The ultimate PHR?

I’ve been interested in the notion of a Personal Health Record for a long time.

I was involved in the development of HotHealth, which launched at the end of 2000, a not-so-auspicious year, given the dot com crash! By the time HotHealth was completed , all the potential competitors identified in the pre-market environmental scan were defunct. It certainly wasn’t easy to get any traction for HotHealth take-up and yet only recently it has been retired. For a couple of years it was successfully used at the Royal Children’s Hospital, cut down and re-branded as BetterDiabetes to support teenagers self-manage their diabetes and communicate with their clinicians, but it wasn’t sustained.

This is not an uncommon story for PHRs. It is somewhat comforting to see that the course of those such HealthVault and GoogleHealth have also not been smooth and fabulously successful

Why is the PHR so hard?

In recent years I participated in the development of the ISO Technical Report 14292:2012: Personal health records — Definition, scope and context. In this my major contribution seemed to be introducing the idea of a health information continuum.

However in the past year or so, my notion of an ideal PHR has moved on a little further again. It has arisen on the premise of a health record platform in which standardised health information persists independently of any one software application and can be accessed by any compliant applications, whether consumer- or clinician-focused. And the record of health information can be contributed to by any number of compliant systems – whether a clinical system, a PHR or smartphone app. The focus is on the data, the health record itself; not the applications. You will have seen a number of my previous posts, including here & here!

So, in this kind of new health data utopia, imagine if all my weights were automatically uploaded to my Weight app on my smartphone wirelessly each morning. Over time I could graph this and track my BMI etc. Useful stuff, and this can be done now – but only into dead-end silos of data within a given app.

And what if a new fandangled weight management application came along that I liked better – perhaps it provided more support to help me lose weight. And I want to lose weight. So I add the new app to my smartphone and, hey presto, it can immediately access all my previous weights – all because the data structure in both apps is identical. Thus the data can be unambiguously understood and computed upon within the second app without any data manipulation. Pretty cool. No more data silos; no more data loss. Simply delete the first app from the system, and elect to keep the data within my smartphone health record.

And as I add apps that suit my lifestyle, health needs, and fitness goals etc, I’m gradually accumulating important health information that is probably not available anywhere else. And consider that only I actually know what medicines I’m taking, including over the counter and herbals. The notion of a current medication list is really not in the remit of any clinician, but the motivated consumer! And so if I add an app to start to manage my medications or immunisations this data could be also used across in yet another compliant chronic disease support app for my diabetes or asthma or…

I can gradually build up a record of health information that is useful to me to manage my health, and that is also potentially useful to share with my healthcare providers.

Do you see the difference to current PHR systems?

I can choose apps that are ‘best of breed’ and applicable to my need or interest.

I’m not locked in to any one app, a mega app that contains stuff I don’t want and will never use, with all the overheads and lack of flexibility.

I can ‘plug & play’ apps into my health record, able to change my mind if I find features, a user interface or workflow that I like better.

And yet the data remains ready for future use and potentially for sharing with my healthcare providers, if and when I choose. How cool is that?

Keep in mind that if those data structures were the same as being used by my clinician systems, then there is also potential for me to receive data from my clinicians and incorporate it into my PHR; similarly there is also potential for me to send data to my clinician and give them the choice of incorporating this into their systems – maybe my blood glucose records directly obtained from my glucometer, my weight measurements, etc. Maybe, one day, even MY current medicine list!

In this proposed flexible data environment we are avoiding the ‘one size fits all’, behemoth approach, which doesn’t seem to have worked well in many situations, both clinical systems or personal health records. Best of all the data is preserved in the non-proprietary, shared format – the beginnings of a universal health record or, at least, a health record platform fully supporting data exchange.

What do you think?

“We have the capability!”

We have had the technology for a purpose-built openEHR-compliant ‘plug & play’ platform for some time; standalone applications have been built, but just recently it appears that the practical reality of an multi-application platform is also about to happen.

“We have the technology. We have the capability…” Stay tuned.

…Reminds me of my 1970′s hero, Steve Austin, the Six Million Dollar Man. With apologies to Steve:

“We have the technology. We have the capability to make the world’s first universal health platform. openEHR will be that platform. Better than ever before. Robust health data…application independent…semantically interoperable!”

Preserving health data integrity

How valuable do we really think health data is? How seriously do we take our responsibility to preserve the integrity of our health data?

Probably not nearly as much as we should.

Consider the current situation of most clinicians or organisations when purchasing a clinical EHR system. What do they look for? Many possible answers are obvious, but there is one question that I suspect very few are asking. How many consider what data they will be able to export and convert to another format, preserving the current data integrity, at the end of the typical 5-10 year life span of the application? Am I wrong if I suggest it is not many at all?

Despite all the effort that we clinicians put into entering detailed data to create a quality health record we don’t seem to often consider the ” What next?” scenario. How much and precisely what data will we be able to safely extract, export, transfer or convert into the next, inevitable, clinical system? Ironically, we are simultaneously well aware that clinical systems have a limited technical life span.

Any and all of the health data in a health record is an incredibly valuable asset to the holder, to the patient (if these are not the same entities) and to those downstream with whom we may share it in the future – in terms of $$ invested; manpower resources used to capture, store, classify, update and maintain it; and not least the future value that comes from appropriate and safe clinical decisions being made upon the integrity of existing EHR data.

Yet we don’t seem to consider it much… yet. However, as more clinicians are creating increasing amounts of isolated pockets of health data, we should be thinking about it very hard.

Every time we change systems we put our health data at risk – risk of absolute data loss and risk of possible corruption during the conversion. The integrity of health data cannot be guaranteed each time it is ported into a new system because current methods always require some kind of intervention – mapping, transformations, tweaking, ‘cleaning’, etc. Small errors can creep in with each data manipulation and which over time, can compromise the safety and value of our health data. On principle we know that the data should not be manipulated, but being limited by our traditional approach to siloed EHR applications, we have previously had little choice.

We need to change our approach and preserve the integrity of our health data at all costs. After all it is the only reason why we record any facts or activity in an electronic health record  – so we can use the data for direct patient care; share & exchange the data; aggregate and analyse the data, and use the data as the basis for clinical decision support.

We should not be focused on the application alone.

Apps will come and go, but we want our health data to persist – accurate and safe for clinical use – beyond the life span of any one clinical software application.

I’ve said this before, but it’s worth saying many times over:

It’s. all. about. the. data.

One of the key benefits of the openEHR paradigm is that the data specifications (the archetypes) are defined independently of any specific clinical system or application; are based on an open EHR architecture specification; and are publicly available in repositories such as the Clinical Knowledge Manager. It means that any data that is captured according to the archetype specification is directly usable by any and all archetype-compliant systems. Plus the data is no longer hard-wired into a proprietary application so that it is orders of magnitude easier to accurately share or transfer health data than it has before.

Clinical system vendors that don’t directly embrace the archetype-technology may still ‘archetype-aware’, and can choose to use the archetype specifications as a means to understand the meaning of existing archetyped data and integrate it appropriately into their systems. Similarly they can map from their non-openEHR systems to the archetype specifications as a standardised method for data export and exchange.

The openEHR paradigm enables potential for archetype-compliant systems to share the same archetyped data repository – along the lines of an Apple platform ‘plug & play’ approach, with applications being added, removed or updated to suit the needs of the end-users, while the data persists intact. No more data conversions needed.

Adapted from Martin van der Meer, 2009

Now that’s good news for our health data.