The issue of fit between the work and needs of clinicians (medical, nursing and other) together with the clinical information systems (CIS) they are expected to use during the course of managing patients, is one that has been gaining increasing amounts of attention over time. This is particularly true of large-scale CIS. That is to say, Electronic Medical Record systems (EMRs).

In fact I’d argue that the attention this issue has garnered on the international stage has increased proportionately to the increase in the number of clinicians routinely using such systems. Of course, that number of clinicians will no doubt bear a direct relationship to the number of CIS installations. Even the head of IT at a major international EMR vendor has acknowledged this issue, as has the Office of the National Coordinator for Health IT (the ONC) in the US.

What is now quite clear is that there is a worrying amount of discontent amongst many clinicians as to the LACK of fit between their needs and work practices, and what many of these systems or implementations allow. This lack of fit is manifest in complaints of, and evidence around, clinician dissatisfaction, patient safety risks, excessive (system driven) workload, excessive cognitive load, and undue complex system interactions (including unduly complex navigation/ interaction patterns).

In turn there are anecdotal reports of – and evidence around – resultant: clinician burnout with this lack of fit and related frustration as the source of: actual patient safety events, malpractice claims / prosecutions, clinicians developing workarounds to bypass systems or some of their intended functions and various other outcomes. The AMA in the US is responding to this burnout problem at a programmatic level through their “Steps Forward” program.

There are MANY, MANY reasons for this apparent lack of fit. It’s a PhD-worthy topic several times over in fact. One set of reasons are those outlined recently by Lintern and Motavalli.

In short – there is an issue – this is clear. Some of the major international EMR vendors are responding to this issue. For example, by changing software design philosophies (e.g. – responsive design built-in as a core approach to product development) including, efforts to boost their internal capability around usability and user experience (UX) (personal communication).

What I think is a much greater and more interesting issue therefore is, how do we as a healthcare industry respond to this very important and global challenge?

In Australia, many of those interested in this problem are working through the HISA UX Community of Practice to try and address it.

For those deeply interested in this topic, I’d suggest that they read a recent article in the Communications of the ACM by Andriole (Dec 2017; Vol 60. Num 12 pp 29-32). Despite this being an IT professional magazine, it’s not a hard read for the non-technical amongst us. Interestingly in this article entitled “The Death of Big Software”, the author is speaking from an industry agnostic perspective. Yet much of what he is saying rings true in relation to the history of CIS’ in healthcare and the core issue of lack of fit outlined above.

Without repeating the entire article, the following key quote says much about the main message of the author… “Software architectures must be blank canvases capable of yielding tiny pictures or large masterpieces”. Whilst such quotes have a tendency to grossly oversimplify the ENORMOUS complexity of building non-trivial software systems, there is a core truth in this statement in my view. I believe that heading in this direction is ONE key way to address the lack of fit described earlier.

Alternate system development approaches and architectures have also been created as one way to address these issues. See the work of Prof Jon Patrick and his team who have built greater configurability – by clinicians – into their core system architecture.

The rise of FHIR, APIs and “ecosystems” is to be commended and should offer some assistance but this may also bring unintended issues around integration of workflows with implications for usability and UX. This is one down side to a “best of breed” approach to CIS’.

Involving clinicians in all stages of the software development/deployment lifecycle (SDev/DepLC) is critical. However, from my experience, current and former clinicians have already been involved for a long, long time inside medical software companies. I don’t believe that the solution is as simple as “adding more clinicians” to the processes. When it comes to health IT clinicians are DEFINITELY NOT a homogenous group by any measure. I could easily write an entire separate paper on this issue of clinician engagement with, and contributions to, these processes. It’s a very complex issue. Lintern and Motavalli offer some valuable insights on this very point – the RIGHT people with the RIGHT skills need to be involved in all parts of the various SDev/DepLC processes.

What can be done?

Educate vendor employees, many of whom already know about these issues and the relevant knowledge bases. But also educate healthcare employees, clinical informatics teams, participating clinicians, IT departments and executives about usability and UX will have some impact especially where systems or implementations allow their subsequent input to affect outcomes in this space.

Whilst there are some usability standards specifically pertaining to CIS’, more functionally- oriented CIS usability and UX standards (or at least some agreed measures to start with !!!) are also important here. This is where clinicians can have a key role.

Let me use a simple example to explain further. I’m aware of a major tech organisation that in recent years (this tech, in different forms, is now widespread) set about developing a rapid tap-on, tap-off PC access system (to avoid the constant typing of passwords as clinicians move between PCs, for example on a ward round).

The idea is, the clinician walks up to a PC and can tap a card on a device plugged into the PC. This action automatically authenticates them on the network and they are logged on to that PC. The stated benefit was to achieve system access (the clinician being able to use the PC) within 10 seconds of the clinician standing in front of the machine and tapping on. So, in this case, THIS key feature (10 sec measure) of the usability of the tech was the primary goal around development… and it was achieved.

If we now relate this example back to a typical CIS SDev/DepLC process. The system requirement would have been expressed as: “the system shall allow the clinician to obtain tap on access through use of an authentication card” or similar wording with no reference to the speed aspect. Not surprisingly, vendors have built and sold to such functional specifications and at a macro level, the industry has historically accepted that as a standard.

The use of such usability and UX standards, tied to procurement processes (“how well / quickly / satisfactorily can the system do X” vs “can the system do X”) should have a role in addressing the poor fit outlined earlier.

CIS procurement approaches to date have been far too focused on pure functionality, corporate risk and cost. Yes, these things are important although the balance is currently “out of whack” and we need to collectively push for MUCH greater integration of usability and usability standards into procurement processes.

Of late even more information has come into the public domain in the form of public anecdote, commentary and research on this topic all of which continues to point to the issues caused by poor usability of EMR systems and the resultant suboptimal user experience.

When set in the Australian context it should give us even greater cause for concern.

I make that statement for 2 reasons. Firstly, despite recent announcements about some Australian hospitals being certified at HIMSS EMRAM Level 7 as well as many ongoing EMR projects around the country, which by the way we welcome with open arms, we are still fairly early in our EMR journey when viewed from a national perspective. So ahead of us lays a great opportunity to take the lessons of more advanced nations about the usability and impact of EMRs on clinical staff and the healthcare system and avoid making exactly the same mistakes. Unfortunately, I see no obvious evidence of that happening however.

Secondly, and of greater concern is that this apparent inability or unwillingness to incorporate the lessons of others into how we implement and manage EMRs is set against the backdrop of an exceedingly low-level knowledge about UX and usability across the healthcare sector. These 2 phenomena can obviously be quite related!!!

So after conceiving of and setting up the HISA UX CoP almost 3 years ago, I have given probably several dozen talks on this topic and made numerous efforts to get the message out in social media about these issues. What is very apparent though, is that even audiences literate in health information or health IT have very little awareness of the knowledge bases underpinning usability and UX. Many have never heard of Nielsen and his heuristics, as one fairly scary example. These “co-located” phenomena should – unfortunately- continue to give us much cause for concern.

The US spent (I believe) $40 billion USD in a push for EMR uptake across their country and now EMRs are almost ubiquitous there. But many usability and UX problems have resulted, as I have outlined above. The ONC openly acknowledges this and states that usability is in their top 2 Health IT issues to be resolved (along with Interoperability… and we all know THAT pain).

So the challenge is in front of us. In effect it’s a long battle to be fought on multiple fronts and which needs to be fought in collaboration with all relevant parties including clinicians, informatics and IT professionals, vendors, funders and other relevant parties. Will we collectively fight that battle or will we, VERY UNNECESSARILY in my view, learn these lessons AGAIN ourselves? Can’t we take an easier path?

Professor Chris Bain

Professor of Practice, Digital Health – Faculty of IT, Monash University

 

Professor Bain is an experienced clinician (former) and health IMT practitioner with a unique set of qualifications, and a unique exposure to broad aspects of the healthcare system in Australia. He also has extensive experience in designing, leading and running operational IMT functions in healthcare organizations. His chief interests include the usability of technology in healthcare, data and analytics, software and system evaluation, technology ecosystems and the governance of IT and data.