse wears a headphone to listen to exactly the same sounds they'd hear through a conventional stethoscope.
there is a slight glimmer of hope for this: remove the doctor.
If you can possibly make this device totally idiot-proof and throwaway cheap, it might just find a use for remote diagnostics, where a patient can self-monitor, upload their "swooshing" noises to an AI which does most of the work.
Of course this presuposes that the patient isn't too busy having a heart attack, and can get their PC connected up to the net, and the battery in the device hasn't gone flat, and the AI at the other end is working, and someone can take the necessary action (if needed).
On second thoughts, scrub the glimmer of hope. There are just too many things can go wrong
politicians are like babies' nappies: they should both be changed regularly and for the same reasons
medical devices running open-source software are extremely rare because of the perceived difficulty in obtaining FDA validation
Do you have any pointers to actual decision makers talking about this "perceived difficulty"?
And why would anybody think anyway that open source software should have a harder time getting approved than closed source software?
They are using a 32bit micro, running on batteries, running linux. It is not a low power micro, and if similar choices were made elsewhere, this thing would need a back pack power supply to be able to run for any extended period of time. They are using embedded linux. So, they really don't know what they are doing with the micro, and will just try different software out to see what works. Fancy system graphs.. just foolin people to think they know what they are doing. But, its probably good for a bit of fun. Probably no commercial sense whatsoever. Had they have used a low power micro controller, use a real embedded system, and get down and dirty by writing the firmware yourself, and then its worth talking about.
FDA approval should not be a problem for a stethoscope. It is not a food and not a drug and not a mission/life critical device. Do stethoscopes need any certification whatsoever? They might get an endorsement from a prominent cardiologist or a cardiology group/association/college/council but given the existence of _CHEAP_ disposable stethoscopes that barely work, I can't believe that there is any requirement for durability, reliability, or sound quality for stethoscopes.
There are electronic stethoscopes available, but they are expensive and noisy. The ability to electronically produce a clean sound from a stethoscope would have some niche uses, primarily in the educational area. It would be nice to be able to allow a gaggle of interns to hear a patient's heart rather than to have each of them have to listen individually with their own stethoscope. I wonder if the designers have included Bluetooth in the original design. As was already mentioned there may be pros and cons to recording stethoscope sounds; I tend to believe that recording the sounds would become desirable for some doctors. Doctors tend to record and document everything as it is, and this could make it practical to record stethoscope sounds, Yeah, the prototype may be big, heavy, and a battery pig; that means that there are many opportunities to reduce its size, weight, and power consumption.
Although the original design does not include waveform diagnostics, analysis, and display capability, that could be added later at little cost.
Quote from Wikipedia:
Thank God the new stethoscope runs on Linux. Who knows what risks for the patient (viruses, worms, etc.) if it ran Windows Embedded... :D
"Sum Ergo Cogito"
There are certainly more windows-based diagnostic devices than other OSes, but I would hardly call Linux "rare" out on the floor.
Disclaimer : I A A MD.
Currently, any exam can be recorded, stored, shared with an expert for additional advices, compared for reference, you can build libraries for teaching or for personnal reference, you can reprocess these data off-line to "clean them", you can mine them to discover whatever interests you, you can reproduce that data using a different way to display it (using false color instead of greys to help visualise contrasts, or more complex math like using fast fourier to transform Doppler-echo noises into a nice colored graphic) etc...
This include everything from the most high tech devices like 3D scanners, MRIs, etc. to the most simple technical examination (electrocardiograms) or even the most trivial step in examination : overall visual aspect (digital photo cameras are a very valuable tool in several arrays of medicine including dermatology and surgery).
And with all this digitalised information, you can even throw computer at the task of helping the clinicians. MRIs comes easily as an example where computer may help making something useful out of the data. But even much simpler exams like electro cardiogramme nowadays include small embed chip that can automatically recognise a couple of pathologies. Most modern machine will print in nice small message like "90% suspicion of grade I atrio-ventricular blocking", in addition of the actual tracks.
(And in fact modern defibrillator heavily rely on on-board electronic intelligence. The iron-like you still see in movies aren't used that much. The modern ones are semi automatic : you stick two big electrodes on the front and side of the chest. the machine goes automatically an ECG, decide the procedure to follow, and announce on a small embed screen and using voice. The practitioner only has to push a button to confirm and do the shock if he agrees with the machine's conclusions. Everything is automatic and comes with voice instruction so even untrained personal could use it).
Everything recordable and digitilized ? No. Nearly everything.
The century old stethoscope isn't. You use it, you hear something, but you can't share it or store it for further reference.
Meanwhile, there's a lot of information that trained doctor can hear in the stethoscope.
But it's hard to show to other (you can store it in a library to keep it for a lecture, you can't point it with a finger on the graph to a student), the student need luck to be around when there's something interesting for her/him to hear. It can't be reproduced.
It's almost some kind of voodoo art, where you have a couple of old expert who can diagnose almost anything with a simple stethoscope and by looking how the patient walk to enter the room. But they can't easily transmit their art. Sometimes you just wish you could keep them stored in formaldehyde.
This gadget that those engineer are working on is a nice alternative to the formaline. A simple gadget that'll help store, archive, send, etc. the noises you may encounter when examining with a stethoscope.
Also, I happen to be a little bit hard hearing. Not much. I doesn't pose any problem in normal life (I can enjoy classical music concerts for example). But when it comes to hearth noises, I start to show my limits and I suck with a stethoscope.
Such a device could enable me to obtain better (cleaned of noise) signal. Or even, as the devices use apparently open source software (uCLinux), it won't be too difficult to bolt in some crypto+network capability and display a visual representation of the sound on my Palm's screen.
I really think that throwing some electronics inside the stethoscope was long due, and I'm happy that the problem is currently being tackled with open source software.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
I have to agree. Most US ERs have a RN doing triage in any situation besides disasters. (Which is funny to watch the poor internist to route people though a hospital during a drill, they don't know where to send people in a larger facility, 500+ beds, sometimes sending patients to rooms that do not exist.) The RN will take the major symptoms and cat1 the patient if there is any cause is tripped as defined by protocols. However, usually triage isn't as bad as people make it out to be, it is the hellhole after a patient passes the triage desk. Normally, even the issues behind the desk are not the nurses fault. We did a test where we set off several cardiac alarms, which promptly got ignored due to alert fatigue. Alert fatigue is a huge issue within Healthcare, where staff gets desensitized to the myriad of alarms presented to them.
Lets put it this way, how many of us just clicked through a series of dialog boxes to find out we had to start over again or screwed something up? Many nurses and physicians have to deal with a massive series of dialog boxes, blinking lights, colored rows, and auditory alarms. Eventually, it becomes background noise, which is horrible. The people in system design need to do a better job with user interaction to help reduce alert fatigue.
In God we trust, all others require data.
" ... medical devices running open-source software are extremely rare because of the perceived difficulty in obtaining FDA validation ..."
That is totally inaccurate. There are plenty of medical devices in use running Linux, and they all get FDA approval without any issue. As long as you go through the same procedures in testing/quality-assurance and documentation it's all the same end result to the FDA. Even in the case of software which is complex relying on other pieces of software, black box testing solves the requirement for reproducing steps to a particular software operation.
What is it going to take to get those reporting on open source to cut down on FUD?
I work for a manufacturer in the pharmaceutical industry and I hear people say things like Linux and/or Open Source is a problem with the FDA, but I haven't seen anything yet to back that up. Period.
To "validate" software with the FDA, you really only need to do three things:
1) Make your software/system/device secure
2) Have control over the system
3) Document #1 and #2
There are people who come from an old guard that have all sorts of misconceptions about what can and can't pass FDA regulations. It seems to stem from a gross misunderstanding of what the regulations actually say.
I've read the portions that deal with computers and electronic information. There is absolutely nothing in it to suggest open source even remotely is an issue. I suspect this is an urban legend started by companies selling proprietary software.
Regardless, the regulations (Part 11) are the FDA Bible and so long as you can satisfy the requirements in them which basically boil down to the three I listed above, then you are in compliance.
To convince yourself this is true, read this.
So does it run Linux?
that has to be one of the coolest headlines ever.
The ultimate advance in stethoscopes would be one with an integrated Peltier cooler. That way, doctors wouldn't have to store them in the freezer between uses.
Have gnu, will travel.
Folks: this is just a senior design project. No need to get all upset. Notice the link on the page to "other design teams" and the news that they "updated the class."
They have to come up with something to design. And I wouldn't read too much into the corporate sponsorships, either. It's another way for engineering departments to increase their revenue.
What's a gf?
That would be "grip f*ck", as in hand job .
Something which all slashdotters should be quite familiar with.
Q: It will run linux??
A: (sigh)....... yes
Oh ! You meant why *this particular* electronic-stethoscope project chose *a full uCLinux* as its particular software ?
If you need to ask such a question, I think your geek is at risk of being revoked here around at
Now, if you don't mind I'll go back at "Beowulf cluster of Linux-running toasters".
More seriously :
uCLinux isn't completely stupid because :
- it *is* a variant of Linux *designed* to run on microcontrollers. It makes as much sense as custom made C-code. (As opposed on trying to cram Debian inside the stethoscope).
- this kind of controller aren't much expensive either : the controller costs a couple of dollar according to the documents on the web site. It's more expensive than a PIC but its still very cheap and would only represent a small fraction of the final cost.
- Unlike a small PIC, this kind of processor+OS environment gives you much more opportunity to add functions. Bluetooth interface to PDA (for display) or to barcode readers (to quickly load a patient ID# to associate the recording with it).
In short, for only a couple of bucks more (that don't really impact on the final price) there's opportunity for much more capabilities, with not so much additional development cost (thank to the various available codebase for microcontrollers such as uCLinux).
It exactly similar to the situation of microcontroller finding their ways into home appliance for tasks that could be done by simpler hardware (internet-enabled fridge depleting the IPv4 address space is the favorite scarecrow of IPv6 proponents) or into small MP3 player where a PIC+hardware decoder solution could do the trick too.
(And as joked before, there's also the fun factor of doing it too. In fact, I'm just surprised that nobody has gotten yet Linux running on a kitchen sink)
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
Thus it may be useful to have a small MP3-player sized (and functionally not much different anyway) smart device that can fill in your pocket instead of needing to connect a dumb device to a PDA or a PC.
Use the device as a regular stethoscope most of the time and only plug to a Palm or a PC/Mac desktop whenever you need it (for more complex processing or visualisation, and to offload the memory).
Most of the health care personal (nurses, emergency intervention teams, etc.) don't work with a desktop computer near them. There are some conditions also in which doctors don't have a computer handy (when consultations aren't held in separate consultation rooms, but a big room where each bed is separated from the other with curtains to form a small consultation box - in this case the consultation is held at bedside, whereas the computer is on the desk in the doctor's office) or would like not having to lug around one (visiting patient at their home. it's inconvenient enough when one have to lug around a ECG machine. Now imagine having to carry a laptop around for any consultation which would require auscultation - i.e.: almost any except maybe psychiatric or locomotor system related).
Also some hospital are much more low-tech than others (smaller and older hospitals out in rural areas, or western doctors working a couple of month to help in a developing country, etc.). A small standalone pocketable device that the doctor carry with her-/himself will be much more useful than a device requiring a computer available at the workplace.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
"...medical devices running open-source software are extremely rare because of the perceived difficulty in obtaining FDA validation."
Horseapples! I've worked in this industry and that is most certainly not the answer. The reason you don't hear more about Linux devices in the medical field is one of two reasons: 1) The high demand for real time systems, which Linux hasn't been that great at historically; and 2) They're not telling you about it! I've worked in the field developing system software, and not once has FDA validation been the cause to choose another OS over Linux. I'm currently a software engineering consultant, and have worked for many medicial device companies. Most are using Linux, with many of the remainder using a BSD. Even in those companies demanding hard realtime, they're still using Linux in the front end.
(*) Why aren't companies advertising Linux use? Probably for the same reason they're not advertising that they're using vxWorks, LinxOS, QNX, etc.
Don't blame me, I didn't vote for either of them!
New GE CT scanners use Red Hat Linux.
If you consider the level of control free software gives vendors, validation should be easier than it is with non free software. Consider the nighmare of Windoze update and dll versioning as a counter example. A whole system like Red Hat will only have what you and Red Hat has to offer. You can be sure validation testing in the lab matches what's in the field. The amount of control over embedded systems is even greater. This inspires confidence and real validation testing reuse.
Friends don't help friends install M$ junk.
It would be nice to hear , and have the analysis of the arrhythmia so that you could adjust the patients medications. E.G. a patient with congestive heart failure has some symptoms (subtle), and has a slight heart sound that might not be heard with a normal 'scope. Computer analysis might hear some lung crackles - indicating excess fluid, and an increase in the patient's diuretics would help them out.
However, a caveat, most diagnoses come from the history. Maybe 5% of the time does physical exam findings actually make a diagnosis. So when your internist or family medicine MD listens to your heart and chest, it's usually just for confirmation. They've already made the diagnosis.
..........FULL STOP.
bdaa85b841febb3c38c5a447efe7fdf736a22b6d6f0629df456ff015b2c4ea42