Your Nuffield Trust Summit speech this year highlights the policy successes and failures that you've seen. Which do you feel have gone well?
It’s clear from my recent conversations on this that many of the things that went well happened quite some time ago. Quite a few of them were actually in public health, with the ban on smoking in public places being one obvious example. Sure Start, the reduction of waiting times, London Stroke, reducing health care-acquired infections and creating national service frameworks were also notable successes.
What links them is that they were often relatively simple, focused and evidence-based policies dealing with particular issues – with policy-making often local, driven by professionals and designed with wide-scale involvement of the appropriate stakeholders.
Which have gone badly?
There are two categories here: things that we should have done but we didn't (such as workforce planning or social care reform), and things we did do that either should not have been done or been done better.
Examples to mention include the National Programme for IT (NPfIT), the Private Finance Initiative, the Better Care Fund, and the 2012 Health Care Act – all of which are highly complex and had fundamental flaws built into both their design and their execution. Another spectacular failure was Transforming Community Services, which was driven by a managerialist approach not very grounded in the evidence.
When we look at the previous answer of what did work, on the whole they were not driven by an ideological theory. If they were theory driven, it had an evidential base behind it rather than a political theory, which was the case with Lansley's reforms.
What else might have been behind these failures?
Some policies, like the Better Care Fund, have had failure built into them. We’ve seen overall problems with the way policies were designed, we’ve seen people with solutions who are looking for problems, and a faulty logic about causality has also often been a culprit – mistakenly thinking doing X will lead to Y.
But perhaps the most common problem, particularly for national policy that's been rolled out, is a tendency to forget that different areas have different starting points, and that context matters. Your policy may work very well in the average place, but most places are not average.
Once the design of the policy is in place, there are still opportunities for good policies to go wrong when they are being implemented. One of the critiques that one may make of the Five Year Forward View, for example, is an absence of a really good narrative about how change was supposed to come about.
It suffers, like many other policies, of optimism bias about the speed with which things happen. For example, our Primary Care Networks pre-mortem identified that not enough time had been allowed for just the logistics of getting practices together to talk about how they were going to implement it.
A particular implementation hazard for health care is that you can't really change culture – which is very often what you need to change – simply by pulling policy levers.
Why do you think we keep having these problems? Is there a particular British problem?
The other problem we have is the highly centralised nature of the NHS and the extent to which politicians get involved in the detail and design of some policies. Alan Milburn got involved in the design of regional structures, for example, and Patricia Hewitt in detailed decisions about licensing and authorisation of drugs, while the issues with Andrew Lansley’s Act have been well documented.
The central nature of policy-making is a problem because England is a very big unit of just over 55 million people. No one else is trying to run a health care system on that scale and design policies from the centre in quite the way we do.
Centralisation is a problem with scale, because with scale comes that complexity. The policies that have failed the most have been in areas with high levels of complexity – bringing an uncertainty about the effect that a particular action will have. That isn’t just uncertainty about whether implementing a policy might produce an expected result, it can also mean a lack of consensus on whether or not what you're doing is actually a good idea.
If we look at the worst failures: the 2012 Act: very complex and very top-down; NPfIT: even more top-down, just as complex. So a combination of scale and complexity at their worst.
Can we learn by looking overseas?
If you look at the Netherlands or Catalonia and some of the other systems that seem to do well, even though governments change there is a consensus on what the model and method is for change and the wider direction of travel. You don’t get the wild swings and reverses that we’ve seen over the years in the UK, where not only governments change but the Secretary of State changes, which can then mean the underlying theory on what we're trying to do also changes. The presence of NHS England has at least provided some insulation from that.
But all countries have serious difficulty in controlling growth in activity (and therefore cost). Germany has had fewer problems on that front in recent years but they've had a huge amount of slack in their system, meaning they've been able to absorb it slightly better.
All countries too have to deal with the fact that doctors are an 'unusual' workforce – in that your frontline staff are better educated and more powerful than the people managing them, and often have a different world view from those people too. That can have an impact when it comes to implementing policy. Don Berwick has said that many health care reforms are an attempt to change the behaviour of doctors but without ever having a direct conversation with them about their behaviour.
What would help to improve things?
Better use of evidence in the design of policy and a better theory of change underpinning its implementation. There's also a tendency to go too fast – you might get better results by going slower. In the end that might be quicker, paradoxically.
We must not underestimate the importance of organisational stability. Having patience with what you've done and not changing it halfway through would hugely help – failing to do that has been a real feature of the some of the structural reorganisations that we've seen. They've been poorly designed and then altered too quickly – that’s a double whammy. People have been too quick to declare defeat.
Fundamentally, a much better understanding of context and history is vital – to learn from what's gone on before. That's one of the reasons why we, with the Health Foundation, will be investing quite a lot of time and effort to try and restore the NHS's damaged organisational memory. Having a better grasp of history would certainly go a long way to avoid the repeated policy failures that we’ve seen.
Edwards N (2020) "Successes and shortcomings: health policy lessons”, Nuffield Trust comment.