Some recent discussions got me to thinking about Public Health again, possibly because I've been spending a significant amount time with the IHE Quality, Research and Public Health technical committee and individual members this week. Some time ago I wrote about "
Taking Cost out of the System". Let's list some of the activities that individual providers and organizations need to engage in with public health:
- Reporting Immunizations
- Getting Immunization Data (if you are lucky)
- Reporting information to disease specific registries (e.g., Cancer)
- Reporting Syndromic Surveillance Data
- Reporting Notifiable and Reportable Laboratory Reports (and Orders)
- Pediatric Lead Testing
- Early Hearing Screening
- Reporting Birth information
Three of these are already part of Meaningful Use Stage 1 here in the US, and others are expected to become part of Stage 2. With few exceptions, all of these different reports go to different silos, using different messages.
When we look at Federal law covering exchanges with public health, it is noted as an exception to the usual rule of "No PHI", although many public health agencies choose to limit the amount of PHI they receive to make their systems less costly and easier to work with.
I'd like us to think about a different model for dealing with public health. It's not the "Send the question to the data" that
Wes Rishel or I talked about
in recent postings, either. Who bears the burden of performing the query? It's not public health, its the individual providers and organizations who would have to receive and process it. That really doesn't make public health's life any easier. There are just as many systems to send the queries to, the data is not "normalized" as it is when public health silos up, and the queries aren't quickly done for ad hoc purposes.
When a doctor treats a patient, the ideal today, is that he treats the whole patient. The system may fall short of that, but it has certainly been expressed in the "patient-centric" focus. Why should public health be any different. After all, public health is really dealing with an aggregate of patients (in the usual sense of the phrase, I realize public health also provides individual patient care, but that is a different sense of the commonly understood meaning).
What would happen if instead of every single provider having to figure out how to extract data from patient records they have on the patient, they were instead required to provide public health with a copy of this information (with a few notable exceptions). It could be through an HIE, through Direct, or through some other transport mechanism. Public health could then take those records, abstract from them what is needed to deal with:
- Immunizations
- Syndromic Surveillance
- Disease Registries
- Birth Records
- Pediatric Lead Testing
- Early Hearing Screening
- Reportable and Notifiable Conditions
What happens to the way that public health is organized? Well, for one, they'd have to pool their IT resources and eliminate redundancies. Two, they'd be responsible for analyzing and routing the data that they need for public health functions to their internal functions, rather than making providers do that for each separate function. Third, they'd have to get a lot better about security of the data that they centrally manage.
What would happen at the provider end? Instead of having one or two interfaces that need to be managed for each public health function that would need to be supported by each provider organization, there would be only two (the outbound push to public health, and a way to query public health where necessary). You could use the same standards being selected for clinical documentation being exchanged between healthcare providers, so that would eliminate another source of variation in the data. Public health can then take the clinical data they need and distribute them where they need to go, with appropriate pseudonomization, abstraction, extraction and roll-up, et cetera. If this were done at the state level, with local access, this could have an even greater impact. Many organizations have to deal with multiple municipalities when communicating to public health. Imagine the cost savings to providers that could occur by ripping out and throwing away dozens of unnecessary interfaces. I've seen estimates of anywhere from $5K to $20K per interface (depending upon who makes the estimates and the complexity of the interfaces being discussed). Even using a lowball figure, that savings could be significant to many provider organizations.
This is a pipe dream. I don't see it happening, but not because it's technically unfeasible. Politically and organizationally it isn't feasible. You'd have to get multiple agencies to agree to share resources, and give something up (control and resources) in the process. I heard someone report at an eHealth Connecticut conference five years ago that they figured the State could spend $5-10 million to set up a State HIE and still have twice as much savings left over if they were able to eliminate public health silos. But could you imaging getting all of those state agencies to buy in? What about the policy ramifications? Actually, because public health has the HIPAA exemption that's not pre-empted by state law, they might also get around some of those. What about public response? What, you are going to send all my medical records to the state? The privacy barons would have a field day over that one, and certainly there are some technical barriers around what would be needed to be put into place to make
them happy. Oh, and then there's all the fine wording of all that public funding running around in public health. That's why we had excellent surveillance mechanisms for H5N1 that completely failed us when we were hit with an H1N1 epidemic in the making. That's not disease-centric thinking, that's even worse: Serotypical thinking, and unfortunately, all to typical.
So of course, it's easier to
create a new standard to work around all those other really complex issues. I vacillate around the simplicity of the query model and the cost distribution that puts most of the work back on the provider. Do I want to tackle an easy challenge and avoid the hard stuff that would provide even more savings? I dunno. It's 2:00 am. Do I really need more work?
But, I cannot help thinking that one standard already supported by several hundred EHRs could replace three, four or even more standards used in public health exchanges. All we need do is to stop public health from organizing themselves into disease-centric silos.
-- Keith
P.S. Maybe what I should have said is that public health should be population centered.
