Most everyone admires the big-thinkers, the ones with the big picture. But take a look at some big-thinkers thinking small…

Leonardo Da Vinci: “Details make perfection, and perfection is not a detail.”
Walt Disney: “There is no magic in magic, it’s all in the details.”
John Wooden: “It’s the little details that are vital. Little things make big things happen.”

So, the big picture requires getting down to assemble thousands of pesky details. And when it comes to offering vaccinations, there are a surprising number of details to get right.

What got us thinking about surprising details was talking with Dr. Leanne Field, Clinical Professor at the University of Texas and Founding Director of the university’s  Public Health Program. She’s been leading a team that created and tested the system they dubbed “VAC-OPT,” as in Vaccination Optimization. To a person who’s never had to administer vaccinations, the decision process for giving shots seems simple enough:  just  0-1, you’re either due for the shot or you’re not. But, as Dr. Field explained, it’s anything but simple: “The CDC publishes an adult vaccination schedule that’s increasingly complex – the 2024 version is 13 pages long, and it’s updated several times a year. The latest version includes seven pages of notes for special situations.” The big problem is that there are plenty of exceptions to the routine schedule – the CDC lists 21 medical conditions that affect immunizations, ranging from cochlear implants to pregnancy to diabetes.

Dr. Leanne Field

We took a look at the CDC recommendations, and while the color-coding makes it look straightforward, if you zoom in you see that the “answer” provided, so easily located, is often to “See notes.” 

And those notes take some reading time as you can see in this one sample page…

Then, if that didn’t require enough searching, there’s another layer of complexity: contraindications to vaccinate, based on allergies or adverse reactions. Those are in yet another place, in an appendix.

VAX OR NOT TO VAX?
In the face of all that complexity, most electronic health record systems offer up vaccination recommendations – what is known as the “vaccine forecast”– based on just two factors: age and vax history.  That leaves a lot of detail up to the provider… detail that is constantly being updated.

Who can keep up with all of this? Dr. Field quotes research that concluded that about half of internists and family medicine physicians found it “difficult to determine adult patients’ vaccination status.” And we’re guessing that most of the other half weren’t paying attention to how difficult it is. It matters, as Dr. Field explains, because “we know individuals are more likely to get vaccinated if it’s a recommendation by a provider. If the process of making recommendations is difficult, you set yourself up for gaps and under-immunizing.”

If that weren’t enough, there’s another maze of detail: medical codes. As Dr. Field explained, “This is a system that takes the clinical language that describes patients’ medical conditions for vaccination and turns them into clinical decision logic – code — that can be read by an automated vaccine forecasting system.” This is where VAC-OPT comes in.

THE STUDY

Inspired by Dr. Vincent Fonseca, a former state epidemiologist and someone currently working with patients who often have chronic conditions that affect immunizing, Dr. Field designed a study to test the feasibility of computerizing the detail work of more complete vax forecasts and thus “cast a wider net” and identify more of those high-risk patients with underlying medical conditions who need to be immunized.*

While we keep hearing about Artificial Intelligence, the problems facing a computer in this case were more of the Artificial Fact-Checker variety: check a person’s medical information against all the indications and contraindications for immunizations. But hold on — the tougher problem was getting all the facts organized and systematized for the computer system to do its checking.

The VAC-OPT team  took the CDC’s codes (the CDSi Code List) and compiled additional codes from SNOMED-CT and ICD-10-CM coding systems to create a much expanded list, called the UT-CDSi Code List, going from roughly a hundred codes to nearly seventy-eight hundred. That took a year and a half.

DOES IT WORK?

Then, the new system used actual (de-identified) patient data to see how accurate the UT-CDSi code list was at  correctly identifying patients with underlying medical conditions who needed vaccination, while correctly excluding patients who did not need vaccines.. Dr. Field and her team were able to report a perfect score, identifying all the complex cases’ vaccination needs.

STChealth was then able to use the expanded UT CDSi codes to create an “optimized vaccine forecast” for at-risk patients and for those with contraindications, following ACIP/CDC guidelines.

THE UPSHOT

Dr. Field and her team can declare a win, with Dr. Field saying, “We have successful proof-of-concept.” She added, “We could really see the impact of underlying medical conditions and contraindications, and were able to “cast the net wider” with the VAC-OPT system to deliver optimized vaccine forecasts tailored to each patient’s needs.

What’s left to be done? “We need to create an interface that is easier for providers to use. We came to understand that the first user is typically a medical assistant and we got feedback on how to make the forecast easier to read. Then, we need to expand our list of vaccinations – we started with just five. And we also want to integrate pharmacy data.”

Given all that they have learned so far, Dr. Field sees this as the future of getting vaccinations, automating the process of finding an optimized vaccine forecast by clicking on the VAC-OPT system and more-or-less instantly making sure those at highest-risk are getting the vaccines they need to stay healthy. 

Meanwhile there’s one old-fashioned problem to finishing the VAC-OPT system: “We are actively looking for funding to scale up our work.”

NOTES

* From Dr. Field: “Any automated clinical decision support (CDS) system, including VAC-OPT, does not eliminate the need for clinical decision making by providers. Clinical judgment is important in making final decisions about vaccinations for some individual patients. From CMS.gov: ‘CDS is not intended to replace clinician judgment, but rather to provide a tool to assist care team members in making timely, informed, and higher quality decisions.’”

“We are grateful to the entire VAC-OPT team including the host of our study, People’s Community Clinic in Austin, Texas, and two non-profit, health information exchanges – Connxus and C3HIE, who passed the optimized vaccine forecasts from the clinic to STChealth and then returned them back to the clinic – in seconds!

---

STATS OF THE MONTH

Thank You Notes for Immunizations
Is It Time?

By Bill Davenhall, Geomedicine Analyst

When was the last time you received a thank-you note for getting an immunization? If you are like me, “never” would be the correct answer. Research results published by universities about thank -you notes and the reactions to them are not surprising:  they found that thank-you notes, if sent on a timely basis, have a much greater impact on repeat business than expected.

Having been an active participant in the recent pandemic response experience, (as a consumer of immunizations), the only incentive of getting the immunization was fear — to protect myself. Fear was my behavioral “push” rather than a preventative “pull” for a possibly lifesaving vaccine. I suspect that a great deal of money and effort was expended by state health agencies across the US in making sure we returned for our follow up vaccine, and, hopefully  future compliance with the best medical advice.  The true costs of this effort will probably never be well known.

Why not a national program to send everyone who receives any vaccination a “Thank You” note within 10 days of that immunization? Impractical, you say? What does it cost for program efforts to get more people to get any recommended immunization in a timely manner? What percentage of all vaccinations occur within the same unique family household? How many households contain adults and  children who do not receive all the required immunizations? What is the likely time lag between receiving an immunization and the receipt of a USPS-delivered “Thank You”?

A media reported story goes that Princess Diana was said to be a well-known “Thank You” sender. After each event she wrote a thank-you note and promptly mailed it!  If you have ever received a thank-you note from a medical provider, you are lucky!  Do you have any idea what someone thinks in getting an unexpected “Thank You” note in a timely manner, let alone even getting a note at all? Imagine how you would feel if you got a thank-you note from your State Health Department Director thanking you for taking the time to get an immunization and encouraging you (or your household members) to stay current. How valuable would it be to suggest to you what your next immunization might be?

If you take this idea to the next step, I suggest you do some serious analytics on your own immunization registration records: see if you can envision how you might select a target test group of the population, figure out the logistics of the process, and identify the barriers to doing it in a timely manner. Determine how difficult it would be to figure out what immunizations they (perhaps the whole family household) will need next; and importantly, offer gratitude for helping protect themselves, their household stakeholders, and the wider community.

All state immunization registries have the data to accomplish such analysis now, a testimony to the foresight of those who created the registration system to build a credible data collection effort that supports the work of the entire immunization ecosystem. The next logical step is to make that system do something unexpected – demonstrate a persistent operational program (24/7 and 365 days a year)  of communicating thanks to the public that accepts vaccines now, and where you can demonstrate the value of saving lives and reducing morbidity. The historical data that all states have today could easily help them determine areas and pockets of opportunity where such a thank-you note might encourage household “stakeholders” to take a greater interest in their own preventative healthcare. Knowing that someone else recognizes your efforts to protect your health or others in your household matters.

Is anyone interested in this experiment?

As always, I appreciate 2^nd opinions.