Blog - Flogging a Dead Horse – Why updating guidelines and training doesn't drive behaviour change

Flogging a Dead Horse – Why updating guidelines and training doesn't drive behaviour change

Introduction

Use of clinical guidelines has shown improvements to survival, quality of care and cost effectiveness for a health service¹, but they are often not translated into practice. An estimated 30-40% of patients receive treatment that is not based on best practice evidence, and 20-25% receive treatment which is either not needed or potentially harmful³.

Common RCA Recommendations

Every time I have sat on an RCA (Root Cause Analysis) panel we have added at least one of two recommendations for improvement:

Add extra staff training. We need to remind staff to do X to prevent Y happening again (according to procedure Z).
Add a paragraph to procedure Z warning about X to prevent Y from happening again.

We would sit at RCA meetings all knowing full well that the training calendars were already bursting, the training coordinator was going to fight the addition of any more content, and even if they did agree, the training itself wouldn't catch locum or agency staff.

Meanwhile we would pass emails back and forth crafting a statement in a procedure or guideline to fix the problem as concisely as possible. All fully aware that any update to the procedure or guideline was not going to be read unless we did a big internal marketing campaign to highlight the change with staff. This was nigh on impossible with the number of competing interests in a clinician's timetable and inbox.

It was a strange situation; we were actively competing with other departments and other quality initiatives for the limited time and attention that the floor staff had available.

Inherent Problems with Behaviour Change in Healthcare

This is not an uncommon problem; healthcare just doesn't have great options to drive behaviour change.

The RCA process itself was originally invented by the aerospace industry, which has the luxury of implementing 'engineering type controls' such as changes to aircraft design to prevent an incident. Engineering controls are on average 1.6 times more likely to report success compared with those classified as 'administrative controls'².

In addition, if an aircraft has a serious risk identified, it can be recalled or removed from service entirely.

You can't do this in healthcare. Care cannot be held over until it is convenient, and healthcare engineering controls such as purchasing a new device or new digital system are very difficult to get through tight hospital budgets.

Our controls are mainly trying to stuff one more item into training, and updating existing procedures and guidelines until they are bloated full of recommendations for management of the endless edge cases encountered when interacting with human patients.

(people are nothing like aircraft FYI)

Guideline Access and Usage

When I was working in the Safety and Quality unit, I had a routine report sent to me showing how often procedures and guidelines were accessed.

I was always shocked to see major guidelines accessed only a few hundred times a year, and some more specific ones accessed only two or three times in a year, (which I can only assume was by the people who wrote and uploaded them).

I used to think, "how sad, they spent months on that guideline and nobody has even opened it...".

Research into barriers into uptake of guidelines has uncovered that the top four barriers into guideline uptake are:

Length or complexity of guideline documents.
Time constraints due to clinical responsibilities.
Concerns that guidelines do not apply to a single patient.
High number of conditional or weak recommendations.¹

I think these are all valid points. On the floor you don't know who has reviewed a guideline, you certainly don't know if they have considered your patient's clinical situation, and if they have considered everything then you can guarantee that the guideline will be a bloated 10+ page document which will take considerable reading to understand how to apply it to your patient's context.

Limitations of Training and Guidelines

Relying on tools like training and guidelines we are never going to be able to match the aerospace industry for safety, because people are complex!

Out of the twelve or so methods we have available to drive behaviour change, there is no denying that enabling desired behaviour (via engineering type controls) is the most effective method, but my argument is guidelines and training can work only until a certain point of complexity.

We have reached that point.

Nobody wants to do the wrong thing, and clinicians fundamentally want to access guidelines more frequently, but are heavily restricted by the four barriers above, and as medicine becomes more complex these barriers will only continue to grow.

If we take away the complexity, time constraints and generalised nature of the guideline, and ensure clinicians can see the review process openly then we will have a much better chance of driving behaviour change.

Introducing Clinical Branches

This is where the bias in this blog post starts.

To do this I think we need a whole new medium; the use of plain text is no longer going to cut it.

There might be other ways to do this, I'm sure many will turn straight to AI, but in my opinion we have gotten it right with Clinical Branches.

Using this platform, we can replace text guidelines with structured decision trees (or pathways) built using a custom canvas for the job.

These are developed to guide clinicians to information relevant to their patient's context. When the clinician views the pathway it converts to a single page application which can be interpreted rapidly, with an average completion time of 48.1 seconds.

Training on unusual but high-risk cases is no longer required. You no longer need to wonder whether the guideline is appropriate for your patient, because you have answered all the branch points and know the information you have is context specific.

Lastly, much like in the software industry, people can view the review process in real time and openly collaborate, building trust in the medium, removing the final barrier to guideline uptake.

Conclusion

I acknowledge that I have a strong bias in this area, but I think people out there who have tried to change behaviour in a public hospital will agree with me, something is needed to better structure our workflow and really drive behaviour change.

I'm hoping that tool will be Clinical Branches for many, and at my next RCA I am hoping to say; "lets just add a new branch for that".

References:

Written by John Shanks - Antimicrobial Stewardship Pharmacist and Software Developer at Kraken Coding