Last week, we took a look at historical efforts to build true healthcare interoperability and what some of the key steps were in the early stages of this drive. It’s important to understand what was tried in the past and how that can inform where we’re going. The Electronic Data Interchanges was an excellent building block, but did not address the treatment and management aspects of healthcare.
With part two, I want to talk more in depth about what hasn’t worked to help define where the gaps were and how we can improve upon them with current efforts. So let’s dive right in, starting with the importance of establishing data standards in healthcare and mandating change.
What Has Failed to Fully Work
Without the standards organization HL7, interoperability could not exist. But HL7 has labored for many years (at a global scale) to make the data flow. The numerous public initiatives, such as RHIOs, HIEs, and Meaningful Use ran in parallel with the progression of needed data formats. Through the years of initial formats, including the HL7 2.0 formats (CCDX), and the HL7 3.0 formats (ADT, ORU, etc), the standards have improved and change has occurred.
But most would agree today that the data to treat and manage healthcare is still siloed. Like the Electronic Data Interchange (EDI), there were standards created, a (partial) government mandate, armies of new “clinical informatists,” enormous expense software changes, and the creation of a new “interoperability” industry. So, what has kept interoperability from working for 20 years? And why is the current effort any different?
Data Standards in Healthcare Should Provide Better Access
First, it is important to acknowledge the limited success. Most physicians do have access to some clinical data through an HIE today (which it is greatly underutilized in many instances). Further, the incentivized adoption of EMRs has moved a great deal of the data from paper to disk. But patients still have to fill out the same forms at every physician’s office, over and over. Physicians seldom know about any healthcare that their organization did not deliver.
Perhaps the most telling indication is this: the reader is almost certainly an expert in healthcare and/or an IT expert, a specialist in this industry. Does the reader believe they can access their healthcare data, or move their healthcare data from one provider to another, or make decisions about their healthcare using their data?
People should expect far better. Today, phones and web browsers are used to manage every other aspect of life: finances, travel, shopping, entertainment, real estate, auto/home insurance, utilities, etc. In every one of these examples, accessing historical records, measuring current status, and most importantly, making personal decisions (informed by the apps and the software available on the internet) is commonplace. But far more than personal sovereignty and convenience is at issue here.
If healthcare data was truly accessible, it would improve quality and reduce medical errors. Perhaps most importantly, it would significantly reduce both the medical cost of care and the administrative cost of healthcare, which in turn would afford better access.
Why the “Old” Efforts at Interoperability Have Failed
The lack of a sufficiently good format
HL7 2.0 (CCDs) had the advantage of allowing any two data trading partners to send each other almost anything, which was also its biggest flaw. The HL7 3.0 data standards in healthcare were still not proscriptive enough and only transported a limited amount of data.
Furthermore, the vocabulary standards and the changes in the format details over time did not have a strong enough governance. Most importantly, healthcare data requires context and the “old” formats (still widely used today) either carried no context or lacked the metadata needed to understand or adapt the given context.
The lack of access
Almost all of the “old” interoperability is either batch delivered through secure FTP or driven as a stream of alerts. To protect privacy, these are walled off and require specific or custom point-to-point integration as well as legal paperwork that confers new legal liability (and restrictions) when the data moves. All providers today use EMR workflow systems, which focus on data created within the EMR and make it hard/impossible to use data from other sources. Finally, virtually all existing efforts have been business-to business (B2B) focused, ignoring the patient.
The lack of an industry-wide mandate.
The ability to exchange data is useless if there is no scalable network of other stakeholders with whom to exchange. A government-driven mandate accomplishes two key goals: a) it sets a standard for universal adoption (thus creating scalability) and b) it compels adoption (also creating scalability). RHIOs and HIEs are voluntary. Once the public funds that created them were exhausted, they sought and largely failed to establish a strong business rationale for their engagement and adoption. The noticeable exceptions to this are the pockets where limited “old” interoperability works today.
The lack of business value. While sharing the data to treat and manage healthcare would reduce costs, the above barriers limit the value of data sharing under “old” interoperability, which in turn makes the above barriers worse. Without a mandate, the healthcare industry would adopt interoperability if there was business value; under the current “old” interoperability methods, the value is limited due to scale and how difficult it is to move data effectively.
We invite you to continue exploring what health plans need to meet the new interoperability rule within the 21st Century Cures Act.
Now that you’ve finished Part Two of our series, you can read Part Three here.