Al Mulley: New care models to capture the critical intelligence needed for sustainability

This content relates to the following topics:

Article information

  • Posted:Tuesday 03 October 2017

Dr Al Mulley, Managing Director for Global Health Care Delivery Science, The Dartmouth Institute (United States) shares learning from Dartmouth’s experience working with six of the multi-specialty community providers and primary and acute care systems in England.

This presentation was recorded at our conference Learning from new care models here and abroad on 3 October 2017. 


Thank you very much.  I mean it’s a real pleasure to be here. I’ve been asked to address four issues; one is involving patients in the public, another is strategic intent of accountable care and overcoming related challenges.  Another is how one creates a sense of mutual accountability within a healthcare system when people bring very different perspectives to their roles, and another is exploring leadership opportunities for the NHS in becoming the key international learner and sharer of learnings in accountable care.  I’m going to trim my time as well so this will be very brief.  Some of you, probably two thirds of you have seen some of the slides, so I apologise for the other third if I go over them somewhat quickly.   

The first issue, the involving patients in the public; I’m going to argue that there are many ways to involve patients in the public and the development of accountable care systems, but perhaps most important is to recognise their engagement one person at a time to capture the critical intelligence needed for sustainability in the health system. I’m going to spend a little bit of time going into that and then move out to the other three topics.  They’re all quite related.   

Before I do that I’m going to spend a little bit of time giving you a sense of why I’m here and the role that Dartmouth has played in informing how one learns from practice variation and how one learns from variation in outcomes and costs.  Dartmouth has actually been engaged with the NHS for decades, in a partnership learning from variation.  Over in the lower left-hand corner here, you can see a paper published in 1982 first Author Colin McPherson, second offer Jack Wynberg, I joined the partnership very soon after that within months.  In 1996 the Dartmouth Atlas was published, the stunning finding was that we were spending three times as much per capita in places like Miami and as opposed to places like Minneapolis and it took seven years for my colleague Elliot Fisher to determine conclusively that the outcomes were no better and function and mortality and in fact perhaps a little both, and both doctors and patients saw the care as having less quality and less value because it was more fragmented.   

In 2010, Muir Grey and Phil DaSilva published the first NHS Atlas of healthcare, modelled on the Dartmouth Atlas and low and behold, you had threefold per capita differences, in cancer care and cardiac care, in 152 PCTs that were studied at the time.  Right care tools emerged from that.  Dartmouth was also involved in the early days of predicted modelling with the combined predictive model developed here at The King’s Fund.  I had the privilege of being invited by Chris in 2011 to be the first international visiting fellow at The King’s Fund, and what he commissioned me to do, and I invited some other co-authors, was to join up the thinking about practice variation, shared decision making, and commissioning.   

We wrote a paper at that time, and chose not to use the term “shared decision making” at all.  Instead we talked about preference diagnosis and we began with an example of two patients, both with breast cancer, one of whom who had a disease misdiagnosis, the other of whom had a preference misdiagnosis and our point was that the consequences for these two women were equally severe.  The scalpels were really sharp, the recoveries every fraught, engaging patients understanding what matters to them, understanding what they want when trade-offs have to be made is a critical part of sustainability in healthcare.   

I’m going to let you hear from two women very briefly.  These are women that I interviewed in the early days of developing shared decision-making programme, the woman on your right is Tina, the woman on your left is Ruth; I’ve learned that 15 seconds from a patient is worth at least 15 minutes from me.  Listen to Tina: 

“I felt the best thing to do and the most conservative smartest thing was just get rid of it, it’s bad, just don’t take chances.  I knew me, I could not live every day wondering, oh is there any left”. 

Listen to Ruth: 

“It’s possible to do the conservative surgery, to take out the lump and the surrounding tissues and still be left with the breast. It’s very important for a lot of women” 

Soon after developing the early decision-making programmes, we coined the term “decision quality”.  Are we sure that people understand the trade-offs that have to be made and are we sure that the treatment they get is consistent with their preference.  A little bit of research related to this is very revealing.  We interviewed scores of women and identified twelve concerns that they faced when making a treatment choice for early stage breast cancer.  Of those twelve concerns, we then asked their doctors: is living without a breast one of the top three?  The response from doctors, 71%.  The response from the women, 7%.  I could give you similar examples about adjuvant chemotherapy, about chemotherapy for recurrence and my co-author Chris Trimble, from the business school said, Al, in other sectors of the economy, there are severe consequences for knowing so little about what the people you serve want.  It’s called bankruptcy.  In healthcare it’s called unsustainability.   

We learned from variation outcomes and costs by making visible underlying variation processes.  We learn from variation practices by making visible underlying variation in preferences and making sure that it’s the patients’ preferences that drive the decisions, not those of the local conventional wisdom that are often shaped by capacity and number of specialists or treatment facilities for a particular intervention.   

Just a couple of examples to bring this home, I know that joint replacement is a priority in the NHS right now, from the Carter Review in Girth you can see the deep wound infection rates vary eightfold among the key trusts doing joint replacements, studied in the Carter Review.  Each is traumatic for the patient and incurs additional costs of £50-100,000.  Quantity of hip systems among trust vary greater than fifteen-fold with one to seven brands and partly because of all of this brand preference among individual operator, average price there is twofold from £788-1590.  You need to make visible these kinds of underlying processes in order to improve efficiency and productivity.  But you also need to pay attention to whether or not it’s the right thing to do for the individual.  Here from the NHS Atlas and Right Care, you can see that there are fourfold variation in hip replacements across the 211 original CCGs here.   

In the United States, we too have fourfold variation in per capita rates of joint replacement for arthritis.  On the right you can see that there’s twofold variation between Salt Lake City and Manhattan, the cities on the right are represented by the red dots over on the left.  They’re roughly twenty of 306 hospital referral regions.  You can think of them as CCGs.  

In 2001, Gillian Hawker did a fascinating study in which she wondered whether or not the fact that the per capita rate of hip replacement on the western side of Ontario was twice what it was on the eastern side could be explained by patient preferences.  What they did was they assembled a group of patients who were perfect candidates for either a hip or a knee replacement by physical exam, by history, by imaging studies.  They then did a little bit of shared decision making among those perfect patients, in terms of the clinical criteria.  Here’s what the rate looked like, without shared decision making, here’s what it looked like with shared decision making.  Look at the discretion.  

Now the headline was that 85% of people in the high rate area said no, 92% of people in the low rate area said no, so it was the preferences, but look at the discretion.  How can this happen?  Listen to this patient: 

“Both my knees are arthritic but this one is now at a pain level where I need to do anything, you know it’s not hurting, you can examine it and x-ray and see it’s arthritic but it’s not bothering, it’s not hurting me so that’s okay, but the left knee was at a point where I just couldn’t take it anymore.  I had to do something” 

So clinically and by image, both knees.  Only one bothered him.  He was at real risk for wrong side surgery, which is a well-known patient safety issue.  What about wrong patient surgery?  Listen to this woman: 

“It’s a pleasant surprise that I can keep it under control without having to get into surgery because it’s seem to me talking to friends and other people, that it was alright, you got some pain you have to have a knee replacement.  That’s not necessarily so.” 

So I’m going to do this very quickly, beyond markets was one of Chris’ phrases in Reform from Within and I’m going to give you a sense of the historical perspective on the market and government failure in healthcare.  Patients and families are critical, it’s often thought that patients and families don’t have a lot of leverage in dealing with health professionals, so there’s voice to policy makers and they enforce a compact and it’s been argued that you can just increase patient power by creating transparency about outcomes over cost and therefore create some competition and contestability.  I don’t think many clinicians buy that, and most of the evidence says that patients don’t pay an awful lot of attention to these kinds of comparisons when they’re feeling vulnerable.  They need support.  What they need is careful attention to help them understand what the evidence says is possible for somebody in their predicament and careful listening to learn what matters most to them.   

Simple measure that we developed at Dartmouth that is a bit of an iconic simple tool that I’m going to talk about later, so I’ll take the time to tell you is three simple questions: zero to nine scale, how much effort was made to help you understand your health issues, how much effort was made to listen carefully to learn what matters most to you, how much effort was made to consider what matters most to you in planning the next step in your care?  If you know that that’s happening, if you have feedback mechanisms for the GPs and others, to improve their ability to do that regularly consistently, not necessarily themselves, but with the team that they’re responsible for leading, then you can be relatively confident about high quality decisions as I defined them earlier.  And when you have choices, made with high quality decisions, you have the critical feedback, you have commissioning intelligence.  If you let the easily avoidable ignorance rest at the frontlines, it permeates a system.  How can the commission be informed?  How can the capacity investments be informed?  How can the workforce planning be informed? 

This seems so simple, at least to me, I’m talking over a long time, but it seems so simple to most audiences. What keeps us from doing it?  There are conceptual challenges that I won’t belabour here because I talked about them in the morning and afternoon session.  They’re all about measuring what matters in managing for accountability.  The managing for accountability as well as the measuring what matters can be characterised as reductionism.  It’s reductionist management around productivity.  It’s Taylorism or Fordism to use that term from those of you who know business school jargon.  It’s assembly line thinking.  But I’m going to focus on here for a moment, just to make the point of what’s potentially possible and what we learned from the vanguards that we worked with is delivering with teams and organising for innovation.  Those are the operational challenges.   

These teams have to be really different than the MDTs that are generally talked about.  We have to include new roles, design for engagement as the front-end learning system for patients and families.  And they’re so different that you can’t incrementally get there.  Somebody has to make a decision, that we’re going to experiment here and the person leading this innovation effort can’t be held responsible for meeting targets using today’s measures.  They need to be held for learning rapidly, using new measures, like collaborate or integrate another measure that I’ll show you in a moment.   

Just very quickly here, if you look at this diagram, level of training in the medical sector it could be the neurosurgeon or the neurovascular surgeon off to the right, and the health coach or navigator down to the lower end of the spectrum.  Difficulty of the task goes up, with the vertical line.  That means the 45 degree angle is very important, everything below it is potentially inefficient, everything above it is potentially unsafe.  That feels like Taylorism because there’s only one finite spot where you’re both effective and efficient.  But, if you really emphasise team work, even measure team work, using something like relational coordination developed by Jody Cortell, that makes things feel different.  Certainly at the frontlines, and even at the organisational and clinical leadership level.  

Dartmouth leads the research in accountable care organisations in the United State, there are now over 800 accountable care organisations, regular surveys are done and early on, twenty interviews of leaders of eleven accountable care organisations, tried to get them what they thought was really important about care management and coordination of care.  At the leadership level clinical leaders and executive leaders made it clear in their interviews that they thought shared goals were important and that frequent timely problem-solving communication was important.  They didn’t get share knowledge or each other’s roles and task inter-dependence.  They didn’t get mutual respect and they didn’t get accuracy of communication.  

When asked about what should happen at the frontlines, they only got frequent and problem-solving communication.  And what happens at the frontline is so much more critical and complex than that.  

In order to test some of this at Dartmouth, when I went there seven years ago, after thirty-five years at Massachusetts General Hospital, I insisted that we build a new model to really experiment.  We had it up and running in six months, we reduced commission time by 50%, nurse time by 50%, we basically eliminated receptionist type administrative support and we used those resources to recruit and train ten health coaches, half of whom ended up having no prior experience in healthcare.  They were recruited for common lived experience with the patients they would coach and passive communication skills, they were trained in shared decision making, motivational interviewing, to understand the needs and wants and challenges that patients face and to avoid the substitution of high acuity care.  When it fails to meet needs and exceeds wants.   

I skipped over reading the line on the first slide about testing the sustainability hypothesis.  The sustainability hypothesis is that a healthcare economy can be sustainable if, and only if, you recognise the opportunity to avoid substituting something that is high risk, high cost and high acuity, for something that is much more basic and could be met with very low-cost interventions of social nature or in healthcare.   

We had the opportunity, in many ways, thanks to the opportunity I had as a visiting fellow here, to really get to understand the Five Year Forward review when it was published, and we saw this triple integration of primary and acute mental, physical health, social as an opportunity to build a foundation for testing that sustainability hypothesis.  We were in touch with the new care models team relatively early on, at one point we submitted this, I don’t expect you to read this, but it will be available to you afterwards if you choose to, and we were given the opportunity to take this and use it to co-produce something with a team of six vanguards, four MCPs, you can see that we were working with people from southern Hampshire, from Birmingham, from West Lakefield, Barnsley, the two packs were Farnham in North East Hants and Salford.  We had guests from other vanguards, we had ongoing support from the new care models team, NHS England, a Dartmouth team of six senior faculty and someone with experience as a chief learning officer from a global service company.  We also had UK colleagues.  We did cycles at one month and six months, they were invaluable for learning local context and local needs.   

We also were asked by the new care models team to commit to being in touch with everyone else who was interacting with the vanguards and we did that, The King’s Fund, Nuffield Trust, the Health Foundation, a Paul Corrigan who was leading the MCP community of practice.  We also had ongoing sourcing of ideas and evidence, from Dartmouth and other international experiences.   

And this is what we ended up with.  We ended up with a narrative.  You’ve got to use logic for learning and learning from variation, delivering what’s valued, measuring what matters, delivering with teams, organising for innovation, leading for accountability and governing for stewardship.  But most importantly, what we ended up with was a set of measures and tools, like collaborated many others, that we tested the relevance of through the eyes of these frontline workers.  We had the opportunity to write about this, this was a paper published the same day as Next Steps in the Forward View and box 3 which I don’t expect you to be able to read, is a description of how these teams were using these tools when we did the return site visits after the eight workshops which produced these eight essential capabilities.  

Here’s a quick set of the key learnings, strategic intent and actions needed to overcome challenges with new measures and tools.  PBC NTs for measures and tools to use as I said, essential capability narrative elicit common patient stories, leading to the sustainability hypothesis.   

There are clear opportunities for improvement, engagement and knowledge will sponsorship for leaders and wasn’t what it could have been this first time around.  We know we can make that better in the future.  Further flipping of the classroom for more actionable learning, we know we can do.  Coaching and technical support for tactical sharing, is a term used by our participants.  Curating examples in evidence of mutual accountability.   

I skipped over an important learning and this was really exciting to us, we were trying to learn together with these people and leave all of the capabilities with them and the early participants were happy to and anxious to serve as local faculty in future versions of this.  We will be doing future versions, we’ve been commissioned by UCL Partners with funding from Health Education England to adapt the PBC and for NCL and NEL, STPs, and we’re working very closely with Right Care to adapt the PBC and learnings in support of STPs designated as ACS ready.   

I’m going to close very quickly now, I think that you really do have extraordinary advantages over certainly the US and many other countries, in the NHS to lead in the learning about accountable care.  Ongoing learning needs in the US and this comes from my colleagues who do this full time, the conditions and capabilities for ACS costs and quality performance.  We know the primary care is important, we know the clinical leadership is important, we know the priorities faced on urgent, low hanging through issues like A&E is important.  Organisational structure is not predicted by role of partners in the system may be.  Conditions and capabilities for new forms of partnering within and across organisational boundaries.  More than 80% of ACOs entered new partnership, motivated largely by need for complimentary capabilities and risk mitigation.   

Condition in capabilities for engaging patients and families in decision making and co-production, we’ve already talked about the emphasis on primary care with new roles.  That isn’t happen in the US as much as it needs to happen.  In fact, ACOs in the US have not been doing a good job of entering into this kind of engagement, integrating mental health, there needs to be more work done there that I think you could lead in.  

Some really priority learning opportunities where I think the NHS could lead because of the far less fragmentation than we have, effective financial incentives and/or intrinsic motivational performance, again echoing Chris’ reform from within.  Effective using new measures of collaborative capacity and tools for mutual accountability and effective new clinical teams roles and populations vulnerable because of complex health and social care needs.  

This is one of the slides that I used in both of the last two sessions, so I’m just going to close with it here.  From the frontlines to system leadership, we experimented with tools from the frontline interface between a coach and a patient, all the way up to executive directors of acute trusts.  Leaders of STPs, the ACSs and even those at the very top of government who sometimes have unrealistic expectations about how quickly meaningful reform can be achieved.  

Thank you.