Antibiotics and behaviour change at the Health Foundation with BSAC

I had the pleasure of attending a meeting at the Health Foundation last week, courtesy of the BSAC. It was a great multidisciplinary event with microbiologists, junior doctors, intensivists and, perhaps most importantly, social scientists - of many different ilks. Here are some slightly unstructured thoughts (NB these are not verbatim notes and may not reflect what was actually said!)

Charis Marwick - Scottish Infection Intelligence Platform (IIP) : an attempt to get innovative ways to integrate data to support clinicians. Some interesting data that changes to surgical prophylaxis in orthopaedic surgery towards aminoglycosides led to more AKI (which was published here). New data shows that this trend has been reversed with a change in guidelines. We haven't seen this in SW England when we looked, but not sure our data is as robust.

Charles Vincent - On safer healthcare. His new book is freely available here thanks to the Health Foundation. We need to move from assurance to inquiry. Are boards and regulatory bodies up for this challenge? It's a profoundly different form of governance. But one that is more honest (what target is not gamed) and the basis of improvement. We need to think about how we build resilience. Patients themselves may be a key resource in monitoring safety and we need to think about how we do this. Is this what the CQC were getting at with their proposal to use social media as the canary in the coalmine?

Esmita Charani  - The importance of understanding culture and team dynamics. There is too much focus on guideline adherence. There is not enough focus on decision making, team working etc. For example, what do you find when you study why prescribers make the decisions they do? You see a prescribing etiquette. There is a reluctance to interfere with the decisions of others; an accepted non-compliance, with emulation of peers a far stronger determinant of behaviour than guidelines; and a clear hierarchy (juniors will defer to consultants). Any attempt to think about stewardship as a set of guidelines and audits is clearly magical thinking. It's hardwork engaging with this sort of deeply engrained behaviour, but we've got to do it.

There is also just too much intervention, and generally too much of everything. There is no time or space for thinking about what is right, or considering unintended consequences.

She finished with a lovely metaphor - all our systems are trying to build bridges across canyons of sub-optimal practice

. 

But this doesn't reflect the reality of chaotic, complex, adaptive systems that will never respond in predictable ways to simplistic interventions. We're dealing in a crowded market place, and we need to get in there and get our hands dirty.

Peter Davey suggested that perhaps we have too much focus on antibiotics...maybe we need to be focussing on higher level principles. In some situations antibiotics are just not the big medication problem. It might be warfarin, It might be insulin. By forcing people to focus on an issue that has little salience to them, maybe we reduce our credibility. Perhaps we need to be wiser in how we pick our battles.

Building on work I've been doing on primary school governance, perhaps there is a parallel in the way I have seen headteachers set a number of 'non-negotiables'. For instance, everyone must follow the marking policy. But within this framework, teachers have complete freedom to deliver purpose. By trusting and supporting people you can profoundly improve performance. There are echoes of Laloux's Reinventing Organisations here. One of the most profound books I've read in recent times.

Nick Sevdalis. Why has the WHO surgical checklist failed to live up to its promise as originally reported? After all, what other intervention has been discovered that could have the profound effects that were reported? He quoted Lucian Leape :

"The likely reason for failure is that it was not actually used"

So they studied the reasons. This was not a compliance audit - it was a proper sociological study of what actually happens at the interface of guideline and practice. A quick nod here to the concept of mindlines that I came across just after this. 

What they found, essentially, was that the checklist may have been implemented in body, but not in spirit. eg. in only a third of cases did the team actually pause. In other words, implementation was not as intended. The consequence of this is vividly portrayed in "Wrongfooted" in which a failure to follow the checklist led to the wrong foot being operated upon.

In general, what we see when these guidelines come out is a complete lack of strategy for how to implement. If you were going to be sensible, you'd focus on the bright spots and use concepts such as dissonance to promote behaviour change. Instead we drop the guideline from a great height and are surprised with the complete lack of engagement on the ground. We perhaps need to differentiate the strategic from the operational. I'm not completely sure about this - I think we need a common purpose at all levels of the organisation - perhaps the difference is how we talk about this depending on context.

So what we see with poor implementation of research or other good practice is :

• lack of fidelity (implementation was not done as described; inconvenient corners were cut)
• lack of understanding of causality. When things don't happen as we would expect we usually lack the data to know whether this is because the intervention is actually ineffective, whether is it effective but poorly applied, or whether it was effective in another setting, but not in this one.
• lack of engagement :  a vicious circle
• poor precedent : spill over effect of negativity towards other interventions

Also, people remember when you get risk mamagement wrong. 

All these things have significant implications for the way we 'do' antibiotic stewardship. We need to be rigorous, we need to get the evidence, we need to understand why things happen and we need to help people feel that they are part of the solution and have a stake in getting it right.

Implementation is a process. Attention to implementation is triggered by problems - timing is everything perhaps. We need to look for the moments of cognitive dissonance and strike when the iron's hot. Banging away at a psychological brick wall is depressing for everyone.

Finally, we need to find ways for senior management and service led teams to share a common language and purpose. Every board meeting I've been to will pay lip service to quality, but in the end only end up talking about money. Most board members do not understand value as they do not spend enough time building understanding in the workplace. We must move away from assurance by spreadsheet towards a deep understanding of the nature of demand on our services.

Detection and management of sepsis - Fabiana Lorencatte and Neil Roberts. There is too much ISLAGIATT (It seemed like a good idea at the time) in service redesign. There is no learning from success or failure. Meaningful change requires understanding of the determinants of human behaviour. This is a complex field and can seem bewildering to the non-specialist. But there are common themes underlying most behaviour change theories. They talked about TDF. Personally, I like the Switch framework. There was a lot of discussion about how we do this sort of work at pace. Quick and dirty...and wrong? This then feeds into the difficulty in knowing (at all levels of an organisation) how to to make sense of competing imperatives. How do we decide whether antiiotic stewardship is most important? Or Sepsis 6? Are these the same thing...they don't feel it. 

Carolyn Tarrant - line infections in ICU. What predicted good uptake of Matching Michigan?. 

• Passionate clinical lead who sets expecations.
• Peer education and persuasion
• Role modelling
• Active use of data
• Embedding and normalising
• Any opportunity to remind people "This is what we do"

We must avoid the temptation to focus on technical aspects of intervention. 

"Improvement and implementation is inherently social."

Discussion.

We need to move away from seeing Sepsis 6 and antibiotic stewardship as competing goals. Stewardship is too negative. Why as, as informed citizens, would we not have amoxicillin for a cold? Or meropenem for cellulitis? We need to find the salient motivators for patients and clinicians, and not see stewardship as a worthy compromise that is good for the public but bad for the patient. 

There was a lot of agreement that our common and unifying goal is optimisation of infection management.

Should we be spending so much time as clinicians worrying about intervening in individual cases? Management may often not be as we would do it, but are the marginal gains always worth the pain? Again, we come back to dissonance- is insisting on a 48 hour IV to oral switch actually always a good use of time. Perhaps better to spend the time discussing 'grey cases' with prescribers. Exploring real life problems and talking about pros and cons of different management approaches. Being seen to be working with people towards a common purpose, rather than against them in a somewhat abstracted form, must surely be a better way forward?

Mike Cooper talked about 'higher order goals'. These are things that are more likely to motivate people to change,and if we can measure them in ways that have credibility, these are powerful ways to create dissonance and promote change.

Another way to talk to prescribers is to ask them "What made you uncomfortable in the last 24 hours?" This is the basis of the 'heads up' campaign, I can't find a link for this, but it sounds a great way of engaging with what really matters to people. And this is really the essence of all behaviour change. Finding that feeling that shows people you get what matters to them, and will be there for them. They are not interested in your problems. So there it is. Easy really. 

Points of leverage in infection management optimisation?

Some work we have been doing in pathology recently shows how profoundly misguided we are to be focussing so heavily on clinical outputs over which we have no real control. These include things like cost. Or mortality rates. These aggregated measures tell us nothing about what happens to individuals, And it is these individual encounters that build up into the aggregate picture.  The other type of measure we often use to describe our services are process measures, such as turn around times. Most of us know these are pretty meaningless to patient care, but because we haven't been able to think of any good alternatives we have increasingly badged the quality of our services according to these variables,

So our only mechanism to improve (our points of leverage) are at the level of input into the care of the individual. In diagnostics, this has led us to the concept of clean in, clean through and clean out. By following this methodology we have seen dramatic improvements in care, mainly by reducing the potential for iatrogenic harm by overtesting. So we've reduced MSU submission rates by 50%, and have now done the same for liver function tests. We have started to develop new ways of thinking about laboratory error that have clinical meaning. And we are beginning to use new ways of displaying results that help patients and clincians understand what they actually mean and promote the correct actions.

So is there something here for antibiotics? If we are not going to look at resistance rates or mortality (aggregated outcome) or time to antibiotics, or guidelines adherence (process and compliance measures), what can we measure? This needs some work, but it might look something like :

Clean in to prescription

• Antibiotics are necessary (or justifiable - this will be a continuum)
• Antibiotics will cover likely (including previously and subsequently isolated) pathogens (sufficient)
• Antibiotics are maximally appropriate (without unnecessary side effects, including overly broad spectrum)
• Appropriate diagnostic tests have been taken before antibiotics (eg Urine culture- UTI; Wound swab - cellulitis with open wound; blood cultures - severe sepsis)

Clean continuation

• Antibiotics are given in a clinically appropriate time frame
• Antibiotics are given by a clinically appropriate route
• Antibiotics are reviewed as soon as results and clinical progress dictates

Clean discontinuation

• Antibiotics are stopped when clinical improvement unless clear need to continue
• Source control is achieved were necessary
• Patient understands the diagnosis and what they need to do to prevent recurrence if appropriate

I think the big difference here with things like Start Smart is the shift away from compliance based approaches to a more forgiving and learning collaborative approach. I think to do this will require infection specialists to act in a coaching role and encourage clinicians to explore their practice in a much more open manner. This will be much more fun. I'm going to start tomorrow with my very nice respiratory consultants, I'll let you know how we get on.