How to build resilience against loneliness in aging (Harvard news)

Harvard researchers have some insights into older adults and some recommendations and concrete actions to take to combat loneliness as we age.


TikTok Education Strikes Again: Be a Haiku Hero!

I have always wanted to wear a cowboy hat, pink with flashing lights. And now, my dream has come true. Here is our latest 60-second education video on how to configure your Epic Haiku iPhone secure chat settings for success.

I am getting into ultra-short form education. One minute to jam-pack a bunch of ideas into a quick (and hopefully entertaining) video. Here’s how I built this:

  1. Set yourself an unrealistic expectation to teach sophisticated Secure Chat settings to physicians / APP’s in 60 seconds.
  2. Broach the TikTok video idea to disbelieving and pessimistic informatics colleagues
  3. Turn off the computer, clear your desk, take out a yellow paper pad and sketch out a storyboard for how this will go, six frames to a page. Sit with head-in-hands, thinking “Make it shorter! Make it funnier! But How?!”


  1. From that, write up 5-second video scripts and clothing and prop requirements for each video
  2. Text your disbelieving colleagues at the last minute, the morning of your next Large PIG (physician/APP informatics group) in-person meeting to ask for a patient gown, scrubs, white coat, other stuff like a cowboy hat, jack-in-the-box, whatever they can steal from their kids that morning.
  3. Run the (exhausting for an introvert) 2-hour Large PIG meeting in the boardroom and then recruit a handful of colleagues to stay late and shoot 5-second videos until the shot-list is done.
  4. Block out the calendar for an entire day of struggle, fire up Final Cut Pro on the laptop, open YouTube to learn how to use Final Cut, and stumble your way to an amateur production of BE A HAIKU HERO.
  5. Show your colleagues the result. Ignore the fake vomiting noises and keep going anyway.
  6. Ask Epic Wisconsin nicely for permission to use a screenshot and post it to on Youtube and the blogosphere.
  7. Apologize to your marketing department for the completely amateurish nature of the resulting video.

See? Totally simple.

Here you go.

I’m sorry Dave, I’m afraid I can’t do that (updated big data and AI talk for 2023) given at CSO HIMSS Ohio

Lovely conference facilities in Dublin Ohio

The Happiest Man Alive shares his 3 rules for life (NYTimes)

From NYTimes: Photo of Matthieu Ricard by Braulio Amado

What can an ordained Buddhist Monk who is also a cellular geneticist and the Dalai Lama’s French interpreter, tell us about happiness? This is a fascinating read. Perhaps all of us in healthcare could learn a thing or two for ourselves, and to teach our patients.

What is Screen Apnea? Why you may breathe less when online (NYTimes)

What do studies show about your breathing when you’re on email on your phone, or doing work online? …. uh-oh, I’m talking to ALL OF YOU out there.

This thought-provoking New York Times article hits me pretty hard. Our breathing is tied into our thoughts. When thinking hard, we will (as I know) stop pedaling on our Desk-cycle under the desk. Turns out, we also slow down and at times stop breathing.


Read the article and figure out what to do about it for yourself.

Will Your Next Doctor Be … A Bot? (SunFest) with bonus uke song

What happens when you put a news reporter, and AI researcher, a Bioethicist and a CMIO together to discuss AI, Chatbots, Bias and emerging trends? You get this highly interactive and entertaining panel. And maybe a song.

Thanks to the Colorado Sun, and XCEL Energy for sponsoring our panel on AI in Healthcare at SunFest, held in Denver on the Auraria Campus of the University of Colorado.

I very much enjoyed this conversation with my colleagues at the University of Colorado, including Dr. Casey Greene, Director of the Center for Health AI, Dr. Matthew DeCamp, Bioethicist at the Center for Bioethics, and practicing general internist.

Among other topics, we covered:

  • AI, Large Language Models and Chatbots, defined
  • Predictive analytics and how they’re different from Chatbot AI
  • The potential dark side of AI in healthcare
  • Using ChatGPT-like tools in summarizing electronic health records, in helping doctors write progress notes, and in helping physicians, physician assistants, nurse practitioners and nurses, reply to patients via online messages.
  • Risks of automation, including Automation Complacency
  • The risk of hidden bias in AI, and how that compares with existing bias in healthcare today
  • Future plans for AI in healthcare

Listen to the end for an updated version of “Hospital of the Rising Sun – Pandemic Edition” with me and my trusty ukulele.

SunFest 2023: Watch every session with Colorado politicians, expert panels and more

AI can’t build a Hi-Rise (NYTimes)

…but it can speed up the job

From the NYTimes.

We, like everyone else, are dipping our feet into the AI chatbot world.

“Lions and Tigers and Bears, oh, my!”

Lots of concern about hallucinations and confabulation by Chatbots.

Also, lots of concern about becoming irrelevant if other companies figure out how to work with Chatbot AI’s successfully, and we do not.

Reminds me of a quote from Prizzi’s Honor, a movie about the mob, where Jack Nicholson, faced with the betrayal of his girlfriend, with a truly befuddled expression, famously says:

“Do I ice her? Do I marry her? Which one of these?”

CMIO’s take: I feel like Jack at this point in our AI/chatbot journey. Don’t you?

Doomsayer, Reformer, Warrior: Which AI Attitude are You? (NYTimes opinion)

From the ethics debate, the AI Wars are heating up and there are 3 distinct factions predicting our AI future. Which one are you?


Work-in-progress. Improving Patient AND Physician Understanding of Radiology Reports

Help us solve a conundrum! We want to teach physicians just-in-time. However, this may worsen a patient’s understanding of a radiology report. Can we make this a win-win?

Image above: From Dall-E via BING chat

To: patients and informatics colleagues out there:

Here is your problem of the day. Physicians and APP’s who order ultrasound tests for “screening for liver cancer” for example, have a difficult time understanding the LI-RADS categories that radiologists put into the radiology reports.

Brilliant! Our always-helpful radiology colleague recently began inserting the following just-in-time education intended for the physicians and APP’s receiving these reports:

Super! Everyone is happy. Physicians have a much better understanding of the interpretation of the report AND there is a helpful index at the bottom of same report explaining the LI-RADS category and also the Visualization score, and some sense of what to do next.

What could go wrong?

This was turned on about a year ago. A few weeks ago, one of my patients, who has full access to her EHR Patient Portal, and this immediate access to her ultrasound report, called me in a panic: “My ultrasound shows that I have POSITIVE – Observation Detected. Doesn’t that mean I have cancer? It says right here!” Fortunately, we had an upcoming in-person appointment and I asked her to bring the printout so we could discuss it.

I looked at the printout she brought in. Indeed, she had a benign appearing ultrasound report. What she was reading was the very bottom of the report that included the index of LI-RADS categories.

I then explained that this part of the report was only a reference tool for doctors and did not apply to her. After about 10 minutes of discussion, she seemed to understand, and then gave me a stern look: “This is very confusing. Why are you doing it this way?”

I took this feedback and sent it back to our well-meaning radiology colleagues, that our patients were now misunderstanding the LI-RADS explanation text intended for our physicians, and over-interpreting it to apply to themselves.

My radiology colleagues responded with the following addition in BLUE:

What do you think? Is it helpful? Is it enough?

In all of our radiology reports, that we have been sending to patients immediately for a couple years now, we have standard text that reads:

If you are a provider with questions, you can call Radiology Help Line at xxx-xxx-xxxx. If you are a patient with questions, please contact your ordering provider to discuss.

We do have an occasional patient who DOES call the Radiology Help Line, and in those rare cases, our radiologists are happy to have that discussion with patients, and patients are satisfied with those discussions.

The addition of the help text for LI-RADs is written about the patient in the 3rd person “the patient”, as opposed to the quotation above, which is written for the patient in the second person “you.

CMIO’s take? For now, the help text on LI-RADS is what our reports have included. We have had no more patient complaints registered.

What do our informatics colleagues and online patient communities think of our approach? Do you have a suggestion on how to make it better? We want our radiology reports to be a WIN-WIN: more easily interpreted by both our ordering physicians/APP’s as well as by our patients.

Advances in PGx (Pharmacogenomic or Drug-gene interaction) at UCHealth (guest bloggers Dr. Christina Aquilante and Dr. David Kao)

Pharmacogenomics is advancing quickly: we can warn prescribers in the EHR when patients have genomic variants that reduce medication effectiveness. We are going from screening populations (18,000 so far), to anticipatory screening for high risk patients (cancer center patients about to choose a chemotherapy). Cool.

Previously, at the Colorado Center for Personalized Medicine…  

In December 2021, our heroes (CCPM in partnership with UCHealth) began releasing clinical pharmacogenetic test results for CYP2C19 and SLCO1B1 to the Epic electronic health records for CCPM biobank participants.

Eighteen months later, our program has flown to new heights.  We have returned results to over 18,000 biobank participants, which have impacted the care of over 2,600 patients.  We have expanded our program to include an additional 5 PGx genes (DPYD, TPMT, NUDT15, CYP2C9, ABCG2), 4 of which went into production the last week of April.  Altogether, these genes impact the effects of 30 different medications ranging from antidepressants to anti-inflammatories to chemotherapies!   


Meanwhile, back at CCPM headquarters…  

Our heroes continue to return high impact genetic variants with potentially life-changing and life-saving impacts for biobank participants and just as importantly, their families.  Our biobank lab and genetic counselor team have returned results for around 30 of these genes to over 250 patients.  As a result of this effort, many patients have been referred to specialists for evaluation and monitoring to identify and treat any concerning conditions as early as possible. In many cases, participants’ siblings and even children are also being tested, often when they otherwise wouldn’t have, giving them the power to battle the villains of genetic disease.     

 Join us next time…  

When we begin performing clinical-first tests for chemotherapies used to treat certain kinds of cancer and medications to reduce the side effects of chemo.  This will be our biggest challenge yet, adding an additional 2 genes, including CYP2D6, which has the potential to affect over 20 medications that treat a host of different conditions. We will start returning non-PGx results to the EHR electronically as well and use invisible data science superpowers within the EHR to identify UCHealth patients most likely to benefit from pre-emptive pharmacogenetic testing. 

CCPM and UCHealth were leading the charge toward use of genetics for clinical care 18 months ago, and our program has grown exponentially since then.  Thank you for joining us for the next phase of our adventure!  

(Photo by Patrick Campbell/University of Colorado)

Christina Aquilante, PharmD
Professor, Department of Pharmaceutical Sciences
University of Colorado Skaggs School of Pharmacy and Pharmaceutical Sciences
Director of Pharmacogenomics, Colorado Center for Personalized Medicine

David Kao, MDAssociate Professor of MedicineDivisions of Cardiology and Biomedical Informatics/Personalized MedicineUniversity of Colorado School of MedicineMedical Director, Colorado Center for Personalized MedicineMedical Director, CARE Innovations Center, UCHealth