Engineers are King

It's a well known fact that talent is the only asset that matters in early-stage technology companies. This is particularly true in software companies such as Pristine.

I just wanted to take a minute to write that down and make it public for any current or future Pristine engineers to read.

If you ever feel that I have forgotten this fact for even two seconds, please shove this blog post in my face and call me an ignorant hypocrite in front of my employees, investors, and publicly on the Internet.

I will thank you.

There Will Be Two Kinds of People in the World...

This post was originally featured on HIStalk. 

…those who tell computers what to do, and those who’re told by computers what to do." - Marc Andreesen, Andreesen Horrowitz.

I have a lot of favorite VC quotes. This one is without a doubt the most profound.

Every modern EHR already supports task lists, messaging, push notifications, and alerts. Medical professionals — doctors, nurses, technicians, therapists, etc. — are logging into EHRs to see which patients they need to treat, where to go, and what to do. Most medical professionals spend hours looking to a computer to know what to do next. The computer guides them through their day.

The vast majority of medical professionals fall into the latter camp of Andreesen’s quote. This was true in the analog era, too. Schedules, sticky notes, and charts drove what and when medical professionals did things instead of computers.

Looked at under another light, doctors, especially surgeons, are revered for how many years they’ve been practicing and how many of a given procedure they’ve done. More broadly speaking, people who fall into the latter camp of Andreesen’s quote are people who repeat processes and steps for a living. I can’t find any statistics on this, but I don’t think it’s unreasonable to assert that 80 to 90 percent of the workforce performs an intrinsically repetitive job. Perhaps 10 to 20 percent are in management and creative functions.

The quote also implies that clinical informatics will become the single most important medical discipline. Clinical informatics is also the newest medical discipline. In fact, it’s so new that no one is even officially certified in it yet. The first tests are this October. It’s an awfully bold claim to suggest that a not-widely understood, not-yet-available, what-will-be niche discipline will be the most "important." Let’s consider why that might be the case, though:

1. The clinical informaticists who select, customize, integrate, and roll out large-scale, intertwined health IT systems are making thousands of decisions on behalf of their users. Their decisions will help or hinder every employee’s actions and behaviors all day, every day. When companies, organizations, or systems fail, it’s because the managers failed, not the soldiers.
2. Building on top of the first argument above, medicine has traditionally been an individualistic practice. There weren’t “doctor managers”. Medical department heads don’t “manage” their “subordinate” doctors in the same way that white collar managers manage their employees. Clinical informaticists are analogous to traditional managers.
3. Computers aren’t going away. You will use a computer every day until you die. Computers may change forms and shapes, but we will continue to integrate computers into every aspect of our lives, businesses, and medical processes.
4. No one user understands the entirety of all of the interrelations between all of the users or systems. To the contrary, most users barely understand most aspects of the parts of system that they’re supposed to know. Every person’s actions affect others directly and indirectly. Mistakes and bad data may flow through two or three people before someone feels the effect further down the line. The design of the interactions between people are extremely important. By definition, good design makes it hard to make mistakes. In medicine, where mistakes can be extremely expensive and deadly, system and organizational designers have the most important job.

Clinical informatics is a growing segment of medicine. I know half a dozen physicians that are eagerly waiting to test in October to become certified. It’s exciting to see the AMA and the federal government recognize the tremendous value these folks bring to the table.

As a former EHR designer, the only problem I have with the scope of the current clinical informatics discipline is that the discipline doesn’t incorporate formal education or training regarding UI design, data visualization, or human computer interaction elements. Clinical informaticists are already designing macro systems and processes. They should be taught and be responsible for designing the micro interactions as well. The macro and micro designs are too related to be disintermediated. Don’t leave design to the programmers. They’ve been designing clinical screens for 30 years and we know how that turned out.

Thank you to all of the clinical informaticists out there who are pioneering the discipline. I must also extend a special thank you to the incredible CIOs and CMIOs of the AMDIS listserv. I’ve learned more about informatics by reading that list serve than anyone could ever learn from reading a book about informatics.

Is Glass a G+ or a Wave?

People ask me all the time if Google will pull the plug on Glass.

No, they won't. Ever. Well, at least not until they develop contact lens computers.

Here's the right way to think about Glass's viability: Is Glass a G+ or a Wave?

G+ is a company wide endeavor. G+ has been integrated into literally everything Google does - search, maps, Gmail, Android, Chrome, etc. It's a layer that lives across the Google-verse.

Does Plus make any money? No. Will it ever make money? No.

Does Google embark on lots of ridiculous projects that will never make a significant sum of money or inflate margins? Yes. See:

Gmail

Maps

Youtube

Android

Chrome

Fiber

In retrospect with 20-20 vision, would you kill any of these businesses because they aren't "profitable" in the traditional sense of the word? No. That would be stupid. All of these businesses are strategically valuable to Google.

Now let's consider some failed projects, such as Wave and Buzz. Both of these were isolated projects that never had any scope within the Google-verse. They were silted. And they never really had a defined use case. They most certainly weren't very ambitious, at least not relative to the projects listed above.

Now let's consider Glass:

1) most Google engineers are sporting Glass

2) Google ran the largest social media campaign ever, #ifihadglass

3) Google forked Android

4) Google "graduated" Glass from Google X into a free standing Glass division

5) Google invested in Glass's screen manufacturer

6) Google wants the concept of eyeware computing to exist, at any cost (data mining galore)

7) Google has the bank roll and the grit to do ridiculous things, hence Google X

8) Google isn't going to let Apple pioneer the next form factor revolution. Apple led the last two, smartphones and tablets, because no one else was structured to: creating new form factors without hardware and software teams is impossible. Google has figured this out, and Glass is a response.

Making any claim that Glass won't come to market is absolutely ludicrous. Glass is one of the most strategically important projects at Google.

"Your margin is my opportunity"

My favorite business quote of all time: 

"Your margin is my opportunity." - Jeff Bezos, Founder, CEO, Amazon

I can't talk about what exactly Pristine is doing just yet, but I can tell you that this quote is at the core of our belief system. Our market has already been validated by existing healthcare IT companies, and we're developing solutions that are 10x better, cheaper, faster, more flexible, and easier to support than what's in use today.

Perhaps best of all, the legacy companies that we'll be competing with are incentivized not to compete with us. Our legacy competitors have business models - and associated cost structures - that are predicated on fat margins.

Margins be damned. It's a race to the bottom, and we'll lead the way.

This is going to be a blast.

The Power of Connectivity

This post was originally featured on HIStalk. 

This blog post was inspired by a recent post by one of my favorite bloggers, Cadell Last. Last is a futurist who writes about biological and techno-cultural evolution. Like Steve Jobs, he looks backwards to connect the dots and tries to extrapolate.

Last only writes about human evolution at a macro scale, but meta-system transformations have been occurring within every industry vertical. It would be impossible to undergo meta-system change at the species level without analogous changes within the sub systems that humanity itself created.

Last identifies the following meta-system changes over the past few billion years of life on Earth. Below, I’ve listed analogous transformations in healthcare delivery since the inception of modern healthcare in the mid 1800s (modern = science based, not witchcraft based medicine).

I’m not going to delve into the first three meta-system changes up through federated regulation because they’re widely understood.

Healthcare in the US is just now entering the age of communication and self-awareness, and we’re struggling as we try to understand the new normal. We all know that robust interoperability is coming, albeit more slowly than many would have liked. Although some early ACOs aren’t doing so well, many of the pioneering ACOs are thriving. Why? Incentives matter.

Putting providers at risk creates meaningful changes. In order to make the right decisions, doctors need the right information at the right time. In time, every EHR will present cost information associated with every order on CPOE screens. At-risk models — which intrinsically foster cost awareness, coordination, and communication — are coming, no matter how loud certain individuals may scream.

Next up, we have the technological singularity, which will manifest in healthcare as IBM Watson-like technology (aka a recommendation engine). Lots of folks already love talking about IBM Watson for healthcare. It’s a novel, sexy, awesome concept and it will happen. Cleveland Clinic and others are already piloting Watson-derived technologies in a few avenues of care, primarily focused around primary care. Other even more powerful recommendation engines, such as Intelligent Artifacts, will expedite computerized automation of healthcare.

Recommendation engines will transform the role of the radiologist, then specialists, and then PCPs. Why that order? Radiology will indubitably be the first because recommendation engines can simply match patterns in large numbers of images (which provide large volumes of precise with little human variability) and patient medical history against ICD codes. The bulk of the data being is analyzed is generated by instruments, not humans, which means the data is standardized, which means that conclusions from it can be drawn more quickly. Specialists will be next because the scope of what they do is inherently less than that of PCPs. As such, the recommendation engines can tap into more focused, less error-prone data sets to provide more accurate diagnostics more quickly. PCPs will be last.

As recommendation engines power an increasingly larger percentage of diagnostics, that begs the question, what will doctors do? Radiologists don’t have many options except to turn to interventional radiology, but that’s obviously not a tenable solution for every radiologist. I don’t have a clue what thousands of laid off radiologists will do with themselves. Specialists and PCPs won’t go away, but their roles will change dramatically. Computer-based diagnostics will in time power at least 80 percent of diagnostics in the long run.

There will still be opportunities for physicians to discern complicated diagnoses, but most of the rest will devolve into clinical mechanics who perform large volumes in-office procedures at low margins. The line separating providers and mid-levels will blur. Physician assistants and nurse practitioners are diagnosing and treating an increasing number of diagnoses, and doctors are competing in a guaranteed-to-lose battle against computer-based diagnostics. And eventually with enough data, recommendation engines could even power robots to perform automated procedures.

And lastly, we have the transition to the global, connected brain. In healthcare terms, that will manifest as mass-scale quantified self and quantified civilization. When it’s cheap and easy enough to record one’s own data all the time, people will. Many will hesitate to quantify themselves at first, but the youth and the 133 million Americans with chronic conditions will lead the way.

As we collect more data about our bodies and our surroundings, we’ll develop far deeper understandings of causality, spread, and manifestation of every disease and symptom known to mankind. Once we have that data, we’ll crunch it every which way with superb analytics tools and tie the recommendations back into decision making processes.

Although I don’t find the global brain prediction to be particularly controversial (unlike the demise of the modern physician), I’m sure many might disagree with this vision based on the current state of quantified self and efficacy of most analytics technologies in healthcare. To them, I posit the Larry Page response: exponential growth. The human mind is inherently bounded by linear thought processes.

There’s an overarching theme across all of the transitions noted above: connectivity. Everything that I’ve outlined above is predicated on ubiquitous, cheap, fast Internet connectivity. Everyone must be connected and we must all share data freely. Your data will help diagnose and treat every other human, and the same is true of everyone else relative to you. Please support and participate in Fred Trotter’s Notice of Privacy Practices revision.

PS, Google, thank you for commercializing Google Fiber and Google Loon. Everyone needs faster Internet.