Sarah Thew: Promoting connected health care for Greater Manchester

This content relates to the following topics:

Article information

  • Posted:Tuesday 11 July 2017

Dr Sarah Thew, Innovation and User Experience Manager, Greater Manchester Academic Health Science Network, discusses how the DataWell Exchange Platform is promoting connected health care for the devolved Greater Manchester health and social care system.

This presentation was recorded at our Digital Health and Care Congress 2017 on 11 July 2017.


Good afternoon everyone, I’m Sarah Thew, I’m the Innovation and User Experience Manager at Greater Manchester Academic Health Science Network.  I’m going to be talking today about DataWell which is a health information exchange platform that we’re in the process of deploying for Greater Manchester.

I’m a geneticist originally.  In about the mid ‘90s I decided that possibly computers were going to be more reliable than working with yeast cells and yeast bugs which kept dying in the incubator so I moved into computing and since then I’ve always worked in the space of sitting between the developers and the software technologists in getting out and talking to the stakeholders and that’s what my role is on DataWell.  So it’s about 80% of my job and I spend a lot of time out talking to stakeholders who care about DataWell, some of that is talking to patients and carers and working with the NHSM’s patient group, I spend time talking to clinicians but our other stakeholders are people like information governance people and the leadership around devolution.  I’m going to be talking at first a little bit about why we need DataWell from some of those different perspectives and then talk you through a little bit of techy stuff about the architecture of the platform but the slides are really not very technical and then a bit about where we want to go in the future.  

Academic Health Science Networks are organisations who are tasked with introducing innovation into the NHS.  Greater Manchester Academic Health Science Network in terms of people is quite small, there’s about fifteen of us.  Lots of what we do is networking, finding out about good things going on outside of Greater Manchester and looking at bringing them in, we also look at what’s going on locally and try to share good practice but also written into the business plan for the Greater Manchester Academic Health Science Network was the delivery of health information exchange platform for Greater Manchester.  So that’s a much bigger piece of work.  I think probably the creation of the Academic Health Science Network and our status as being within the NHS but not a provider and not a commissioner is quite helpful in doing this piece of work because we’re not seen as directly having skin in the game.

So this is my referral spaghetti slide.  On one side the grey segments are all RCCGs in Greater Manchester and the brightly coloured segments on the other side are our providers and this is about three months’ worth of referrals data and firstly it’s quite complicated, secondly what I’d like you to take from this is the fact that our patients don’t stick within the localities.

We’ve got ten localities within Greater Manchester but people go all over the place and that’s not surprising.  We have hospitals that specialise in different sorts of care, but unfortunately their records don’t follow them and I’m sure this is a story that you all recognise from your own localities.

Within Greater Manchester we’ve got some pockets of really good data sharing, for example the Salford integrated record, Wigan, Central Manchester, I’m not going to try and list them all or I’ll get in trouble for missing somebody out, but we’ve got these long term pockets of really good data sharing but even where that’s happening patients are travelling in and out of those localities.  So just to take Salford Royal for example, and there’s been the Salford integrated record in place now for about twelve years which allows Salford GPs and Salford Royal to see aspects of each other’s records.  That is fantastic and has enabled really good care and lots of great research as well, but about half of Salford Royal patients are not with a Salford GP.  So those Salford Royal clinicians have a big gap for the other patients who are coming from out of area. 

That’s a little bit about the system.  I’m not going to go through this in lots of detail because I’m sure you all recognise this, people are living longer, they’ve multiple conditions, people are being seen across specialities and across lots of different hospitals.  Within Greater Manchester as part of the devolution agenda we’ve set ourselves the task of improving the health of a population who, to be honest, are not one of the healthiest in the country.  We set ourselves challenges around rates of recovery for cancer, mental health, early years and in order to address all those improvements that we need to make and yet still acknowledge that we have challenges around budget we need to change the way that we’re working and fundamental to that is recognising that across Greater Manchester as a system we need to start to work together, we need share our expertise to try to only build things once where we don’t need to build things over and over again and to share data support those new ways of working.  

I do quite a lot of work talking to patients and carers and this is a really beautiful quote I think from one of the carers that I was talking to.  He was describing the excellent care that he gets from individual organisations but he feels that because we fail to coordinate across the system we’re not looking after him and his partner.  So he’ll find that he has two appointments at the same hospital on two days consecutively or he’ll find that on the same day he has to travel from south Manchester to north Manchester to get to two different hospitals.  He’ll work with one care provider to agree that they make sure they don’t schedule appointments against his respite time and that information isn’t shared.  So although the individual organisations are doing well we as a system are not doing so well. 

I’ve worked with our patient group to put together a video and our patients are talking about their experiences of care.  I’m not going to show it now because it’s about five minutes long and I don’t think I've got time, but I’ve pinned it as a Tweet at the top of my feed so if you’d like to watch it please do they’re a really good engaging bunch and it’s a really interesting video.

A little bit from the clinician’s perspective and again I think lots of this will be familiar to people in the room.  We’re not doing well at sharing information when it comes to transfers of care, we are not doing well when it comes to people arriving in hospital and having to tell their story multiple times.  I think from the point of view of someone who does a lot of thinking about system design and user research I’m well aware that the problem of having no information is probably not that much better than the problem of having absolutely all the information, and I think there are going to be some really interesting design challenges for those who are interested in user interface design in thinking about how we handle the move from having no information to having lots and lots of information and presenting that in a timely and effective manner to clinicians who are trying to achieve particular tasks.  

Fundamentally what I hear when I go out and talk to clinicians is the amount of time that they spend at the moment chasing results, trying to track down information about patients, trying to understand what kind of social care package someone is under and again there’s big opportunities there for us to enable the system to work more effectively if we start sharing data.  
I want to talk a little bit about research as well because DataWell is absolutely about thinking about audit and service design and research as well as thinking about direct care.  This is an example of a piece of work where we needed to understand what was happening to a population across their journey of getting a cancer diagnosis and their treatment as they travelled across lots of organisations and although there was massive willingness to do this piece of work across the system it’s really difficult to do because it’s really difficult to get the information governance in place and then it’s really difficult just to do the data linkage and to track the patients and it’s difficult because we’re just not doing it on a routine basis, but for Greater Manchester to work effectively as a system we need to understand our population and what happens to people as they move between organisations as well as what’s happening in individual organisations. 

So all this brings us round to DataWell.  DataWell has been a long project from initial specification to going out through procurement and getting to the point we’re at today, long enough for me to go off on mat leave and now have a child who’s about to start school in a few months’ time and that’s because although the technical work is hard, we worked across the system trying to gather requirements, it’s also because this is about hearts and minds and system change and people becoming more comfortable with the idea of data sharing and more willing to understand the benefits of data sharing. 
So we worked across Greater Manchester when the NHSM was first formed to pull together a set of requirements around what the exchange should do and we used that as the basis of a procurement and went out to public procurement in 2015.  My first and hopefully last ever experience of public sector procurement.  But after about nine or ten months we ended up with a contract we’re working with a company called LumiraDx along with IBM and EY and they’re delivering the technology for this system working with the Academic Health Science Network.  

This is my technical slide so it’s not too scary looking, is it?  I want to talk about some of the fundamental design principles that sit underneath DataWell.  So firstly there is no central data warehouse.  DataWell is an information exchange and information moves at the point of need.  So should someone turn up in accident and emergency and we need to find out the background history on that person it’s at that point that a query goes out across the exchange to go and try to find all the information about that person from the individual organisations who are part of DataWell.  

So these nodes down at the bottom of the slide are the individual organisations participating in DataWell, so the hospitals, the GPs, social care, mental health.  They each have a server of their own and they choose what information they want to put on to that server and they remain the data control of their own data throughout its use in DataWell.

The big orange bar in the centre is the DataWell exchange and that’s where we do lots of the work of building things like a master patient index so that we can track people around Greater Manchester.  The second principle of DataWell is that we don’t ask organisations to change the way that they’re producing information, for example we don’t ask everybody to change their naming convention so that we’re using the same naming conventions across Greater Manchester because we would get nowhere with that approach.  Instead what we do is take the data that organisations want to share into DataWell, and we’ll talk a little bit more about this in a moment, but the first data sets we’re looking to share around pathology, and we worked to map the pathology data against ontologies and data standards, international open data standards so that we know for example that at Salford Royal they hold white blood cell count as white blood cell count and the second hospital holds it as WBC.  We map that data into the exchange knowing that both of those things are the same thing so that when a query comes in needing a white blood cell count we go out and find that data from the two different organisations.  

The other side of this is that we also don’t expect organisations to connect to consume the information in a particular way.  So organisations working to take data out of DataWell and to display it in their own records can choose how they want to do that.  So where they already have an EPR in place, for example Salford Royal have their allscripts EPR, they can simply query against the DataWell exchange and pull out the information they need and display it locally in the system their clinician is used to using.  

We do also have a web user interface that’s a longitudinal record so that for organisations who don’t have an EPR or don’t want to embed the data directly into their own EPR they can still participate in DataWell.

That’s it for the technical stuff.  

So this is a little bit about where we’re up to now.  We now have our four major teaching hospitals in Greater Manchester connected up and starting to share pathology data which we’re really excited about.  We’ve got two CCGs involved in the exchange at the moment.  It should be clear that for the moment the GPs there and the CCGs are consuming data from the exchange rather than sharing their own data in.  That’s starting to change and we’re starting to get GPs and particularly GP administrative staff seem to be very keen to share results into DataWell.  I think they spend a lot of time faxing data around so I’ve been getting chased by practice managers who are really keen on DataWell.  

We wanted to develop a couple of example uses for DataWell and the first project that we’ve started work on is a shared pathology view.  So this is in support of direct care it’s recognising that clinicians and their staff spend a lot of time chasing results.  So I’ll talk you through a particular use case.  We have a neurology team at Salford Royal who support patients who have multiple sclerosis, they have about 30 patients who take a drug called Lemtrada.  Lemtrade has an extremely positive effect on your quality of life but comes with quite a high risk of side effects.  So the patients taking Lemtrada have very regular blood tests to track their platelet counts and out of that cohort of 30 patients they’ve spotted two who have needed to come in for hospitalisation.  So it’s reasonably high incidence of side effects.  

At the moment that tracking is being done by two experienced nurses and it needs to be experienced nurses so they can spot when something is starting to go wrong, but those experienced nurses are tracking with phone calls, faxes, logging in to independent systems, it is really not a good use of their time and Salford Royal would like to expand that service out and offer that drug to more patients but they don’t feel able to because they just don’t feel like they could staff that safely at the moment.  So I think end of July we’ve got training in place to start training those clinicians in the neurology department to get them using DataWell and starting to access these results across that system and we’re going to be working with the team there to follow the impact of starting to be able to access those results and understanding how we hope it’s going to start saving them time and changing the way that the nurses work.

We’re also working to support our local 100,000 Genomes efforts within Greater Manchester though I will just talk very briefly about this.  Fundamentally 100,000 Genomes is both a research project and a project to transform the way that the NHS works with precision medicine.  So whilst the mechanisms that we’re building within DataWell are in the first instance support collection of data for patients who are consenting into the study, the other side of this is that we’re building the mechanisms that will support referrals and transfer of results back from our local hospitals into our genetic medicine centre.  

The last part of this puzzle is our accelerator pipeline and that is the support that we provide to NHS and social care organisations as they want to start to make use of DataWell and we hope that DataWell in the end will just become background infrastructure rather than a kind of important consideration starting to share data.  We’re beginning to get some really exciting projects through that accelerator pipeline looking at social care, GPs sharing data, some work around stroke audit and looking at rehab.  So we’ve had some highlights over the past year. 

I particularly wanted to highlight one small piece of work that is really in test at the moment but I think is fundamental to the future of this work.  We’ve started looking at how we can link out to LPRES which is the Lancashire equivalent of DataWell.

So I said our patients don’t stay within one place within Greater Manchester they also don’t stay within one place within the north west.  So we have similar sorts of shared records in Cheshire, in Leeds and West Yorkshire and in Lancashire and we need to start thinking about what happens to patients as they move out of Greater Manchester.  So we’ve just started doing a little bit of sharing of test data out into Lancashire to support our patients who are based in Wigan and who often go north for their care rather than south.  So really although it’s early days I’m pretty excited about that piece of work.  

I wanted to talk briefly about consent and ownership and patient access to records.  I feel like it’s a bit of a cliché having an onion diagram with a patient at the middle of it, but fundamentally this is how we would see access and ownership of data within DataWell moving into the future.  We think that patients should have access to their records within DataWell, that you should just go to one palace to be able to access your data and very agonistic about what that app looks like and I suspect we’ll have multiple different apps.  I’ve been working with a company who make apps for teenage diabetics and it’s probably not the app I’d choose to use for me and my family but for that audience it’s perfect and all those apps should be able to start sharing their records with us assuming there’s appropriate IG and data quality work in place.  

So the patients need to sit at the centre of this.  They need to be able to see who else is accessing their records so they can have trust in the system, they need to be able to have control over that record and then around that we need the layers of people who can access that record because they’re providing direct care to that citizen for social care, for mental health, for primary or secondary care and then outside of that a layer that supports the work that we need to do around population health whether that’s research, service design, audit and planning. 

I think I could probably do a whole talk on learning health systems but I won’t, just to say that I think it’s clear that learning health systems are the way that we  need to go in the future to start to pull together more effective evidence basis and more real time evaluation.  So that idea of closing down the feedback loop between research and real world practice by feeding the researchers that real world evidence, getting their research out into practice more quickly and evaluating that and whilst DataWell is one component of that I think devolution and things like Connected Health Cities start to bring us more of that.

With that through devolution we have an organisation set up in Greater Manchester GM Connect to support us in thinking about cross system information, governance, we have a clinical leadership group that can support us in thinking about innovating clinical practice, we have Connected Health Cities across the north of England to start looking at how researchers work more effectively with real world health evidence and all of those steps are taking us closer to this vision of learning health systems.

I’m going to finish and say thank you very much for sticking with me on a whistle stop tour.