#OpenDataSavesLives: Session 39 - Virtual Wards

As we came out of the pandemic, we kept seeing a huge demand for our online sessions and we are still delighted with the impressive turnout and amount of engagement and discussion that comes out of these sessions. This time, we talked about some of the challenges trusts are encountering with patient flow, from admission and treatment duration through to discharge and the more recent pilots of remote monitoring and virtual wards in the UK.

We’ve narrowed down our focus quite a bit since our initial sessions, with a particular emphasis on inequalities. This was also relevant this time, where analysis of data on people waiting for treatment, as well as those who are waiting to leave hospital, are showing some predictable trends in terms of inequality by ethnic group and socioeconomic background.

Clara shared some of the problems hospitals are encountering in identifying and monitoring ‘which patients should be where’. Stephen then introduced a project that aims to solve this problem by piloting remote monitoring for 120 patients in the Leeds region. 

Clara Wessinger - Head of Planning & Performance, Kent & Medway NHS FT

Kent & Medway Foundation Trust have been developing a sophisticated model to answer questions such as ‘which patients should be where?’, ‘who is ‘fit to reside’ and who should be discharged from hospital?’.

In technical terms, they are modelling demand and capacity for discharge pathways, intending to build on this for forecasting and prediction of demand on hospital services.

One of the challenges we hear so often is that trusts and organisations have been working in silos, developing their own approach to the same problem with a lack of sharing and interoperability. Up to this point, the 4 acute trusts in Kent & Medway have had their own models of discharge and patient flow, with varying degrees of success. Clara and her team are now building a much needed single source of system-level data on hospital demand and capacity.

There are limitations and weaknesses to building such a model of course, the usual suspects - gaps in data collection and varying accuracy and availability between trusts - however, as Clara put it so rightly:

“Sometimes you just have to make the best of what you have and work with it, rather than constantly seeking to find the most complete picture… Whilst this may not be the perfect solution, we need to crack on and work to try and find solutions to the problem with the data that we do have at our disposal”.

Since the pandemic, trusts have seen a significant increase in the number of patients no longer fit to reside but remaining in hospital - people are staying in hospital longer than they need to. In recent months, this has started to tip back into a better direction.

To understand what is happening here, Clara and her team are pooling together data on average duration of stay over time. They have built an impressive model using data across three pathways (ways of being discharged from hospital) in the South East. Here is a link to the model - the downside here being that the platform is only accessible by those with an NHS login. If only the data were publicly available on the web in an open dashboard, then other trusts could access the data and code, adapt and build on it for their own operations...

You can see an overview of the model in the slide below.

Points from the discussion:

We are developing a good understanding of what the problems are, but not really why the problems are there in the first place. What are the processes hospitals are using to discharge patients? Are trusts doing it all in the same way? If not, where are the inconsistencies?

A lack of standardisation is part of the challenge - there is a lack of digital record of these processes to enable comparison, coupled with wildly different models of commissioning services that enable patients to be discharged. We need to understand which patients are leaving hospital and when, in order to understand and tackle the inequalities in patient experience I mentioned above.

With access to the platform, this analysis can be easily shared and reproduced between regions. However, without approved access, this isn’t possible. Analysis is so often locked behind logins and paywalls that make it difficult for others to pick up and interact with the data without a formal agreement with the solution provider. Whereas, if the data is open, everyone can benefit from it...

Stephen Blackburn, Innovation Relationship Manager, Leeds City Council

The technology used for remote monitoring offers a solution to some of the problems raised in Clara's talk. Virtual wards aim to reduce the amount of time that patients spend in hospital by providing technology to monitor them in their own home after they have been discharged, whereas remote monitoring aims to prevent people from being admitted to hospital or visiting their GP in the first place, by providing technology for home monitoring.

Leeds City Council are piloting remote monitoring in the region, for people who have diabetes and two other long term conditions and would otherwise need in-person care and monitoring. Over the next year, the pilot will include around 120 patients and the technology will be assessed for its effectiveness in helping them to manage their condition, improve their wellbeing and ultimately reduce the number of GP visits and time spent in hospital.

They are working with a number of solution providers, including MiiCare and Luscii - which collect data using monitors and sensors such as gait, blood pressure and heart rate, and the MiiCare platform is operated by voice and collects audio data. Each patient will need different data to be collected according to their condition. Information is fed back through the app and can be picked up by the GP on their patient dashboard.

Points from the discussion: 

How do we ensure equality in how patients are selected for the pilot programme? The wards selected are some of the most deprived regions in Leeds, and patients were chosen based on prevalence of diabetes and two other long-term conditions. Leeds City Council will be working closely with GPs to identify the full cohort, however this raises the same important question about process that we discussed with Clara. Like any medical trial, how do we know that patients are selected fairly and equitably for the pilot programme, and that regional differences in pilot success are not down to inconsistencies in this process? To do that, a digital record of the selection process will be needed.

And the data - a huge amount of which is being collected - where does this go and how do we maximise its value? To address this, Leeds City Council will be sharing as much data as possible with the Leeds Office of Data Analytics for insight into population health. We hope that these insights will be made publicly available as much as is possible with patient-level data and we’re looking forward to hearing the outcomes from the study.

Conclusion

Another lively session, with a huge number of questions and comments from people in the chat. We’re glad that our network is still so passionate and engaged with these topics. Lately, we’ve been working with bodies such as the Association of Professional Healthcare Analysts (AphA) to support those in the relatively new healthcare analytics profession and to ‘professionalise’ the industry. We’ll be holding a session on this later in the year so stay tuned for our updates on our website and Twitter.

Our next session will be 13th September with fantastic new speakers. Here’s the event page, we look forward to seeing you then!

As always, thank you to our sponsors and supporters that enable this initiative to continue. If you’re interested in what we do, then get in touch at hello@opendatasaveslives.org.