Re-envisioning Electronic Health Records

Okay, so this “forward looking” image is somewhat self-serving.

Recently I was asked by a colleague, how would I size up current EHR’s: what are the major wins and major needed improvements? In about an hour, and off the top of my head, I wrote this list. It is neither comprehensive nor deeply reasoned. However, it IS a compilation of frustrations and grateful moments that come from having help design, implement and USE an EHR over the past 2 decades.

  1. See Youtube, from a Boston EHR interest group, CRICO: Provocative; I don’t agree with all of it, but it provokes a lot of discussion, see below.
  2. MAJOR win: engaging the patient using the portal: online communication, access to records, Open Notes, Open Results, is just the beginning of a needed improvement in information transparency. Anything that eliminates the telephone tag circus is good.
  3. MAJOR win: having the chart accessible ANYWHERE ANYTIME avoids missing data (medical errors from handoffs of care) and paper filing and paper shuffling costs. As one colleague wrote: PAPER KILLS. In the late 1990’s between 1/3 and 2/3 of patient appointments did NOT have the relevant paper chart pulled in time for me to see the patient. How embarrassing. This no longer happens. Also, our paper filing, despite a team of 50 in medical records, was always 2-3 weeks BEHIND. Even if you DID receive the paper chart, chances were good you did not see the recent report you needed.
  4. MAJOR win: improvements in legibility (sometimes at the cost of unreadability due to note bloat). How many times per day did used to I take a paper chart to a colleague or nurse and ask “what you think this says?”
  5. MAJOR win: easy narrative documentation using Speech Recognition, and increasingly, Natural Language Processing (detecting codified concepts using machine learning). We only just starting to see the fruits of these technologies. A typical physician’s cost for human transcription, per year is about $15,000. And the turnaround time can be days, resulting in missing data during that time. Speech rec is instant. AND NLP has the potential to create instant alerting and reminders.
  6. MAJOR win: reminders and alerts improve the frequency of doing the right thing at the right time more often (when well designed). This is the FLIP SIDE of of Alert Fatigue (see below). I love when my system reminds me to vaccinate, or screen for colon cancer, or screen for depression, particularly when I catch and prevent an illness that I would otherwise have missed.
  7. NEEDED IMPROVEMENT: alert fatigue: poorly designed, terrible signal-to-noise ratio of alerts. Enough has been written about this. Our Physician Informatics Group (PIG: yes, we don’t take ourselves seriously), constantly struggles to improve the SIGNAL to NOISE ratio of these alerts, and to reduce alerting. I consider it a personal failure if we have implement at “Best Practice Alert” that stops a doctor’s work, instead of designing a smarter EHR that “guides” and “nudges” a doctor’s behavior, so that we make the RIGHT THING EASY.
  8. NEEDED IMPROVEMENT: Better ways of capturing physician-patient interaction (see #1), maybe full video recording instead of typing out a history, and having the machine collate into a timeline, concise narrative.
  9. NEEDED IMPROVEMENT: user interface design: (see #1), why can’t the electronics disappear into the wall until it is needed and then pop in with reminders and context-sensitive help just-in-time?
  10. NEEDED IMPROVEMENT: how to eliminate communication barriers and snafu’s based on nurse-physician-patient ping-pong messages.
  11. NEEDED IMPROVEMENT: an appreciation from clinical leaders that an EHR is NOT THE SOLUTION: instead, need to focus on a clinical re-invention that uses an EHR as tool to create better teamwork and communication. How to get that across? Our biggest successes come from clinics that realize this one fact. See previous posts on SPRINT  EHR Sprint team: work hard, persevere, sometimes you get to build a dream team and TRANSFORMATION Politico (and HuffPo): The Doctor of the Future (with stuff about us, and Care Redesign at UCHealth!)

CMIO’s take? Send me a message! What’s missing? What would you take issue with? Let’s craft a message to our EHR vendors and demand innovation and something better. I’m convinced we’re in version 20 of something that will need 50 versions to get right.

Book review: Aurora by Kim Stanley Robinson


This is a fascinating speculation into the future of human space travel to the Stars. What would happen if you were born, the sixth generation to live on a starship? What if your planet destination was questionably habitable? What if your cohort decided to return to star travel? What if things progressively went wrong? What would star traveling humans had never seen planet-side sunrise?

Robinson does a nice job setting up the scientific premise of a multi-generation ship heading out to the stars. Ho would the 6th generation feel about being brought into a world not of their choosing?

How would the travelers deal with the inevitable technology and mechanical breakdowns over hundreds of years? How can any small society have the resilience to self-repair long enough to reach their goal?

And what if there were major decisions to be made upon arrival? What is the leadership and cultural environment of a ship of 2000, and could it withstand a life or death decision without degenerating into anarchy?

Sure, the storytelling at the individual level is perhaps less compelling, but the grand scale of Robinson’s vision, sweeping across centuries and across vast reaches of interstellar space, is a great journey.

CMIO’s take? The human psyche is fascinating, at the levels of society, community, organization, family, and ultimately individual. Stepping outside our own familiar environs lets us see ourselves more clearly. Sci fi helps me do that. How do you do it?

EPCS: Morphine (ukulele video)

This guy thinks he can sing. How unfortunate.

Has your organization adopted technology to electronically prescribe controlled substances (EPCS)? If so, good for you! If not, what are you waiting for? UCHealth is kicking off this project on April 9, 2018, a “big bang” for all providers who have a DEA number to be able to e-prescribe and avoid that horrible, bifurcated workflow of “electronically transmit all the regular prescriptions directly to pharmacy, but remember 1) keep your tamper-proof prescription paper stocked in the correct drawer of the printer, 2) to go get that narcotic prescription off the printer, 3) sign it with pen-and-ink, 4) then tell your staff to track down the patient to come pick it up or remember to put it in certified mail to the patient, especially if the patient called to request, or worse, you forgot to get it off the printer before the patient left clinic.

After the DEA ruling to permit EPCS a few years ago, our EHR vendor developed the 2-factor authentication tool to meet the regulation, and now we have finally contracted with a company to provide that service, link our physician identities to their smartphones via an encrypted app, and now a smartphone alert and a password (something you have and something you know) are our two factors that permit EPCS.

Here’s a little something I wrote to commemorate our EPCS go live:

CT on ukulele singing “Morphine” inspired by Eric Clapton’s classic song “Cocaine.”
(Part of my YouTube channel)

I also created an internal-only training video for Duo Mobile as second factor for EPCS.

I learned the structure of a well-told instruction video from the original “Apple Business” video tips back in the early zips (2000’s). You’ll see a human, then a quick screen video, then close with a human. Perfect format:

CMIO’s take? Sometimes a video tip (or video musical tribute) is what gets an audience’s attention.

Amy Cuddy and the coming wave of scientific recriminations

I have seen Amy Cuddy’s presentation; she speaks well, has a compelling argument, and a research study that she published in 2010 that backed up her assertions. It made sense, and “power posing” became something I thought about at times.

Since then, there has been this backlash; the article speaks about it in detail, that is a personal attack on Dr. Cuddy, which is mostly unfounded, and is tragic for her career, and I am hopeful for her resilience and rebound.

However, much more concerning is the implication for all social science research and indeed all research; the critique about replication of findings and the over-reliance on the p-value in research. Most studies rely on a p-value of “0.05” indicating that the likelihood that a finding in a study represents a real effect in the entire population that was not studied (for example, studying 100 patients where there are 10,000 patients who were not studied, and it would be unrealistic to study all 10,000), should show that it is highly unlikely that the results would occur as a result of sampling error.

The critique is that, before publishing, researchers should REPLICATE their findings, or better yet, have others replicate their findings. The challenge here is that research is already slow, plodding (it can take years to write and submit a grant, suffer the indignity of a very small chance of getting funded months or years later, conduct the study over a year or 3, then spend time writing up the results, submit the paper, hope that some journal somewhere accepts it), and with long delays before results are accepted and published. Now consider the requested addition of replication: do this AGAIN in a different setting, a different population of patients and then combine or compare the results.

Shall we double the grant funding of all results so that we can do this? Is research funding not scant enough as it is (some colleagues have a 1:500 chance of funding their research at current funding levels)?

CMIO’s take? Not an easy question to answer. I think there is a lose-lose proposition coming out of this discussion, with no clear path forward.

Amazon’s AI Flywheel, and what we can learn from it

from amazon dot com

A fascinating look inside Jeff Bezo’s strategy at Amazon, pivoting the entire company on Artificial Intelligence, and how it is driving Echo.

More interestingly for me, how does AI connect throughout that massive company, how has one person brought that single technology under one roof, one connected integrated team, and make the results of that knowledge available to all the other teams previously developing such tech independently?

Finally, how does the AI group create the virtuous flywheel, described in Good to Great. 

Somehow, Amazon has cracked one of the great business puzzles; how to create a small innovative team within a much larger bureaucratic business, that is nimble, learns fast, develops expertise in many related areas (surrounding AI), and then serves up that knowledge to the rest of the organization in useful, coherent, collaborative ways, driving up the value of the entire organization?

Whatever it is, Amazon is bottling it, and Amazon is selling it.

CMIO’s take? The cycle of the virtuous flywheel is our aspirational goal.