**This is a blog post to accompany my presentation at the LAA CGD on 6/5/24

I’ll start with a story

The White Star Line Fleet
In 1912, the RMS Titanic set sail, deemed ‘unsinkable’ by its creators.

We all know how it ends – a tragic collision with an iceberg, leading to more than 1,500 dead.

However, there’s a lesser-known part of this story that often gets left out.

The SS Republic
Several years before the Titanic set sail, another ship in the White Star Line fleet, the SS Republic, sailed the same icy waters of the North Atlantic.

The SS Republic encountered numerous icebergs. Its captain, Captain Sealby, negotiated them successfully, and meticulously documented each encounter. He recorded detailed notes on the relevant coordinates and manoeuvres needed.

All his learning points were stored within the company’s logbooks, with a view to informing future voyages.

Unfortunately, these records were either lost in the archives or not effectively communicated to the Titanic’s designers and crew.

Captain Sealby’s documented knowledge could have informed better design and navigation protocols.

Instead, the same waters claimed the Titanic, a disaster that might have been mitigated.

The lessons of the titanic story are usually the dangers of human overconfidence and the brutality of mother nature.

However, arguably even more importantly, it illustrates the dangers of poor institutional memory [1].

What is institutional memory?
It is the retention of knowledge in an organisation.

Institutional memory ensures high performance is sustained despite the challenges of high staff turnover, flash teams, new evidence and guidelines, a changing political landscape, and of course unpredictable patients.

The institutional memory of an organisation is the anchor that keeps it steady while navigating through the changing tides of time.

In the PRU there are important changes on the horizon in terms of our geographic footprint, and the separation from LAA in the next few years. Growing pains are sure to be present, along with a sense of new beginnings. My belief is this heightens the importance of reinforcing our institutional memory.

If lasting institutional memory is the goal, we need to optimise how we manage knowledge.

Knowledge management
Ultimately, this presentation is about the creation and management of human knowledge in emergency healthcare organisations.

Formal strategic knowledge management (KM) is a relatively new concept in healthcare [2].

We are knowledge workers. Our work involves complexity, ambiguity, interdependence, and is constantly evolving. Therefore, KM has been informally present in our world for a long time.        

However, the deliberate integration of formal KM processes (inspired by the corporate world) into healthcare systems has only gained momentum in recent years. Its emergence has been driven largely by technological progress – machines [3].

The ideas I will be sharing today may, therefore, be new to you.

My recent journey
I’m going to talk about my own journey in the past two years.

I will share project work from my time with Sydney HEMS, and work that is ongoing with the PRU that my team and I are proud of.

Marginal gains
The prehospital arena is the perfect muse for project work in this area.

We have the advantage of attending to one patient at a time, coupled with the relatively protected time for debrief in the car after each case – arguably the perfect workflow for knowledge creation.

I believe that every unit of knowledge generated and processed by the organisation that ultimately leads later to another clinician internalising it is a marginal gain.

Marginal losses
I also believe that literally hundreds (no exaggeration) of opportunities are missed every week.

Each time a learning point is discussed but not captured, the institution loses a piece of its memory.

What I’ll cover (broadly)

• What is knowledge?
• Knowledge creation: the “SECI” model
• Coffee and Cases
• Improved knowledge management in the PRU
• The (near) future

What is knowledge?
Firstly, I’ll say what knowledge isn’t. It isn’t data or information.

Data is simply raw values or measurements. For instance, the raw numbers seen in a patient’s systolic blood pressure readings.

Information is data with added context. It encompasses the Who/What/When/Where of that systolic blood pressure.

Knowledge is information enriched with experience, which is actionable.

It is subjective and incorporates how an individual perceives the world—through their lens and values [4,5].

For example, knowledge in this context could be the mental checklist of potential causes of undifferentiated hypotension.

Knowledge is about action. An individual always has knowledge “to some end”.

If there is no associated action, it is merely information.

Knowledge can be broken down into two types.

Tacit knowledge
Tacit knowledge encompasses highly personal and deeply rooted experiences, insights, and intuitions that are difficult to articulate and formalise (i.e. difficult to write down). It is context dependent.

It includes practical skills (“know-how”), as well as cognitive dimensions such as mental models and nuanced perception, which are inherently subjective.

Explicit knowledge
Explicit knowledge is easily articulated and communicated in words, making it simple to process, understand, and store in databases.

It is “codifiable”.

Good examples are our SOPs, checklists, guidelines, and clearly documented learning points from educational discussion.

The “SECI” model
The “SECI” model is a foundational knowledge creation theory developed by Japanese academics Nonaka and Takeuchi in the 90s, rooted in Japanese business practices from companies like Honda, Canon, and Nissan [8].

Despite its nation-specific origins, it has profoundly influenced how modern organisations worldwide manage knowledge.

The premise of the model is that knowledge creation results from fostering continuous conversion of tacit into explicit knowledge and vice versa.

It describes four modes of conversion: “Socialisation”, “Externalisation”, “Combination”, and “Internalisation”.

I hope to convince you that my QI work in Sydney and here in London aligns with this model!

“Socialisation” (社会化 – Shakaika) is where tacit knowledge is shared directly between individuals. This happens via face-to-face interactions, direct observation, shared experiences, and debriefing sessions.

“Externalisation” (外化 – Gaika) converts tacit knowledge into explicit knowledge via dialogue, reflection, and deliberate documentation.

“Combination” (結合 – Ketsugō) aggregates and organises different pieces of explicit knowledge into more complex sets (in a streamlined knowledge management system). Examples are SOPs, guidelines, and searchable databases.

“Internalisation” (内化 – Naika) embeds explicit knowledge into individuals’ tacit knowledge through doing, and so a deeper understanding is achieved. In other words, an individuals’ mental model evolves for higher performance.

This continuous, cyclical process – the “knowledge spiral” – fosters continuous knowledge creation in an organisation [5].

So with all of that theory in mind, let me take you to Sydney…

Sydney HEMS
As mentioned, I was a HEMS fellow in Sydney last year. A dream of mine for many years was to work in that service… and so I arrived full of positivity, expectation, and a considerable case of Imposter Syndrome!

I then sim’d and workshopped and drilled during a phenomenally intense, and immersive month-long induction process. Without a doubt the standout educational experience of my career.

However, after being signed off I had fairly significant performance anxiety.

“Coffee and Cases”
To my rescue came Coffee and Cases. A voluntary meeting for all clinicians on base, and anyone wanting to join remotely, which (as the name suggests) involves caffeine and talking shop.

It happens every single day at the Bankstown Base at 10am. Including Christmas day.

The atmosphere was intentionally casual, with participants offering cases spontaneously for discussion. Conversations frequently deviated off on tangents and down educational rabbit holes, moving far from the original case. Both successes and insightful failures were enthusiastically acknowledged.

I’d never encountered such candour and open vulnerability from colleagues, especially from leaders in the organisation.

I felt brave enough to speak up about the things I was particularly nervous about or lacking confidence in and ask questions that previously I may not have asked through fear of appearing incompetent.

With each question, I was able to resolve a little more ambiguity in my own mind.

“Socialisation”
The questions I asked were always welcomed and stimulated loads of discussion, exchanging of ideas, sharing of mental models, and further questions.

I felt like I was pouring quick-drying cement between the bricks that had been laid during my induction.

Learning the SOPs and simulating cliff face RSIs were obviously essential for being mission-ready, but Coffee & Cases was the key ingredient for building my confidence. It was a daily shot in the arm.

That confidence came from absorbing the tacit knowledge of experienced colleagues. We were “socialising”.

Japanese car companies have a similar process called “Ba” (場). Except they drink Sake instead of coffee.

Interestingly, whenever anyone said “I don’t know”, particularly if it was a senior person, the knowledge exchange that followed seemed to turbo-charge. I think it’s an underestimated phrase, especially when said by an organisational leader.

Psychological Safety
Coffee and cases is a psychologically safe environment. That was the engine that powered its magic [6].

I think psychological safety is commonly discussed in the context of reporting incidents and CRM, but it is equally important for fostering knowledge creation.

During C&C all attendees, up and down the hierarchy and across disciplines, felt safe to speak up in the face of reputational damage (i.e. appearing incompetent). Ideas and questions, however stupid, felt natural and welcomed. Remaining silent (the easy option) always felt like a missed opportunity.

I felt fearless in my pursuit of knowledge.

However, I noticed a problem
Knowledge created during C&C was staying in the room.

If you weren’t at the party, you didn’t get to “socialise”.

I could only attend so many sessions. Sure, I could tune in via Microsoft Teams, but it’s a big commitment to do that regularly when not at work, especially with a 1-year-old at home.

In some sessions I noticed an online form being filled out – an electronic record of cases discussed, supposedly with documented learning points. Perhaps I could read the highlights for the sessions I missed…

It took some digging to find out where this record was kept.

Signposting not documenting
When I eventually found it, I felt C&C documentation (if done at all) was largely the bare minimum, and certainly no reflection of the awesome learning conversations taking place.

There was a lot of this…

“Discussion around transfer to regional trauma centre vs MTC.”

“Challenges of managing the patient who is agitated due to hypovolaemia.”

“Human factors issues – decision making, directing traffic.”

Flagging/signposting that discussions took place, which is of no educational value.

“Externalisation”
Tacit learning was not being adequately “externalised”.

I proceeded to design and lead a quality improvement project.

My team and I campaigned for better documentation of learning points and opportunistically coached colleagues.

Best practice was decoupling learning from the case, so that it could be disseminated throughout the organisation. Colleagues would benefit from the knowledge created without needing to be present for the case discussed.

“Shoulder dislocation? Methoxyflurane works really well for muscular relaxation and is safer than sedation.”

“There is crossover between anaphylaxis and asthma, especially in kids – IM adrenaline is a good idea if the patient is sick.”

“Don’t make assumptions about a patient’s level of intoxication. External evidence of a head injury? TBI till proven otherwise.”

Coffee & Cases “Snippets”

We started disseminating knowledge as branded “Snippets” through a variety of means, including a rolling slide-deck on screens across all aeromedical bases in New South Wales, a monthly digital newsletter, content on the Sydney HEMS website, and ultimately a searchable Snippets database.

Post-production
All content is vetted by one or two HEMS consultants (if a second review is requested by the first reviewer) before dissemination.

Approximately 30-40% hits the cutting room floor.

Knowledge is linked to relevant SOPs, external guidelines, and pertinent further reading materials. Additionally, individual points are tagged with items from the prehospital and retrieval medicine curriculum in Australia.

“Combination”      
This is an example of “combining” explicit knowledge items to create more reliable knowledge creation.

The Physician Response Unit
My foray into knowledge management continued when I joined the PRU in London, with another quality improvement project starting in February this year.

I have iterated the system of capturing and processing knowledge that I built in Sydney and applied it here.

My team and I have rolled out this system to all educational platforms in the PRU.

This represents a broadening of the project’s scope compared to Sydney, where it encompassed Coffee & Cases sessions only. Although just as I was leaving the M&M team also started employing the process.

PRU Learning Activity Log
We have introduced standardised Googleform for recording learning for all educational encounters. From moulages to D&D.

It ensures learning points are itemised, and relevant SOPs can be linked.

It emphasises documentation discipline:

• decouple learning from the case
• don’t just signpost, describe learning points in full

All content goes through the same post-production process that was established in Sydney (i.e. consultant vetting prior to dissemination).

Approved content is disseminated via email, MS Teams, and uploaded to the governance app for D&D sessions.

CEM (Community Emergency Medicine) Coffee & Cases
We have introduced a new educational platform – CEM C&C (paying homage to the process in Sydney) – that is currently happening once per week. My hope is we manage to increase that frequency in due course.

It is an opportunity to discuss cases in a more protected and decompressed environment than PRU D&D, which usually happens in front of an audience of ED doctors in the ED hub (an unavoidable obstruction to maximum psychological safety).

Cases are offered up by participants informally. There is freedom to go down rabbit holes, play “what if”, and share intelligent failures.

Ample space for maximum vulnerability.

PRU case load
Our case load in the PRU is far more weighted to complex medical and frail patients than critical care.

Understandably, critical care often takes centre stage in our D&D process. CEM C&C provides a valuable additional opportunity to exchange tacit knowledge about our more routine cases, which can be extremely challenging in terms of human factors and from medical ethics perspective. These cases often carry a significant emotional toll.

For me personally I have found it a great opportunity to unpack some end-of-life care cases that I felt conflicted over. Sharing the emotional burden with colleagues has been powerful.

Anecdotally many of my colleagues, particularly EMTs (who are less used to the doctor-centric ED hub environment), have found CEM C&C dovetails neatly with D&D.

Examples of approved learning points from CEM C&C:

“There may be situations where you are unable to action your preferred plan (e.g., EOLC pathway vs treating an unconvincing infection to satisfy fixed beliefs of patient/relatives). Your wishes may not align with your patient’s. In these situations that you must remain committed to patient-centred care and reach a compromise.”

“EOL conversations are challenging. A strategy that works with one family may not work with another. One strategy is to use a simple analogy the family can relate to. See “The Paper Boat” article (link below).”

Hot debrief
We have also started logging knowledge created during hot debriefs, where tacit knowledge is fresh.

Some particularly intriguing points have been documented which feel likely to have been lost if not logged immediately after the event.

“A non-invasive strategy for getting the attention of a colleague during a heightened scenario is to place a hand on the shoulder. A good example is when the PRU doctor is locked in a discussion with eyes off the patient, and the EMT needs to re direct them to a clinical priority.”

“When you encounter frail/demented patients, don’t assume they want you to talk to the relative/carer during the clinical encounter. A good opening gambit is to ask the patient who they would like you to ask questions to and proceed from there.”

PRU Knowledge Library
On MS Teams there is excel table containing all approved content since February earlier this year. It contains a total of 140 learning points (or units of knowledge).

However, it is clunky.

I felt strongly that there must be clearer, cleaner, and more powerful ways to present the knowledge to colleagues.

App prototype
So I have created an app prototype on Google Appsheet.

It was very straightforward (no coding knowledge required). Rolling this out to the whole service is full of obstacles of course, not least the cost.

Nonetheless I put it together to demonstrate what is possible when knowledge management is done deliberately, and it is prioritised by the institution.

Internalisation
It is worth returning to the SECI model and looking at the fourth mode of knowledge creation – “internalisation”.

This occurs when individuals apply explicit knowledge in real-world or simulated scenarios, leading to more tacit knowledge—i.e., enhanced intuition, appreciation of nuance, and a more sophisticated mental model.

A digital knowledge library, developed through the processes of “socialisation”, “externalisation”, and “combination”, serves as an invaluable training tool, discussion aid, and reference manual (an adjunct to SOPs). It fosters the practice-driven “internalisation” of contemporary explicit knowledge by facilitating all staff staying up to date with every educational moment occurring within the organisation.

Moreover, it preserves valuable knowledge shared by staff who move on, ensuring their insights are not lost but continue to benefit the organisation.

Preserving legendary minds
By optimising our knowledge creation process, we can capture and codify legendary wisdom before it retires.

Our SOPs and workflow protocols are indeed built on the shoulders of historical giants within our organization, but their tacit knowledge extends beyond these documents.

For instance, a 15-minute chat with Bill Leaning, who is considering retirement in the next few years, yields extraordinary insights on almost anything related to prehospital care. His understanding of all aspects of the field is something only achievable after many years of honing his craft. We now have the machinery to codify it.

I also think about my dad, a 70-year-old GP who is desperately dodging retirement out of fear of the gaping hole it will leave in his life. He is also a giant in his world, and his tacit knowledge must live on when he eventually hangs up his stethoscope.

The tacit knowledge of Bill and my Dad must not meet the same fate as that of Captain Sealby of the SS Republic, whose crucial insights were lost in the archives.

Final point.. artificial Intelligence
The emergence of AI has massive ramifications for acute healthcare. While we can’t be certain of its precise role at this stage, there is little doubt that our complex decision-making will be augmented by machines in a matter of years.

I don’t view this as an existential threat, but rather as an opportunity to enhance patient safety. The question is, when we do arrive in that moment, how do we ensure we can trust the machine?

I believe the answer lies in building a digital library of knowledge, ensuring the content is up to date, vetted by senior people, and aligned with organisational values, policies, and procedures. This will ensure that when we employ AI technologies – such as natural language processing and machine learning – to interact with the digital library as if it were a human, we can trust the answers it provides.

In the PRU, we are already taking steps to achieve this.

Summary
I’ll finish with what I believe to be the key ingredients for organisational knowledge creation, and building institutional memory:

• Psychological safety. This is always a work in progress for all organisations.
• Coffee and Cases (or similar). Informal professional chatter should be an organisational priority.
• Disciplined documentation. Decouple the knowledge from the case.
• Post-production diligence. Ensure senior vetting prior to dissemination.
• A digital knowledge library. Regularly updated. Combined with SOPs. Ready for deployment with AI comes for us.

References

1. https://www.whitestarhistory.com/
2. Wager, K. A., Lee, F. W., & Glaser, J. P. (2017). Health care information systems: A practical approach for health care management. John Wiley & Sons.
3. Fong, B. Y. F., Fong, A. C. M., & Li, C. K. (2016). The role of technology in knowledge management. Journal of Knowledge Management, 20(4), 738-758.
4. Polanyi, M. (1966). The tacit dimension. University of Chicago Press.
5. Nonaka, I., & Takeuchi, H. (1995). The knowledge-creating company: How Japanese companies create the dynamics of innovation. Oxford University Press.
6. Edmondson, A. C. (2018). The Fearless Organization: Creating Psychological Safety in the Workplace for Learning, Innovation, and Growth. Wiley.