Monday, October 21, 2013

"HIT Startups suck"

Sort of a response to this article:

Philosophical debates of healthcare IT (HIT) are always interesting. "Mr. H" had a long diatribe, cutting down and lumping all HIT startups into a single group of 30 year old, over confident, abrasive, swaggering, naive, etc. Now, I personally am not a fan of a bulk of startup culture that chases investment instead of actual customers.

So, I agree on the uselessness of many of those attributes. But, to lump everyone into that cookie cutter is quite dismissive.

Being [over] confident, though, is something you have to be to be launching any type of a startup, HIT or not. It takes a lot of gumption to quit a well paying job with a baby on the way, a house, a wife, and start something from scratch [yes this is me]. You must have the confidence you can do it, or else why would you risk it all?

Being humble is something more founders should embrace, my wife knows I hate talking about my accomplishments to anyone.

The state of HIT

In its current state, HIT is broken. EMR deployments are a pretty big indicator of where we stand. There are quite a few hospitals that are hemorrhaging cash deploying EMRs. Physicians by in large hate the EMR at the hospitals they work at. Find a doctor that is part of a group and visits multiple hospitals, he will be using a different EMR at every hospital he visits.

Best of all, none of the EMR systems actually communicate with each other, no matter what we've been promised by CCHIT and the governments free $44,000. Heck, doctors within the same hospital can't get patient records of their own patients.

Don't get me wrong, we need EMRs at hospitals. But, it is a fantastic indicator that the government 1) had to force them on medical institutions and 2) subsidize the purchase of them. Perhaps without #2, we'd actually have EMRs that people want to use and purchase out there, instead of the daily headaches that HIT staff and users have to deal with.

My own vision of IT is it should make the users life easier, make costs go down, and make the business case that it provides a ROI. This is not the case where we stand today.


But, back to startups in healthcare. There is a massive and distinct separation between what is going on. There are more consumer based digital health startups than you can shake a stick at. The market is flooded with them. But, the number of digital health startups focusing on the enterprise, is a very shallow pool of entries.

The reason for this goes a bit beyond what "Mr. H" claims.

The biggest of all, is compliance. HIPAA and HITECH are immense barriers to entry for those new to the market. Granted, they aren't big hurtles once you actually understand them and know how to properly mitigate them. From my point of view, HIPAA actually isn't all that useful since it is basically a minimum set of what you should do.

Next, are EMRs. Part of the EMR carrot stick was interoperability. Anyone who has done anything with HIT knows what a hilarious joke interoperability is.

With the same EMR issue comes HL7. Out of all things, HL7 has been the biggest headache of all. It is an immense standard that takes a special level of finesse to comprehend. The bigger issue with HL7 is, there really isn't a whole lot of information out there on it that someone in a startup can use to build their own hooks to EMRs. MIRTH is the de facto standard, but as their recent acquisition shows everyone should be looking for alternatives.
Note: I am not saying HL7 isn't useful, it is quite necessary. It is just a lot.

There are no easy to use APIs for HL7, which is the big issue. As a developer by trade, I am used to just dragging a reference library into a project, making a few calls, and being up and running. With HL7, you have to write your own parser. You have to dissect the message, you have to learn the difference between segments, repeating segments, and avoid bashing your face into a desk.

But then, lets assume you develop for yourself the best API out there for HL7 the world has seen. There is no guarantee that it will work from hospital to hospital. In fact, you should assume it won't, at all. This isn't a fault of HL7, it is a problem back to the EMR vendors.

Which, the cynic in me thinks is all by design. If the large EMR vendors can keep any and all competition out, what do they have to worry about as far as challenges?

HIT is fundamentally broken. So, when "Mr. H" talks about writing enterprise systems, there actually is no bar set to hurdle.

Barriers to entry

One of my favorite lines I've heard over and over is
"a low barrier to entry that virtually guarantees that any degree of success can quickly be replicated by competitors."

The author correctly states ideas are a dime a dozen, but it is execution that matters. I know exactly zero founders who are concerned about this.

If anything, it is welcomed because it forces you to push forward.

One more thing

if you are new to healthcare selling a product whose target customer is hospitals, it’s a near-certainty that you’re going to justifiably fail, with your only hope that a better company will buy yours before it splats to the ground. That’s not anti-innovation at work – it’s the reality that healthcare attracts a lot of flaky, poorly thought out startups that don’t deserve to succeed. Come back when you hit $1 million in annual revenue.
This piece is a bit much. The author's stream of conscious diatribe ends on an extremely negative tone. Instead of being a Debbie Downer and hoping for destruction of people pouring their lives into a company, you should offer actual advice to them.

The issue I have here is the two words "justifiably fail." Being new to HIT doesn't mean your idea is bad or you deserve to fail. You should fail only because you are new to the field? Please. That isn't anti-innovation, it is just vitriol.

Sure, the harsh reality is that it isn't easy. But if we want HIT to actually solve issues and make the lives of everyone using it easier, everyone needs to work together.

Yes, there are a lot of bad ideas out there [not just in health care] that deserve to fail. If we applied "Mr. H's" standards for startups to EMR vendors, it would be quite an empty field.

Enough complaining

"Mr. H" is correct about a few things though.

The enterprise sales cycle is quite long. It isn't as simple as buying a few Apps on the App Store and being up and running that instant (But, hint: physicians are doing this anyway and are violating HIPAA and BYOD policies which hospitals have put in place). I come from building secure systems at the Department of Defense and Department of State. A 18 month sales cycle, was considered pretty quick for anything other than a 30 person pilot.

Understanding HIT problems is another. Enterprise class issues that hospitals face aren't out there in the open. I am a silent partner in another HIT startup and the CEO is a physician. He has been an amazing help in getting me to understand the depths of the issues which face physicians and nurses on a daily basis.
As a note, our sales cycle with that product is short and long as it sells to individual physicians, groups, and hospitals. We've been chasing a few hospitals for the last 12 months and they are finally jumping on board. 

Who am I?

I am one of these 30 year old HIT startup founders. I earned my chops building highly secure enterprise systems for the Department of Defense and Department of State.

Since 2011 I've had two software acquisitions, I sold my first piece of mHealth software to Walgreens in 2011. In 2012, a piece of mobile security software I wrote was acquired by the the Canadian mobile security company Fixmo. Since 2009, I have been a silent partner of another mobile health startup that in the last year has gained crazy traction.

I don't give the "Steve Jobs" wannabes the time of day either, but those are few and far between. Or am I just hanging out in the right circles?

But, I have met a lot of great HIT startup folk the last year or two. To dismiss the lot of us because you've met a few immature brats, well, you'll be in for a surprise.

1 comment:

  1. HL7 is a lot, because healthcare covers so many things. And yes, HL7 has lots of optionalities, because healthcare isn't standardized. The debate around "contents of an electronic record" have been debated by clinicians since the early 1960s, and there still isn't any agreement.

    As for HL7 tools: have you tried looking at ? There are tons of tools, with HAPI being probably the best freeware API around for HL7 v2.