Hacker News Reader

My unsolicited advice: rather than being "fairly certain" that you can make $500/mo with your idea so that you would prefer to start a company first, I suggest that you reverse the order of operations.

First, be certain that you can make $500/mo by building it yourself and get the hospital paying $500/mo.

Then, if the opportunity looks good, start organizing (aka building an organization, a company) around that.

You'll short circuit a lot of pain that way.

Motion seconded.

It sounds like you're seeing a lot of problems in healthcare. That's good you're noticing them. It sounds like you also want to solve some of them, mostly because you're fed up with them being problems. That's also good. The best kinds of solutions often come from scratching your own itch.

Another similar recommendation: get more specific about what you want to solve. "Disrupting healthcare" is a platitude anyone can agree with because the news runs a segment every day about inefficiency in the health system. If you see 10 problems, does 1 stand out way above the rest? Work on that problem.

Just building a quick & dirty solution and shopping it around like smalter suggests is the best thing you can do to prevent working on a problem you'll burn out of in 6 months. You'll very quickly realize whether you really want to do that thing or not.

I'm taking your advice and I'll hack together a prototype and see if the administration will bite.

> First, be certain that you can make $500/mo by building it yourself and get the hospital paying $500/mo.

I'm currently chewing on ideas from "Lean Startup". It suggest to try and validate your idea before even building a product, for example with a interactive prototype. If the OP has a good contact with his hospital, he can show the prototype and ask: "If I build this, will you buy it? Will you pay upfront for a discount?"

I highly recommend reading Steve Blank's "4 steps to epiphany".

The guy is a first year resident and should be working 70 hrs a week perfecting his skills. If he is cheating the system now by not being fully committed, I can only imagine what he will be doing 7 years from now.

Since when has the technology community at large considered automating menial data entry to be "cheating?"

Hospitals spend hundreds of thousands of dollars training residents (med students learn practically (as in practical knowledge) nothing in med school). Therein lies the cheating.

If you gain 10 minutes/week for every surgeon in USA, then you've done more healing than you could by simply being a great surgeon for your whole life.

This guy has one shot at learning medicine, his residency. If he wants run a trivial startup, so be it. But his first responsibility, at this time, in his training, (no he is not really a doctor yet) including his residence and fellowship.

His first responsibility is to whatever he wants to make with his life, not neccessarily his training. If he wants to run a medical startup that you call trivial, it's just as good (if not better) way to go as focusing on his residence and fellowship.

The default career directions for anyone, including medical residents, are just that - simply defaults, not some oath or moral obligation to follow that exact career choice 'till death do us part'.

Have lower level people, such as a Medical Assistant or a 1st year resident do the data entry.

ahhh, feels good to do more healing than a great surgeon, before I've even had my Sunday coffee.

Unfortunately, the note that has to be entered in the EMR has to be started and signed by a fellow (a person who has finished residency but still in training before becoming an attending doctor for those who don't know), and this is the note with the data, and so I (a first year resident) would have to be sitting next to the fellow, he would have to start the note, then I would have to switch seats with him, do some data entry, then he would have to write his impression ("normal", "COPD", "asthma", but in more doctor like terms) and sign the note. The way the system is set up makes this impractical.

Isn't technology meant to make our lives easier? If this can increase the productivity of doctors, I reckon it's a good thing. That way they can spend those full 70 hours a week perfecting their skills, instead of wasting 20 hours to enter and retrieve data every week.

Why don't doctors have clerical staff to do the data entry?

They do. Both do. No one is immune to data entry in the medical world.

depends on where you work and how much staff is hired for this type of work, but with the way the EMR is set up, and who legally has to start and sign notes (a doctor), you cannot avoid some data entry unless it is somehow automated for you

exactly what I'm trying to do

Residents don't need to spend every waking hour focused on their job. As long as he is doing his job properly, what's wrong with doing some coding / business on the side?

A First year resident know jack shit about medicine and works 13 hrs a day making mistakes along the way in hopes of learning enough so not to make these same mistakes when they will be in command (5 years from now).

You are very sure of how he should learn and contribute, why is this?

Working Smart > Working Hard

If a Dr. sacrifices their productivity for a short time, with the result that he increases the productivity for all Dr's; then the Dr. has done the opposite of cheating.

>I can only imagine what he will be doing 7 years from now.

OP could follow your advice and be an average Dr. Or OP might actually succeed in improving wasteful processes in hundreds of hospitals.

> should be working 70 hrs a week perfecting his skills

I already work 60-80hrs/week, should I be working more? Would 100hrs/week make you happy?

> I can only imagine what he will be doing 7 years from now

6 years from now if I continue on this road keep my head down and work hard I'll be making 300-400k/year as an attending physician. I can work 4 days per week and make 200k. I don't care too much about money as long as I can pay my loans off and eat out once in awhile. That's all nice and dandy but I'd rather spend some of my free time building something that helps the healthcare system.

I'd suggest instead talking to the people who actually make buying decisions, and find out what other purchases they have approved recently (and how many other people needed to give approval -- e.g., the IT department), and why. Unfortunately I suspect the news will be discouraging.

Don't just try selling something that interacts with medical data unless you have serious quality control (you may need various ISO certifications in place) and a good grasp of privacy law. That's what many of the frustrating hospital purchasing rules are trying to ensure.

The reason that the healthcare world has so many problems with obvious software fixes that haven't been fixed is because they're very rightly paranoid about using unsafe software, because sometimes it kills people even when it's designed by a corporation with lots of quality control (cough cough THERAC-25), and not just a side-business with a doctor and a hired coder.

I will be talking to the purchasing dept after I build a prototype to show the director. I am aware that this will be an uphill battle since everyone in healthcare is paranoid including doctors and admins.

Your advice is duly noted. I am planning to talk to the director of the pulmonary team sometime this week and see if he is interested. I know most of the doctors who work in the lab will welcome the software. Seeing that most large hospitals pay 50+ million for their EHR/EMR setup, $500/month sounds reasonable.

I looked briefly into building it myself (just in case this thread did not get a good response), and I can parse out the text using the pdf-reader gem and now I just need to use some regex to match what I want and then stick those values into a report.

Thanks again for the advice.

> Seeing that most large hospitals pay 50+ million for their EHR/EMR setup, $500/month sounds reasonable.

When I worked for a construction project management startup, this too sounded reasonable: many software systems for this audience were in the six-figure range... we were only asking $100/month or so. ("And a free trial!!!")

We learned that when your monthly price point is the size of an accounting error, no one can take you seriously.

This software was also holistic in nature: everyone, everyone, everyone has to use it for it to work well and as a core part of their organization. It's a superset of the marketplace business problem.

To address these we had to charge more. Much more. This raised the stakes on both parties and also filtered out less serious buyers who would probably fail with our software.

The construction companies who were serious buyers really considered thoroughly whether they could implement the change; we worked with them for, sometimes, months going through how the change would take place. And training, training, training. Usage metrics. Account hand-holders. Keep them using the system.

--

As others have said, medical software is some of the hardest to break into. Medical staff HATE CHANGE, and need to. Change in the short term leads to mistakes and mistakes cost lives or careers. The system is bureaucratic, politicized and slow moving. Anti-disruption.

I believe that a guerrilla approach to insurgent medical software is ultimately what will work - things like https://www.radiologyprotocols.com/, where a radiology tech wanted a common repository of knowledge for others in his field. Then, through word of mouth and using it with his colleagues, it gained first use in his hospital, then international use, and now it's starting to take off in the US. (Reminds me of "Big in Japan" first, or conversely Japanese artists having to become popular in the US before being taken seriously back in Japan.)

--

Okay, advice.

Sell it before you make it. If you have multiple hospital/medical contacts, play with pricing. Try outrageous things. If not, just try to work out any sort of deal and have them sign a piece of paper that says they'll pay.

Then build it.

Iterate with the first customer until it's "So Good They Can't Ignore You".

After this, you might get better traction outside the US. There are hospitals in other countries who are desperate for good software and also don't have the money for the large EMR software many US hospitals use.

GOOD LUCK!

I've done some work in clinical research and I can tell you that you might need a legal cofounder more than a technical one! I had ways to automate dozens of employees jobs away just by integrating data systems, but in the eyes of the hospital it was too risky. They didn't want to violate HIPPA. I wrote a little thing about it for The Scientist: http://www.the-scientist.com/?articles.view/articleNo/35249/...

I'd love to talk to you more about this because these are problems I'd love to see solved. I think its a real up hill battle though.

It seems liability would be another reason to have a firm legal foundation. You wouldn't want to miss an edge case that results in a patient's death. If an investigation showed that your translation software made a mistake that led a doctor to make a bad decision, you'd be in a pretty ugly spot.

This is part legal, and part QA, I believe.

Yea, this is why I didn't want to start a company in healthcare. It's ugly.

I think I can avoid the whole thing by having my software just be a tool, not a final read, and doctors will still have to double check the numbers against the PDF hardcopy. It will just take 1 minute or less to glance at 10 values or so and make sure they match up compared to typing it in themselves

I would generally recommend against getting a legal cofounder, even in the medical world. They're not contributing nearly as much as the other cofounders in the beginning, especially when you're building a tech-focused startup.

Make friends with a lawyer. Read up on everything legal you can. Pay your lawyer friend when you need lawyer advice and lawyer action. But don't give them not only full-time, but founder, level status unless they are truly a full-time founder.

Plenty of others share their early-stage lawyer advice freely so startups don't need to get in this situation. Check out:

* http://startuplawyer.

* http://rockhealth.com/resources/startup-elements/hipaa-compl...

* http://rockhealth.com/resources/startup-handbook/

Help me find a research position! I am a fresh MD graduate. Thanks!

As a physician of 15 years in internal medicine all I have to say is residency must have become ridiculously easy if you have time to start a company.. For me, 100 hours a week made that impossible so I waited until I was done. Your idea is an old one. It's been deliberated and contemplated. You forget that you're writing code that is essentially a medical tool that potentially affects peoples lives. You will need massive insurance, and lots of regulations to contend with. You will take on all liability with regard to patient outcomes. In the end, it will cost you 10x more than 500 a week. I don't even consider healthcare with my coding projects.

I would think automating it would vastly reduce the likelihood of error vs. manual entry.

It's a different type of error that's possible, though.

Manual entry errors happen at predictable rates, and with good processes you can get them down quite low.

Software error rates are not so predictable (unless you have full control over all variables... and that's surprisingly hard to get). You could deploy this software in the context of a slightly different version of the radiology software which used a different font in PDFs (or whatever) and find that all 8's are now read as 3's; no human would do that.

If a human started getting input that wasn't sufficiently readable, they'd talk to someone about it. Could the software do that?

Bingo. This guy should quit commit or quit.

Hey there! Do you have a research position for a fresh MD graduate? Thanks!

I know one registrar who built his house while doing the program. I know another who got a few medals sailing whilst doing his etc etc. I'm not sure that you can generalize about how much time people have.

Two pieces of advice from a very drunk technologist.

One, never approach a community that lives and breathes startups using words such as "fairly certain," and "try." They have no power here.

Two, start the company, make a respectable business plan, and then whore yourself out. No one is going to be interested in a joint venture with a few paragraphs of information to go on.

You don't have to be perfect, nor do you even need a full roadmap of ideas or plans, but people need to see that you're willing to put effort into it. No one makes a successful startup, especially in the medical field without considerable research and effort. If people don't see that you're willing to put in the time, they'll never get behind you.

can't agree more

That's not something I would call "disruptive". That's an improvement over what currently exists, but it's trivial and doesn't disrupt anything in any way.

I was thinking the same – is 'fix/improve' a new meaning of 'disrupt'?

I agree. It was late at night and I couldn't think of a shorter headline. I apologize.

I would be willing to bet the hospital you work at is paying several million a year for their EMR and all sorts of smaller contracts that probably nobody knows about. One of those contracts probably supports the integration you are looking to do and has not been configured yet because it has not went high up enough the food chain. I think you are better off asking contacting the IT department of the hospital and have them log a ticket to the EMR provider (probably Cerner or Epic) and have them look at it before getting too invested.

> have them log a ticket to the EMR provider

Good advice. Although sometimes, the best way to get a fire lit is just to set something up and wait for IT to complain...

I second just telling the IT department about it so they can tell the vendor. I work for an EMR company and I'm pretty sure UCSF is one of our organizations. I don't know if the hospital will pay you extra for an interface solution that the EMR company is supposed to take care of. I'd check with your hospital first.

Good advice. I heard somewhere that they did have this ability but it was a very pricey add-on. It was just a rumor though. I'll check with IT.

Founder of a SF-based healthcare-related startup here. Emailed you. I'm not looking for a job but happy to help how I can.

EDIT: To anyone in the Bay Area (or really anywhere) also thinking about joining this industry, feel free to reach out to me too. My cofounders and I have learned way more than we want to know over the past few months about the healthcare industry + the startup side of things and are happy to pass on our knowledge.

I'm based in Singapore, Asia. Will be happy to share too what we've learnt about building a B2B platform for private clinics to order medical supplies.

Hi, what's your email? I'm a recent college grad who deferred my med school for admission for a year to learn Rails. Both of my parents are physicians and I've been working in a clinical setting since the age of 15. I'm seeing many opportunities to build tools that help improve physician workflow and I would love to reach out to you for advice. Please let me know if you're available for a short chat. My email is fatoki09(at)yahoo(dot)com.

My email's in my profile. Just shot you an email.

Is the time spent entering in data considered billable? If yes, then I doubt the hospital will be interested. If no, then you may be onto something.I was in a presentation not long ago at my former employer and we were being pitched a pretty cool piece of software that would probably have saved me a couple of weeks of tedious interface building. I pointed out to the VP how much time I would save. He said that we bill by the hour. I don't work there anymore.

Hospitals don't bill per hour; they bill per procedure. Data entry isn't a medical procedure.

However doctor's compensation is measured in a variety of ways depending on the hospital. Some doctors might get paid hourly or by a fuzzier metric of how much time a procedure takes, so they may not want to use this type of system. Similarly, hospitals might be able to bump up the procedure if it takes more time (ie is more involved). For example, a doctor who has to review charts and do some data entry might be able to raise a standard level II checkup to a level III one because they had to do more work. (checkup names are made up, but you get the idea)

Hi, I'm a 3 time healthcare entrepreneur (10 years) who has been down this path, and many like it, several times already. The problem you are describing is well known and in your hospital's IT queue, just so low on the list to be invisible.

As others have rightly pointed out, the technical issues are perhaps 10% of the problem. On top of that, because each EHR is configured so differently, your solution will be a consulting effort not a product (read: expensive)

My suggestion - thoroughly evaluate how difficult this would be to get approval for at your hospital. If you get past this stage, talk to other hospitals. Learn what the implementation process will be at each hospital. This will likely dissuade you but, if not, you've found a niche.

The reality is that this is not an IT problem but a process flow problem. Somebody with a lower pay grade should be entering this data, say an MA or a 1st year resident, until the IT team can prioritize this known issue.

Agreed.

What this means is that usually in the medical world you can't just build a product or service and sell it to EHR/EMR providers. Their (e.g., Epic, eClinicalWeb) customer is hospitals and clinics. So you need to get hospitals and clinics to demand/buy what you offer. Only then EMRs will pay attention and only because increasing amounts of their customers are knocking down their door asking for your product.

Very rare is the case where an EMR will see the market opportunity you're carving out and buy your product. They don't care. They'll either buy your company once you get big enough to represent a sizable addition to their revenue (single millions is not even close) or just build an imitation feature once they're forced to.

Just to add to what you are saying, it isn't just docs wasting time with systems that don't play nicely. Radiology is a hotbed of IT chaos with dozens (zero exaggeration) of systems that just do not work properly. Cardiac CT reports, MRI Spectro, CT dose reports, RIS systems and the PACS (there are about 20 data points on this one alone) etc. Basically anything that has any post processing done is also displayed horribly by viewing systems. Comparing what I see when making reformatted images to what some poor doc gets when I send him the reformats is very depressing. We have a radiologist who codes - he is mostly mortal unfortunately and has a finite amount of time is his day. Despite this he has made some amazing improvements in crap radiology workflow practices.

Hi, Would you be okie with sharing some specific problems and painpoints over email (but leaving out confidential details)? ryan.teo@gmail.com

  We have a group of friends who are familiar with healthcare and have programming backgrounds.

Ryan

Forget coding for health care. MD and coder for 15 years. You will make 10x the profit if you avoid the regulations and liability of health care software.. In addition, I'm not sure you have any idea how little few time you're about to have.

I've struggled with this for awhile. I've had some thoughts about improving medical education that would avoid the legal hassles of healthcare and generate a good income. That idea is for another time though.

Sorry to piggy back on your post. But Indian Medical Doc here pursuing Orthopedic Surgery Residency who also codes in Python, Django, Frontend and Excellent in UI/UX is also looking for some help !! If there is anyone in the audience who is a full stack python guy located in India or USA and can help me disrupt Health care Kindly get in touch.

Have lived and worked in some of the finest hospitals in US (viz: Ny Presbyterian and KCH brooklyn) and also in rural India this has given me a very good understanding of health problems that can be fixed with technology. Have already done some good amount of work and things are already in place. Need someone who can complement me with the right skills.

HealthEnclave.com

Am looking for some one good with-

- Django / Python / Flask - Tastypie / Django Rest framework - Fronend Desing (HTML5, CSS3, Jquery, Bootstrap, Less) - MongoDb / PosgreSQL

additional skills: (not necessary but it would be great if you have 'em)

- Objective C (iOs dev) And JAVA for (Android) - Some understanding of ML , Data analytics - Electronics prototyping with Arduino

I am familiar with and have some amount of practical working knowledge of all the above listed technologies.

Thanks Asingh [at] HealthEnclave [dot] com

Hi Asingh,

  I've dropped you an email. We're based in Singapore, but we're happy to work together.

  Btw, is HealthEnclave meant for wellness or do you plan to develop it along a more serious angle?

Ryan

What is Health Enclave? I can't figure it out when I look at the site.

Looks like a content farm?

Well it's just Version 1.0 needs to be improved much much further. Content is just a small part of what I am trying to solve. Get in touch if you are interested

Your a first year resident. How about finishing your training? This sickens me.

You should not be making assumptions about people you don't know.

With that said I am not trying to make a quick buck here.There are real life pain points that I am trying to solve. The longer it takes for me to create my vision the longer people will have to live with pain, suffering, anxiety (because of lack of proper information to make choices upon).

If a good surgeon who took Fees for his services die tomorrow then it would not matter because his patients who just go to another surgeon. But if this doc creates something that solves a real life issue that no one is able to solve or maybe he creates the best one.. Only then would IMO would it even matter that this person even existed

Unless you have paid for his training, why does this bother you?

> Many of the problems in healthcare involve two systems that need data from each other.

This. From my limited experience, even with all the systems able to talk HL7, apparently what message segments the system sends, the message segments the system reads etc. are different from system to system. The result is entire subsystems or heavy customizations for redirection, massaging, and transformation of messages. Something to think about systems integration.

Another approach is to buy subsystems from as few vendors as possible (naturally subsystems from one vendor would all talk among themselves very nicely) - but that is obviously not without any repercussions.

I'm a pharmacist based in Singapore working in IT side of things for the pharmacy department in a public hospital here (means: pharmacy system, EMR, inpatient automated system, decision support). Still learning but would be happy to keep in touch.

I started my career making medical devices talk with PCs back in the 80s. Sounds like not much has really changed. Usually these machines have a serial protocol or in newer machines, a webserver and API that can be used to get data in a better format. You are lucky when you get good documentation. Sometimes, the vendor sells PC software as an add on and doesn't want competition.

This problem is probably a goldmine if a decent middleware could be written that grabs data from the device, transforms it to a standard format, and then loads it to EHR/EMR/HCM whatever software the hospital is using.

Here is one question. How will you do with all the tape? This creates a potential lawsuit point in case of failure (or even in case of blamestorming, when the software worked correctly but a scapegoat is needed). So you will need some form of insurance or regulatory approval.

I can create a very few extremely simple solutions in my mind that will be quite precise. You don't even need much coding. In the worst case OCR can do with simple extraction script if the PDF is too unruly. And are a weekend job. But selling it to someone will be a pain.

He'll need to carry E&O insurance if he ever gets a client. He'll need to keep that insurance while the product is in use.

Get out the US. Other places have less crap in the way and have better performing services (ie their lack of regulation in the IT area is actually capable of producing better results). Are the regulations just protecting incumbents? Or am I overly cynical?

Any software that handles/transfers this sort of data is going to be (to the FDA) a Class II Medical Device. It seems like there would be a overwhelming amount of work for a one or even two of you :/

http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRS...

The MDPnP group has done a lot of work on interoperability and securing interactions. I don't know how much device-specific information they make available, but some of the testing procedures could be helpful for validation:

http://www.mdpnp.org/

What I understand that Disruptive process means that you are to offer inferior product for a price that matches the quality. You work hard and grow your customer base and after a while you can go after higher quality and more lucrative layers of the product market.

What you describe is an incremental improvement.

As a responsible hospital, I will have to ensure your software will work 99% of the time and that if anything goes wrong you'd be paying the loss. That's a dangerous field to get into.

That's one of the reasons why it matters so much that the guy is actually a doctor. Not only does he have technical expertise in the target domain, he already knows what's going on with safety regulations, liability insurance and suchlike.

He is a first year resident with 3 months experience. Trust me, he doesn't know jack.

It would have to be better than that. A 1% failure rate would give several errors (and potential lawsuits) per week.

Yea, it's an messy area to start a company in. I'm trying to learn how to cover my bases as I go along but if I'm paralyzed by fear forever I'm not going to move forward.

I'm from India but can understand where you are coming from. Think it a useful idea that can work. Let me know if you want to explore.

As a fellow doc, i say welcome! Have you looked at DICOM? Most medical s/w, h/w support it. And congrats on the radio residency!

Yea DICOM is for imaging though (CT scan, MRI, etc ...) but I think you know that already. Even though DICOM is a standard, there are still stupid things that go on (e.g. radiologists have to hand type radiation doses that the patient received because apps do not talk to each other).

HL-7 is the standard for other medical apps. Even though these standards exist, the problems are still there and solutions are sometimes too pricey for hospitals.

Thanks for the encouraging words ... it's been a long road.

Some systems automate the sending of dose data, but send it as an image with the data hard burnt into it. Amazingly crap. I highly doubt that radiologists are recording dose as they very rarely do any examinations that give dose. It would usually be radiographers/Medical radiation technologists/Technicians etc (pick your naming convention!). DICOM is capable of handling data other than images, I believe some neurology systems produce data into the format. It has a ton of fields that seem to be able to be filled with user defined data and yet not break the DICOM spec. There are many things about hospital IT that are way more crap than hand recording of radiation doses.

Thanks for the informative reply. It might be a long road but it's a part of your journey. Enjoy every mile.

Hi,

  We're based in Singapore and previously built www.MedF1.com, a B2B platform for medical supplies ordering for private clinics. We did a health report generator from some Quest Diagnostic lab results previously for clinic chain in Singapore. Dropped you an email - ryan.teo@gmail.com.

  Interested in work with healthcare disruptors, we can design the software (mobile, web) and hardware (Ardiuno, Linux).

Best, Ryan

This doesn't make sense. He says he is a first year resident and is starting a new program next year?

Radiology residents do a 1 year internal medicine residency first and then they enter their radiology residency. Makes perfect sense.

This question clearly reveals your lack of understanding of the US medical education system.

I hate responding to trolls, but your comments through the thread have been a bit abrasive to me. Accusing someone of 'cheating the system', saying someone has to work 70 hrs a week, only having one shot, saying they should quit, telling them they must work 13hours a day to learn.

I'm assuming you're in (or rather I hope you are) medical school and not just someone applying. Here is the rub. Some people don't have to try to do well. There are people who never studied for the MCAT and scored 36+. They didn't have to really try to do well.

A radiology at the place he mentioned is very competitive, so it is safe to assume this guy was at the top of his class. The intelligence/knowledge difference between the top of the class and the bottom is enormous. Not everyone is the same. Don't fault him for what he wants to do in his free time.

Thank you for writing this. I hadn't taken the time to look at the usernames among all the negativity in this thread and realize much of it was from the same person.

Side note when looking at your profile: looks like both hnofficehours.com and hackernewsers.com are down/no more? Dead links.

Thanks for this buddy.

Is there some reason the pulmonary machine doesn't do HL7 or equivalent? Is it just not configured?

It might be because the machine is old. Hopefully, it does have the option to generate a CSV/ XML file. From the description, it does seem to have some kind of computer connected to it and a printer.

Otherwise, you might have to parse the data from the serial port.

Yea, I have to look into this, but even if it does do HL7, you need to write software to extract what you want and present it in a meaningful way to the other application.

I would be interested in this, but I am from Europe. I know my way around PHP and RoR.

I'm a doctor - help me disrupt healthcare