Quality Improvement Traps
Nick Downham – Oct 2018
© Cressbrook Ltd 2018 – All Rights Reserved
Using the English NHS as an example, this article focuses on the weaknesses of much current Quality Improvement (QI) thinking and application within healthcare settings. It details six areas of weakness in many of the QI approaches I have observed.
There is huge investment by the English NHS and other healthcare settings in developing and bringing in QI capability. Almost every hospital and region has QI plans or transformation plans that feature QI as an important component.
The Health Foundation, in their useful mini guide ‘Quality Improvement Made Simple’ use Dr John Øvretveit’s definition of Quality Improvement:
Improvement as better patient experience and outcomes achieved through changing provider behaviour and organisation through using a systematic change method and strategies’
Despite massive investment, the impact of many QI programmes is not clear-cut. There are of course some improvements at the level of the process, but there is little evidence of it being as ‘transformational’ as desired. In fact in some cases it appears to be slowing down NHS transformation rather than complementing it.
There are a number of traps that prevent healthcare systems making the most of QI. These problems are not because QI is flawed as a methodology, but due to poor application and positioning. I say this from the perspective of someone whose career is based on this subject area. I trained as an industrial engineer at university in 1996 (think quality thinking, QI, reliability and flow), worked as an engineer applying it in industry and then spent 13 years working for the NHS, and independently, running a number of £million QI based programmes for different NHS settings. So I get its use, but I have learned its limitations, especially with regard to system change.
If not avoided, many of the traps I list below will lead to QI becoming known for failed improvements. Eventually the term will become toxic. This would be a tremendous waste given the investment the service is making.
It is not all bad, there are great examples of QI out there, and core QI thinking is of huge importance, but what are the traps?
The Quality Improvement (QI) Traps:
1. Inadvertent reinforcement of the current structures:
The lineage of QI thinking comes from bounded, specified processes, like in factories. Hence it is of most immediate interest to similar organisations like hospitals.
QI is often implemented through the training of front line teams (more on this below) or by training QI teams who then run QI projects with front line teams. As QI is predominantly process focused, when it is applied like this it accepts the existing functional and professional boundaries at play i.e. it makes the current better. This is all very well, but if our hopes are pinned on transformation it is problematic. With the huge investment in the NHS with things like Sustainability and Transformation Partnerships (STP) and other transformation initiatives, it seems that transformation is very much the aim. If that is the case, then making the current better is simply not enough. What if we need to do better things? [*] What if we need a fundamental change in model in many settings?
The application of QI rarely brings this level of challenge. There is often a level of acceptance that the process or structure is broadly correct, it just needs to be made better. This is single loop learning. What we need for transformation is to do better things i.e. question the work we are doing fundamentally, down to the core assumptions that shape it – this is called double loop learning. Otherwise we end up trying to make the wrong thing righter. This brings us in a loop back to Øvretveit’s definition. His emphasis on changing organisation is key – i.e. challenging the assumptions that form our work.
2. Lots of focus on wastes, far less (or even zero) on failure demand:
What if much of our demand is generated by us (healthcare organisations) rather than patient need? This is called failure demand and it surrounds us in epic proportions. As my blog article Sources of Failure Demand in Healthcare explores, much of this is down to the deliberately fragmented nature of how our services are designed. Again, this brings in the notion of fundamentally questioning the work – this time at the leadership and system level. This is something not often done with QI, as it is often introduced by leaders and applied down through organisations i.e. it is done to others (generally the front line).
The way the NHS, and other healthcare systems, are shaped is driven by ideologies. As the above article Failure Demand in Healthcare explores, there are five dominant ideologies at play (economic / industrial, political, bureaucratic, moral / ethical and professional). It is these that truly shape (namely fragment) our systems – even down to the level of the process. It is the interplay of these that are the root cause of much of the system dysfunction we see on a daily basis. This fragmentation of our system is at the centre of the difficulties in improving towards the six dimensions of quality (safe, effective, patient centred, timely, efficient and equitable) as detailed in the US Institute of Medicine’s ground breaking report Crossing the Quality Chasm.
Despite this, the application of QI in the NHS commonly lacks challenges in these ideologies. Most notably because the thinking behind these ideologies is so engrained in the NHS (like the NHS’ DNA) that we are often unaware off their presence.
Challenging underpinning ideology is not missing from QI as a methodology, but rather it is often missing from the healthcare interpretation of QI we often implement. After all, one of the forefathers of QI, W Edwards Deming has understanding Theory of Knowledge (in other words ideology) as one of the four lenses of his System of Profound Knowledge.
3. Assuming all healthcare settings operate like a hospital:
While they may have some merit in some of the transactional parts of hospital systems, standard QI methods are being applied and trained, at scale, in the primary care system. In many cases QI practitioners fail to realise the difference between the two systems.
Primary care systems are generally purposeful systems – i.e. they are reliant hugely on patient context – whereas the majority of the secondary care systems are, while hugely technical, transactional. QI cannot be applied in the same way in these very different system archetypes. No amount of QI process improvement is going to transform a health system reliant on non-healthcare determinants of health – systems which are network based. This trap is amplified at the moment as hospitals and hospital QI teams become more prominent (due to having the lion-share of the system budget) and train those in the community. The tendency is to apply hospital archetype method to primary care problems…..
4. We focus on the tools:
The quest for scale in QI application can result in very heavy standardisation of training and method – because it is easier to QA / train practitioners. This standardisation and training at scale often results in a loss of meaning. An example of this is one of my old programmes, the hugely influential NHS Productive Ward programme. While it achieved a huge amount in many places, I often saw problematic implementations of it. Often leaders had not understood what was beneath the surface and had just taken the tools. So I saw many hospitals end up boiling the Productive Ward programme down to the application of 5S in store cupboards….
We also train the tools as they are easier to visualise and quantify. Hence the obsession with workplace organisation (5S), process mapping, driver diagrams, basic demand and capacity and fishbone diagrams. As we can list the tools, we can judge individuals on how much they know, and possibly even create jobs around this gradient of tool based knowledge. But consider this – just because I have a bag of spanners and other tools doesn’t mean I can fix a car. Even if I can name those tools and roughly understand what they do still does not mean I can fix cars. The value is in the mechanic’s knowledge of how a car works, not the tools. So, when considering health and social care, the value in improvement is in understanding the process, system conditions and ideologies that shape the work [**], and then having the skill to change them; something many QI initiatives and practitioners are not positioned to do.
5.We don’t contextualise for clinicians:
There are countless examples of websites and publications full of QI tools and guides. Even from those from organisations within healthcare systems, much of the content has very little contextualisation for the audience. They rely on clinicians to contextualise the principles or the tools for their setting. My experience is that our ability to do that is not great. I have found that one of the barriers to implementing new systems is sometimes not down to the content and quality of the new thinking, but that many professionals find it difficult to contextualise a theoretical framework to inform their work. It is the concept of a theoretical framework that is the issue. This is driven by the lack of any headspace for clinicians to think these things through and it is also not something we do very often. Being able to contextualise a theoretical framework, and much of the core of QI thinking is in the form of frameworks, is a learnt skill. Without help busy professionals don’t have time to do it properly and we end up losing the real depth of application. We end up just skimming the surface.
At the level of tools, these are often over complicated in their presentation (it keeps management consultants in a job!). We often don’t start by thinking about what clinicians already know. For example with run charts, we know that many clinicians already can use observation charts, so they already have the core skill – we just need to link to it rather than present run charts as something new.
6. We never give it time:
On a recent study tour I attended to high performing Intermountain Healthcare in the USA, the headline message was how long it took to properly apply QI. In their case it has taken decades to put in place, and even they would say that they have only just begun.
Take the core QI theme of variation, as I discussed in my article on the subject the really challenging thing about healthcare is that we start from a place where there is little agreement about what ‘good’ is, especially beyond the scope of single interventions. Even NICE guidelines are often contested. So when trying to understand variation, the question of ‘variation around what?’ is incredibly difficult to answer without diluting things by resorting to rough proxies. Intermountain spent decades co-producing and testing the efficacy of their core pathways for their major condition groups. Only once these were in place could they begin to build data systems to understand the clinical variation around these pathways and the cultural foundations in order to constructively discuss any unwarranted variation identified.
I rarely see any NHS implementation of QI which is given the luxury of such time. Expectations are set and programmes must meet targets. This emphasis on quick results forces us to avoid confronting the systemic issues. As said previously, it forces us into single loop learning where we don’t challenge the underlying assumption (driven by ideologies) that shape our work.
This lack of time also rubs off how on we seek to influence others. We all know that you own what you create. That for lasting change we need to involve those doing the work and those leading the work in testing their own assumptions (this is vital for unpicking underpinning ideologies). Yet, without time we are forced into relying on mass training or even incentivising (for example QOF [***] payments and CQUIN [****] payments – both of which are essentially a form of coercion). Using hard or soft forms of coercion is a long way from the co-production required for lasting change of mind-set.
The so what?
So much of this is OK IF the aim of the huge investment in QI is to maintain the status quo (to tread water). If the intention is that QI contributes to a real shift in the system, then our efforts in many cases are missing the point.
There is much to learn from core QI theory, but my learning from specialising in this field for the past 18 years is that we need a much more mixed model. It is not about one tool or approach. The problem is that the NHS is incredibly faddy when it comes to new initiatives. Hence the current system obsession with QI as the ‘answer’.
We need a model with good process rigor, but one that systematically challenges the core ideologies that shape our institutions. Deming had it right decades ago by encouraging us to start by viewing the current work through his incredibly well balanced System of Profound Knowledge.
Unfortunately, in many cases we choose to pick the bits of his teaching that are easy to replicate and things that give us more immediate results (the QI tools and single loop change). A choice that is unlikely to get us to the transformation we seek.
If you liked this article, you may also find the following articles useful:
[*] Anderson-Wallace, Blantern & Boydell (2000), Modes of Organising.
[**] The work of John Seddon and his Thinking, System, Performance framework is particularly useful in unpicking this.
[***] QOF – Quality and Outcomes Framework – a General Practice incentives scheme within the English NHS.
[****] CQUIN – Commissioning for Quality and Innovation – a commissioning incentives payment scheme used within the English NHS.