Interoperability in healthcare has become a widely used topic in recent years. Whereas the industry addressed the challenges related to health data exchange as an opportunity to integrate (clinical) IT systems, to date one cannot read or hear about such challenges without coming across the word interoperability. A lack of interoperability is often seen as the cause for limited health data exchange and hence the many inefficiencies in healthcare. Solutions are sought in "increasing interoperability", and in referring to the use of internationally adopted communication standards and protocols in healthcare.
In this series of articles, Founda’s CPO Andries Hamster will address different aspects of interoperability and share some of his learnings in dealing with healthcare interoperability over the course of many years. The main learning? In order to achieve successful interoperability, the real challenge that needs to be addressed is the "mindset" of those involved in health data exchange.
What these "mindsets" are will be further discussed below – in no particular order.
The first mindset is about whose perspective you start from when addressing the need for data exchange. Considering that data exchange is an activity that, at a minimum, requires two participants (the sender and the receiver), many of the traditional exchanges have been built from the sender's perspective.
Generally speaking, the industry distinguishes between push based and pull based exchanges. An example of the first is the Direct Project. Pull based exchanges are also referred to as "Query-based" exchanges. In such exchanges, data is made available for "discovery". The TEFCA framework is a good example of this.
In both types of exchanges it often is the sending side that determines what the recipient side is allowed to receive or discover. This often limits the recipient to effectively use the received health data in its own (clinical) processes. When such limitations are too big, the recipient doesn't perceive value from the health data exchange – often leading to underutilization of the exchange infrastructure that was put in place.
I found a good analogy in "The Patient will see you now" from Eric Topol. Topol explains that it is the doctor who needs to change to accommodate the needs of his patients in order to stay relevant. When translating that to health data exchanges, these should be built around the needs of the data recipients to be successful.
Health data is, for many good reasons, considered to be among the most (privacy) sensitive data types. No one likes to see their very private and intimate data fall into the wrong hands. As a result, and often amplified by health data regulations, we have built quite a number of "defense walls" into our health data exchange infrastructures that are based on zero-trust.. Explicit patient consent and complex multi-factor login methods are just two examples that are built on the assumption that "the other side" of the exchange cannot be trusted. Hence, each "defense" is forming an additional barrier to the health data exchange value recipients perceive.
To lower these barriers, it makes sense to build an exchange on the basis of trust. Trust being the ability for a health data owner to unambiguously verify who the recipient requesting access to its health data is. Technology and standards to do so are available in the form of attribute-based identities in combination with trusted identity providers. National governments can play a pivotal role as – like with our passports – they have the authority to hand out (electronic) identities for both healthcare professionals and patients.
This third mindset may sound like an "open door". However, many health data exchanges see the first and last mile of an exchange as a technical challenge – i.e., how to connect to the data sources and data consumers. The reality is that the first and last mile are all about changing behavior of people involved in the act of data exchange. That makes this mindset a tough one, as there is nothing more difficult than changing the way people work when they are used to doing things the way they do for years. A famous and applicable quote from Albert Einstein is that "If you do what you always did, you get what you always got". Changing behavior is a challenge for healthcare organizations only they can tackle, and it requires an investment in project management capabilities.
Whenever you read the word interoperability, the words international standards and health data exchange are probably found in the same sentence. As a result of 30 year of hard work, it has become commonly agreed upon that in order to solve the challenges associated with health data exchange, the solution is found in applying international communication standards such as HL7v2, DICOM, FHIR, CDA, etc.
Unfortunately, each of these standards leave room for interpretation – leading to a lack of interoperability, despite the claims made by these standards. Glancing over the various health data exchange initiatives around the globe, conflicting choices are made about how to interpret the room these standards allow. These conflicting choices have led to numerous "dialects" (a.k.a. localizations) of the same standard.
The mindset in this case is to be extremely cautious when localizing an existing standard. Each localization carries the risk of becoming a future legacy issue. It is therefore extremely important for organizations such as the international Integrating the Healthcare Enterprise (IHE) to participate. IHE offers a methodology to come to an agreement on how to apply interoperability standards in the context of a given use-case. Although this may feel like a slow moving process (which it often is), it guarantees long term success.
Health data exchange has shown to be a complex challenge in healthcare. Many different systems expose different interfaces, adopt different standards and use different data models. A natural reaction to such a chaotic landscape is to start harmonizing all aspects of data exchange. Many interoperability projects and programs around the globe aim to remove the chaos by working their way backwards from a, sometimes, ideal "model of the real world". Unfortunately, the real world exists – and will continue to exist – of many different actors that take part in health data exchange. The landscape will always be chaotic. Advancing interoperability requires embracing the chaos and dividing the challenges into smaller ones. Each "small" solution is an interoperability building block. Larger challenges can be solved by stacking the building blocks together.
An often witnessed pitfall of interoperability is the quest for the perfect solution. A data model that includes all possible aspects of clinical information, a data exchange solution that fits all use-cases. Of course we should not settle for bad solutions. However, the other side of the spectrum is the risk of over-optimizing solutions for scenarios that rarely occur – to the point that they become useless. I call this the ‘good, better, best, useless’ syndrome, where good solutions are discarded for the fear of not being the absolute best.
Over the years, I’ve learned that the essential elements for successful health data exchange are:
With Founda we are taking these learnings into account to build a new and modern approach to health data interoperability. Check out our white paper if you’d like to learn more about our platform.