Here was the assignment: “Write about health care data transfer.” OK, sure. But there was a second part to that instruction: “And make it interesting.” Whoa. Possible? We took up the challenge. Early on we came upon an encouraging quote from a Google economist:

“I keep saying that the sexy jobs in the next 10 years will be statisticians.”

That sounded promising, but upon further digging, we learned that the quote was from 14 years ago.

Still, it’s fair to say that there is a direct correlation between career advancement and the amount of data a person is expected to master. Indeed, data mastery can accelerate career advancement. So maybe we could argue this… 

If you love data, it will love you back.

Sounds good, but it doesn’t get us to the topic of data transfer and making it interesting. Undeterred, we persevered, and eventually came the breakthrough: we sat down with Joe Kelly. Here’s one of the first things he said:

                                                “Health isn’t just about health.”

He went on to ask, “What is the true view of your health?” Then, “Every patient has a clinical view of their health, but the view changes.” Kelly listed more than one “you” (one plural for “you” is “youse”) — depending on who’s doing the viewing. He enumerated some youse:

“There’s the actual you, the clinical you, the policy you, the food handler you, the student you.”

Expanding on that thought, he said that your health depends on who’s asking, and that gets us to data and its transfer. He said, “The answer to a question about your health will change if it’s a doctor asking for vaccination records, versus the McDonald’s Corporation asking as part of your employment, or the school system asking as part of your enrollment. Everything is context.”

This means that the challenge for people providing data is to be able to adapt to context. This is where the world of data has changed and is changing. “When people talk about ‘data sharing,’” Kelly argues, “what they have meant is data transfer. That’s been the bane of all systems – you transfer terabytes of data and hope it doesn’t go wrong.”

He paused here to ask if we remembered something called “sneakernet.” No. Turns out that not so long ago the transfer of large quantities of data took so long to send via transmission wires that it was faster to put it on a disk and drive it to the location and walk it in – hence, the “sneaker” in sneakernet.

The point being, of course, that data transfer has come a long way. Nevertheless, there are enough problems that Kelly compared data transfer to a bridge, and not one you’d want to drive across: “You build a lot of bridges and every now and again you lose a few cars, but that’s just what you’d come to expect.”

The solution? Kelly insists,

“You don’t transfer data, you transfer access.”

He added, “The home of the data is the home of the data. And we can move that data at the speed of light — we can send it into space. We don’t need to load paper into a truck.” And bringing the topic home to vaccinations, he said,

“When it comes to vaccination forecasting, everybody knows who needs what… or do they? You have clinical policy and the doctor’s view, but there’s also state policy or school policy.  And the data on a person’s vaccination status might be in Indiana and also in Arizona. Both have ownership claims on the data, but they can transfer authority to use some of it; they can simply agree upon transfer keys.

“That means it’s now possible to sync the data, almost in real time, and put it in the context that’s most useful to whoever is asking. In the system we are developing, you can say…

“‘Tell me everything about Bob – give me all that we know and all that they know.’”

This isn’t just theoretical – the first part of the system is now being tested with the IIS in Louisiana. Kelly says, “It’s called Atlas and it’s part of STC/One. What I love about it is that it doesn’t try to tell you how you should look at your data. It’s based on the premise that there is no wrong way to get data. You can look at your data any way you want.” A quote from the country’s best-known statistician, Nate Silver, is relevant here…

“On average, people should be more skeptical when they see numbers.

They should be more willing to play around with the data themselves.”

So, putting this all together, the goal is to have data sharing (not transfer!), with a system that retrieves data matched to the context, whether it’s the first you or the second you or the tenth you, and no matter how many youse there might be. Thus, the goal is to create a system that lets you play around with data until you love it and it loves you back.

* * * *

Joe Kelly led the development of the Atlas project while CTO Officer at STChealth. He continues to oversee its implementation, serving as Senior Contractor.

---

STATS OF THE MONTH

A New Demographic? The Immunization Partisan Gap

By Bill Davenhall, Geomedicine Analyst

 In a nutshell, the “immunization partisan gap” is  the impact that a person’s political leanings may have in influencing the receiving of an “updated” Covid vaccine.

Twenty years ago, it was unlikely that political differences among the population of the United States would ever  become a major threat to the immunization ecosystem; however, we now see evidence that the Covid vaccines brought with them some major concerns that differed substantially due to various political leanings.  A recent study (2024) conducted and reported by the Pew Research Center concluded this: “Levels of concern for getting or spreading the coronavirus are about the same as they were in March of 2023 and remain down dramatically from the earlier pandemic.” Now, even though the virus is still a concern in many areas of the US after almost four years since its arrival,  a person’s partisan political affiliations remain a “gap” in vaccine “uptake.”

While the usefulness of these new research findings, drawn from a sound sample survey of approximately 11,000+ persons, will be subject to debate, there is an ancillary source of more precise data that could potentially offer another “oar in the water” over the concern over declining vaccine “uptake “rates  – household voter registration files. Each state presently maintains robust data collection systems that identify every household member that is registered to vote along with key demographic information such as year of birth and street address. Using this data source, linked to immunization registration files, could offer immunization programs with even more precision in their immunization promotion and vaccine distribution network. Below is one interesting finding that the Pew Research team identified – and perhaps suggesting where greater precision could be applied.

Linking publicly available state voter registration data with state immunization registration data at the household level could be a new way forward in finding more precise ways to target  those households that wish to protect their family members from preventable diseases — now and into the future.

In next month’s column I will explain how immunization advocates could begin to look at ways to leverage existing data through the applied use of household geography!

As always, I appreciate 2^nd opinions.