#OpenDataSavesLives: Session 41 - Tools & Methods in Open Health Data
Open Data Saves Lives was set up to address the need for better, faster use and sharing of data relating to the pandemic. It provided a space to convene and make progress rapidly, when there wasn’t time to wait for governance measures to be put in place. We found through our network of friends and colleagues, and through Twitter and the web, that lots of people were doing great work with COVID data, and they needed a platform to share it.
One of the lessons that came out of the pandemic was the importance of acting quickly, sharing data efficiently and working openly and collaboratively. The lack of data available in the first several months brought this into the public conversation, and our goal is to ensure that it stays there.
In this session, we’ll be talking some more about the value of working in the open. When it is shared, data can be used perhaps in ways that you wouldn’t expect. We heard from:
- Richard Irvine - Chief Data Officer, Leeds City Council/NHS West Yorkshire ICB . The role of Data Mill North.
- James Perry - Director, Egnida Innovation. They are developing a vulnerability tool to help end fuel poverty using open health data
- Wang Zhao & Matthew Stammers - The Whole System. They have been working on a new open source modelling tool 'Slowing the pace of Alzheimer's'.
Amazing work has been done in health, but often progressing this work is curtailed because organisations do not share their data. There is fear around publishing data that could be potentially disclosive and so the default behaviour is often to be closed.
Leeds City Council have been part of our network for some some time, and they bought into being open by default over 10 years ago. They were one of the first local governments to develop an open data platform. They now look after two open data directories - Data Mill North and the Leeds Observatory - and they have a growing number of partners who routinely share data on their platforms.
Richard made an interesting point about ensuring data quality. People often just upload data to the platform, but being a consumer of their own data makes it much easier to identify and resolve data quality issues - rather than relying on user feedback which is often few and far between.
We (Open Innovations) wrote a blog post about feedback on open data portals, that we recommend taking a glance over.
So what’s next for Data Mill North? Richard has many ideas for the platform, but is interested in hearing your thoughts. They aim to stimulate curiosity, and encourage more people to share data on the platform. There are hundreds of datasets on Data Mill North, covering local health, economy, culture and more.
The Council has a plan to make use of sensors in the region - capturing near real-time data (up to every 10-15 seconds) on things like air quality, humidity, footfall, transport. This data will be pushed into Data Mill North and openly available for use. With this data, they aim to understand wider determinants of health and care, as well as predictive factors for illness, health and wellbeing. For example, is there a higher incidence of respiratory conditions in regions of Leeds where air quality is low?
High resolution data will enable better, more informed strategy and delivery of services in the region. They aim to make this data easy to access, understand and use - therefore Richard wants to hear your thoughts on the data, and where to take the platform next. Get in touch with him on firstname.lastname@example.org.
James Perry, Egnida
We met James in quite a different context relating to the energy sector. James has been building a tool for identifying vulnerable customers that helps to solve fuel poverty using several data sources including health data. We were interested to hear from James what his experience was like working with primary care data from Fingertips, as a non-health domain expert. We’re also working with James on other projects and we see a lot of potential for a tool like this to identify and manage vulnerable customers in a health context - potentially customers that need to be contacted in relation to their health appointments, or communicated in extreme weather conditions.
The tool isn’t yet available to view online, but we are working with James to include new open sources of data in the tool such as warm spaces, and we see potential to create a rich resource for making geographical comparisons of metrics such as morbidity and mortality, incidence of illness, referral wait times and measures of outcomes.
This tools exists because people share their data openly, and so someone like James can find it, process it and combine it with other datasets to enable modelling across domains. When publishing the health data included in the tool, its owners may not have anticipated that it would be used for identification of vulnerable populations for the energy sector. In short, when the data is open, people can use it in innovative ways that you wouldn’t expect, and this vulnerability tool is an excellent example of how this can be done.
If you’d like to try out the tool for yourself or learn more, then get in touch with James on email@example.com.
Wang Zhao, The Whole Systems Partnership
Finally we heard from Wang from The Whole Systems Partnership. We’ve heard from them at Open Data Saves Lives before - last year, they joined us to share some of their information modeling on the impact of COVID-19 on cancer pathways. Their systems allow in-depth systems modelling and ‘what-if’ scenario analysis, which is difficult to understand in a system as complex as the health service.
This time, they shared a modelling tool they have developed for demonstrating the potential effects on the UK population of a new drug which slows the pace of Alzheimer’s. The tool will help the NHS plan Dementia services and is available for all to use - and they want people to use the tool to understand the implications on health services following the likely introduction of this new drug.
The model is available on Really Useful Model’s github. Matt Stammers and Wang Zhao are the data scientists who’ve linked the systems dynamic model built on Stella Architect to Streamlit using Python and AI. They also use Agile Systems Dynamics Modelling which is an open-source, python-based tool also on Github.
In the session, they talked through how they built the model - you can view the recording here, and explore the code on their Github. We love their approach to building tools that help healthcare professionals make decisions and model outcomes, whilst working entirely in the open.
We want Open Data Saves Lives to continue to act as a platform for people to share the great work they’re doing with open health data. There were three things that came out of our last session:
Data needs to be shared openly - platforms like Data Mill North and the Leeds Observatory enable this. Feedback helps to make them better, and the more people use and share their data, the more they become a rich and useful resource for making decisions and delivering services.
When data is open, it can be used in ways you wouldn’t expect. James’ vulnerabilities tool demonstrates this. If the data is available and on the web, then it can be used by anyone - not just healthcare professionals and domain experts. It enables innovation across sectors.
These models can therefore be used to help healthcare professionals to make decisions and model outcomes. As we saw from Wang at the Whole Systems Partnership, better modelling of impact and outcomes allows more informed conversations and decision making.
Our next session will be the last one hosted by Open Innovations - so we’d love for you to join us and celebrate all that we’ve achieved in the last 3 years since the inception of Open Data Saves Lives. We’ll be joined again by people who’ve been working on impressive projects with open health data. Here’s the link to sign up and we’ll see you there.