[anon4this1]

As a Doctor, the potential for improvement begins from very basics. For instance - a radiologist sees a funny looking line on an xray and reports "this is probably nothing, but we should repeat the xray in 6 weeks to make sure." Sometimes these requests slip through the cracks and in a fraction of cases the patient presents 2 years later with terminal cancer from their metastatic lung Ca.

Ensuring blood thinners are prescribed post coronary bypass surgery or stenting is another one. It should happen every time, but sometimes by human error it may not.

A good startup idea would be to make a hotlist of the top ~1000 of these obvious, preventable errors, combine IT data sources to predict when they might be occuring, and have checking systems in place to prevent them. If it worked even adequately, hospitals would be seen as negligent for not using such a system.

[charlesism]

I'd be interested in your opinion on what errors could be prevented by intervening less. With all the legal, financial, and emotional pressures for doctors to "do something," I'd be surprised if the most common errors are errors of neglect.

[PavlovsCat]

A lot of pressure on doctors and nurses (in hospitals) is due to the fact that there are a lot of actual and potential patients, but not so many beds and people to take care of them. So corners have to get cut, less is already being done.

[charlesism]

My reasoning is that it's probably easier to harm someone by botching an operation, or by administering an overdose, then it is by their not having the surgery at all, or skipping medication. Not that either case is ideal, of course!

No doubt doctors and nurses are overworked. It occurs to me that our culture may set the bar too low on when to intervene. Things seem to go wrong more often than we like to acknowledge.

Heck, just sleeping a few nights in a hospital, without any treatment, already exposes a human to unhealthy hospital food, infectious disease, and a small, but non-zero, risk of a nurse mistakenly pumping you full of insulin.

[ch4s3]

Then you would be very surprised. Patient data is fragmented and doctors are overworked, accidentally missing something routine happens constantly. But unnecessary treatment happens too, if you're interested in that, read up on Choosing Wisely.

[rzzzt]

There was an article making the rounds here on HN about medical checklists, similar to what pilots use for takeoff and landing procedures, for example. Can't remember good keywords to search for, though...

[hidroto]

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

Another example of doctors advocating checklists.

[maxerickson]

This article is by the same author:

http://www.newyorker.com/magazine/2007/12/10/the-checklist

I like the quote from the doctor doing a lot of the work with checklists, answering a question about when they will be widely adopted in medicine:

“At the current rate, it will never happen,” [Pronovost] said, as monitors beeped in the background. “The fundamental problem with the quality of American medicine is that we’ve failed to view delivery of health care as a science. The tasks of medical science fall into three buckets. One is understanding disease biology. One is finding effective therapies. And one is insuring those therapies are delivered effectively. That third bucket has been almost totally ignored by research funders, government, and academia. It’s viewed as the art of medicine. That’s a mistake, a huge mistake. And from a taxpayer’s perspective it’s outrageous.”

[DanBC]

There's the WHO Surgical Safety Checklist.

http://www.who.int/patientsafety/safesurgery/tools_resources...

Here's one article about it: https://news.ycombinator.com/item?id=436575

[nchammas]

It was probably about the work of Peter Pronovost, or something closely related:

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

[api_or_ipa]

As a side note, I've long advocated developers use checklists during releases and more involved git tasks. Why? The cost of using a checklist is minimal-- maybe a few moments glancing at a sheet of paper. The upside is huge though, since you can avoid catastrophic error.

I was fortunate enough to earn my wings before becoming a developer, and as lame as it sounds, learning when and where to use checklists is one of the best lessons I learned.

[Retric]

"Anything that can fail, will fail."

If someone is getting a procedure, it seems like they should automatically be prescribed required medication unless they have been flagged with an adverse reaction. Sure, review this stuff, but the default needs to be less harmful than the supervised action.

I am guessing there are a lot of historical and cultural reasons why this does not happen. But, why can't we apply basic automation.

[amenghra]

A good database that patients can access might also be helpful.

Until you get something fully automated, you have an extra person to understand/review decisions.

[lostlogin]

Unfortunately hospitals don't have "a database" there are often dozens (maybe more?) of vaguely connected systems with numerous logins required. Getting the one thing you want (e.g. patients phone number) can mire than one computer and several programme logins. Having your accounts randomly blocked and calling IS to unblock them isn't uncommon either, same goes for random outages of periods of painful slowness. Maybe it's just the system I'm in?

[TheSpiceIsLife]

Last time i was in a treatment room at a hospital (broken thumb) I was able to open a browser on the PC before the doctor came in a told me off.

He told me his group of physicians had refused to treat anyone until the IT department removed the requirements for login to basic systems because they were sick of not being able to use their tools because a login didn't work for whatever reason - forgotten password, login system not working, it was all the same to them, they just wanted to be able to access their tools.

That was in Adelaide, South Australia, about 4 years ago.

[chris_va]

Selling IT to hospitals is very very very painful, just an FYI for anyone here thinking about this line of work.

[lostlogin]

Just to add to this as someone on the other side of the equation. Buying health related software in a fairly corporate environment is the most painful thing I've ever been involved with. Accountants and IT departments who range from unhelpful to hostile and have ridiculous demands. This for a small system in a very small department.

[jchoca]

One of my friends started a company that basically does that. It's called Qualaris (http://qualaris.com/).

[yitchelle]

'A good startup idea would be to make a hotlist of the top ~1000 of these obvious, preventable errors.'

If a hospital is run like a start up, the possibilities are rather mind boggling.

[Noseshine]

Where exactly is the appeal?

A startup is an experiment, most of them fail. It will probably also have an oversize share of people who don't mind risk but do like (monetary) profits, very much so. This means they will take risks with patients health and lives that we have already chosen not to accept except in cases where bad outcomes don't matter (patient is dying).

The "politics" can also be a lot worse in a fast-growth company with a lot of success-driven people. The risk of losing out is a lot lower in a "boring" company where everybody just gets their salary and advancement is slow and predictable. This leads to less focus on the product/service, because people are occupied with "positioning themselves".

There will also be investors, and they don't really care about the product at all, or about long-term success - as long as they can unload their shares at an IPO like the hot potatoes they often are they are fine. Which is not so because people are evil, but because the selection function is set up in a way that favors this behavior.

In a sense though you are right, the possibilities are rather mind boggling indeed. I just don't think in a positive way.

[fzzzy]

Fail fast?

"He's got a cold -- let's just kill him"

[cmdrfred]

Perhaps you can cater to the hipster demographic.

"Startup hospital offers artisan cardiac catheterization"

[icantdrive55]

'A good startup idea would be to make a hotlist of the top ~1000 of these obvious, preventable errors.'

I think that's a great idea.

I don't think it needs to be a start-up, but maybe the AMA, FDA, or whomever could put together a simple list of medical preventable errors, and publish that list in a easy to read format.

Just put the information out there. After my bypass; that will be the first book, or website I'm looking for.

Someting like, "After the first heart attack, an aspirin might prevent the second heart attack."

Something that would legally protect the Doctor.

It seems like the entire system is so afraid of lawsuits they just give the bare minimum of advice. And then there're the doctors who just don't really care.

[sandGorgon]

Isn't this already done ? I thought that the us had a voluntary , anonymous database of medical and airline pilot errors

[samesense]

Yes. I've seen several examples of decision support since 2010. It's not easy to train models because the codes for procedures can be ambiguous. Furthermore, since hospital classifiers favor medical error recall over precision, docs can be swamped with so many warning messages that they tend to ignore them.

Take adverse drug interactions as an example. The training data for drug interactions mostly come from adults, so the resulting models do not apply in a pediatric setting. When the models are let loose in pediatric hospital, a high percentage of the drug interaction warnings are false positives, so these type of warnings tend to be ignored.

It seems that the trend is to use decision support with a lot of human oversight and investigation of the raw data to see if the model conclusions are correct.

[cmdrfred]

What you describe to me seems to be a simple UI problem. I'd add a button "False Positive, Don't show this warning again" and flag the "drug interaction" for review. If I get enough of these, I'd change the behavior for all users in the next update.