October 2024 updates
***Something new***
This month, we thought we would try something new in AnalystX. Something in true AnalystX fashion: an initiative that celebrated plurality of thinking, divergent ways of doing yet all grounded in the convergent ethos of let’s continue to use data and analytics to drive better patient health outcomes and better citizen care.
?So, we decided to hand the pen for the monthly newsletter to one of our initial founders Hassan Chaudhury HonFAPM to provide his sagely perspectives on the world: local and global (or glocal) and private, public and third sector fusions. We are immensely grateful to Hassan for his perspectives below:
The Next 5 Years: 5 Big Trends in Health and Care Data and Analytics
There has never been a better time to be part of the health and care data and analytics community; more recognition of our value, more of our insights built into decision-making, more steps towards our professionalisation, more data available than ever, and more powerful tools.
If we want to take full advantage of these favourable conditions, it’s important to take stock of what the trends are, and where it could lead us over the next 5 years so we can maintain our momentum. Here are 5 trends that I am coming across, as well as what each might mean for us, with the proviso that my perspective is just as biased as anyone else’s:
1.???? More effort to curate or pre-curate data (rather than leave the data as is until its required)
2.???? Embracing multimodal data insight (because we are digitising different modes of data)
3.???? Sharing digital infrastructure and resources (due to the high set up costs)
4.???? Data now has value again (especially if we bargain collectively)
5.???? More industry partnerships (with clarity on our terms)
1. Curation
Data curation is a big topic and one worth investing time into, touching data stewardship and more. The point here is that health and care data in its raw form isn’t always usable for our specific use cases without putting in detailed effort, as if we were curating art or historical artefacts in a museum. This is more than placing data assets into a data lake, and could include ingesting cleansing, structuring, transforming, and indexing it to raise its level so it better meets our needs. Pharmaceutical companies engaging health and care systems often speak of requiring real world data at regulatory grade, by which they mean real world data curated to regulatory grade. Curation at scale means we don’t have to curate piecemeal on demand, but that requires forethought and strategy.
As we enter into a world where messy data that isn’t held uniformly has less value than high-quality curated data, the trend will be for more curation and not less. Data curation is an investment in capability and capacity. It means, for example, that we can reliably get data out of our EPRs. It might also mean a locked room for us to look through paper notes in privacy.
This means we should plan early and invest in curation to reap the benefits downstream; it’s far harder to curate on demand.
2. Multimodal data
Health and care data has many ‘modes’. These include direct structured data from an EPR, digitised imaging, digitised pathology, molecular and genetic profiling, and preference data direct from patients. We could even include sensor data from wearables. The point is that the insights we can derive by uniting multiple data types, i.e. a multimodal standpoint, are far more powerful and compelling than a unimodal one. It’s far simpler to join data before we have a query to make, than to try to join it after.
The more data types we digitise, the easier it is to imagine more unified and conjoined data assets and queries. This is not a simple task, and much credit has to go to the organisations with the vision to be working on multimodal data. As an example, Leeds Teaching Hospitals NHS Trust (LTHT) hosts DATA-CAN; the HDRUK Hub for Cancer, with the National Pathology Imaging Co-operative (NPIC) already located as close as can be at St James’s Hospital in Leeds. Add the fact that the Yorkshire and Humber region already has ambitions around data, led by their Health Innovation Network (HIN), and we can imagine what might be happening there.
This means we should consider how to join and also conjoin different modes of health and care data from the outset.
3. Shared infrastructure
It’s not controversial to state that the infrastructure required to truly leverage data and analytics in health and care is going to cost, which is difficult when there NHS has to contend with trolley waits in A&E, and delayed transfers of care (DToCs), and our care sector has to worry about early detection of deterioration, and prediction of emergency events.
?Infrastructure matters more now as we have moved decisively to a model where data remains in situ and the query comes to the data, rather than have sensitive data move with all the inherent risks. This infrastructure requires digital maturity as the components and related aspects are not ‘plug and play’.
?However, can we really expect digital maturity in data and analytics across the NHS and care sector anytime soon? HIMSS has set up ANAM - the Adoption Model for Analytics Maturity. It covers 7 stages like other HIMSS models like EMRAM. Consider where your organisation might fit on the continuum:
(0) Fragmented point solutions
(1) Data aggregation and initial data governance
(2) Centralised database with an analytics competency centre
(3) Internal and external repost production and agility
(4) Measuring and managing evidence-based care, care visibility, and waste reduction
(5) Enhancing quality of care and population health
(6) Clinical risk intervention and predictive analytics
(7) Personalised medicine and prescriptive analytics
More information at: https://www.himss.org/maturity-models/amam/
Digitally elite centres might be aiming high, but this is not going to be replicated across the NHS or other elements of the landscape - that is simply not how we have ever or will ever operate. This is why the trend is going to be for those centres who have invested and put in high sunk costs (in the millions), to be offering other public body organisations and centres the opportunity to share digital infrastructure and resources. It simply doesn’t make sense for everyone to have a pool of experienced data engineers and data scientists plus all the supporting structure, and this is something I am working to create at Great Ormond Street Hospital for Children NHS Foundation Trust (GOSH) [https://www.gosh.nhs.uk], one of the best children’s hospitals in the world and the most digitally advanced NHS Trust. Its DRIVE unit [https://www.goshdrive.com] is already leading the way and could ensure that others do not have to invest upfront or go through the pain of learning how to do what is required.
This means we should consider joining forces and conserving our limited resources early on and not waiting for opportunities to present themselves before we explore these partnerships.
4. Data has value
A few years ago, EY attempted to define the value of a curated NHS data set. The report stated: “We have estimated that the value to the NHS in terms of operational savings, enhanced patient outcomes and generation of wider economic benefits to the UK could be as much as ￡5bn per annum and deliver around ￡4.6bn per annum of benefit to patients” - https://www.ey.com/en_uk/insights/life-sciences/how-we-can-place-a-value-on-health-care-data
Not everyone agreed. Part of the disagreement was that data has no intrinsic value. While it sits inert and unused, it just costs us to store it; it must be leveraged for us to extract value. This meant the key became the insights rather than the data. Neither side won the argument conclusively, but it was clear that the value of data as the new oil was being challenged.
Now we have AI Foundation models, the pendulum has swung back viciously the other way. These models act as a base for other models and can have applications built on top of them e.g. how OpenAI’s GPT-4 model underpins their conversational chat agent, ChatGPT.
Vendors are now trying to join in on the hype, often to bolster their investment stories ahead of a raise and possibly increase their valuations, by offering NHS organisations deals to build AI Foundation models for them. These models require tonnes of data for training, but they also require data of a certain type and pedigree to successfully answer the inevitable questions; who did you work with and what did you train your model on? NHS data, collected across multiple care settings from cradle-to-grave for decades and for millions of lives, lends itself tremendously well to AI Foundation model development. This offers operational and clinical value to our health systems, which has evident value for our patients and populations too.
This now means we can argue that NHS data has more intrinsic value, and we have now come full circle. The upshot of this is that collective bargaining (especially where we pool curated, multimodal data on shared infrastructure) will help us far more than if a few elite centres strike out on their own to sign deals. Industry is often torn between wanting to work with the data assets of many NHS and care organisations to increase the power of studies but negotiate separately knowing that they have an advantage of knowing more about the value of the project that each individual organisation.
?This is important as the information asymmetry when we make deals with industry can hurt, as we may not quite understand the value of our assets and may end up underselling our time and effort.
This means we should consider how to maximise the value and return on our data assets while ensuring we stick together rather than be picked apart.
5. More vendor partnerships
Now we have established that curated, multimodal data has immense value (operationally, clinically, financially), and we have the shared infrastructure to enable collective bargaining, what is left is that we are careful to only work with industry partners who share our values and appreciate our ‘red lines’ and dealbreakers.
Red lines could be as simple as not wanting data to be used to target healthcare professionals or for insurance purposes. Red lines could be that we ensure patients, and the public are fully involved and engaged in the purposes behind secondary data use and can veto projects that do not meet the right standards. Red lines could be that there is genuine fair value exchange, after all if there’s value to be generated why shouldn’t the NHS and care sector have its fair share?
The right vendors and partners are out there, but the onus must be on those who hold the responsibility for data, or in GDPR terms data controllers and processors, to set the terms of the discussions. This is not just about service wrappers, although that it also an essential element. It is about having the confidence to work together with the right bodies for example techUK, AnalystX, AphA, BCS, HDRUK etc. working in concert to design a framework that means we all achieve our goals.
Conclusion
Our data and analytics community of communities is full of health and vigour. I can see that if these 5 trends continue and converge over the next 5 years, we will see momentum for the NHS and care sector when it comes to using its own data and getting value for it.
It will require us to understand curation of multimodal data needs to happen upstream and not on demand. It does mean we need to consider sharing digital infrastructure and resources because we can’t all absorb the sunk costs. It means we appreciate that the value of our data assets requires some form of collective bargaining so we can do better in our negotiations with industry, and it means we will inevitably have more of these agreements.
I may be completely wrong about the landscape at the end of the following 5 years but we all must make hard choices about which trends we can see, which we will attempt to influence, and which we will attempt to take advantage of.
What matters for us in health and care data and analytics is that we understand these trends and we can get ahead of them so we can take full advantage.
***Continuation of something new***
We would love to hear our readers’ perspectives so please do react to the above and provide any great examples of ongoing work that is aligned to the five trends. One of the great strengths of AnalystX is the ability to allow the community to share examples of great pieces of work or even challenging ones.
Now over to what the community has been doing during the month of September.
AnalystX update:
Workforce professionalisation:
September has been the countdown to the first anniversary of the launch of the National Competency Framework. We have successfully launched our Fika Bites podcast series, 8-minute chats with different people involved in the development of the framework. We have also released 3 case studies around sites that are adopting the NCF, a series of articles and of course some Andy Lavelle videos about the framework.
October sees us kick off the new development work on the NCF. As we expand it to include 4 new roles: Analytics Engineer, Machine Learning Engineer, Data Ethicist and Data Governance Manager. As always, we want the community to get involved so, please register you interest here. We will also be out and about with Sarah Blundell speaking on a woman in healthcare data panel at Snowflake’s World Tour event in London on 10th October.
Learning and Development:
September also saw us begin the work on School and University Outreach as part of the pillar 1 New Entry talent development for the NHS Data and Analytics Academy. Work is beginning in October on the development of the online challenge materials, with a launch event planned in December, and other events planned for January, February and March 2025. There is still time to get involved in this work as a volunteer so please register you interest here.
Huddles:
September has been a busy month for Huddles; we had three huddles including Process Mining in the NHS/NHSE, Uncovering? Amazon’s culture of innovation and Data Science, AI and Technology in the OneLondon Secure Data Environment. All three huddles had great interaction and engagement from the audience, but we unfortunately ran out of time. All recordings are now on the AnalystX Huddles FutureNHS workspace for you to access in your own time.
We have more huddles coming up in the October starting with a refreshing repeat of the Multimorbidity and Markov Modelling huddle which happened in April. We had technical issues at the time so you will be glad to know that @Adrian Pratt was happy to come back and share his expertise on this important topic again on the 8th of October 2024. You can sign up here
Please see our Huddle schedule for more details of our upcoming huddles.
Communities update:
Data viz and BI community
Sonu Mittal Alistair Bullward Rozanne Addams Farhad Hossen
We are a tool agnostic?community.?Our?community?is open to anyone who is working in Healthcare, Data, Analytics,?BI?and?Viz?space. We believe that together we can help each other to overcome our challenges?and?to achieve our goals. The purpose of the?community?is to
·?????? Provide a framework for knowledge sharing.
·?????? Support people working across a range of tools, technologies and?platforms.
·?????? Make learning?and?development?material available?and?accessible to upskill?and?to professionalise?our workforce.
·?????? Provide an innovative, collaborative and?safe space for users to generate ideas?
·?????? Share best practices
·?????? Share your journey with others
·?????? Offer free hands-on training sessions where possible
·?????? Membership:?all health and care sector
Meet-ups are every 6 weeks for an hour.
You can join us by completing registration?form.
Our slack channel is open to all?and?is a place to share?and?pick the collective brain, you can join us at: Data?Viz?Slack
On the NHS MS Teams: BI?and?visual analytics specialism?Community?of Practice | General | Microsoft Teams
The AnalystX Data?Viz?for Health?and?Care channel: https://data-viz.analystx.uk/
For?further information?please contact: [email protected], [email protected]
Update:
We had our "BI?&?Viz?community?of practice (COP)" meet on 10th?Sept 2024.? It was lovely to see all the participants?and?our presenters.? Jennifer Struthers presented "open-source tools plotly and dash".? Janet Broome?and Rebecca O'Connor?from Snowflake presented?"Modelling Data with Snowflake".? Thank you both for this useful hands-on training ??????. ??All the sessions are recorded?and?recording links are available on request.
Upcoming trainings?update:
29/10/24 - Introduction to Snowpark - how to programme against Snowflake using Python?and?other languages. How notebooks are supported in Snowflake.
10/12/24 - Generative AI - Working with Document AI
28/01/25 - Working with Geospatial in Snowflake
We are working on to arrange trainings for PowerBI, SQL?and?basic statistical concepts. If you would like to present, then please give us a shout?and?we will arrange it.
The next UK Health care Tableau user group will be held on the 9th October 1-5pm hosted by Charlotte Shearsby and Steve Adams for more information, please click here
?
Data Engineering Community of practice
领英推荐
Data Vault & Your Feedback
This month, our meetup was focussed on Data Vault which @Paul Ellingham and @Nick Lefley walked us through.
Matthew Bishop prepared and shared a questionnaire to help us continue to shape our community in the right way.
We’re in the midst of arranging some demonstrations from Microsoft, Oracle and AWS for upcoming forums.
Our data engineering steering group would like to hear from our data and analytics colleagues to refine our deliveries. Please complete the questionnaire here
Federated Data Platform
Want to learn more about the Federated Data Platform??
Why not explore the resources on the FDP Centre of Excellence (CoE) Resources for analysts, developers etc
We ran the first “How To use Contour session”; and we had a great turnout with over 70 attendees. The feedback was very positive and the questions in the chat indicated a good level of interest. The recording is available on the CoE here, so you can watch and learn at your own pace.? NB.? Contour is a point- and -click application that enables users to perform large scale data analysis on tabular data bases and is an excellent starting point for analysts wanting to start their journey into FDP.? You can practice your new skills and build your own contour using synthetic data, on the new FDP Solution Exchange.?
Coming up next on FDP CoE: ?We've got FDP tooling webinars coming up including Code Repositories and How to Use Pipeline.?? We also have a fortnightly drop in where you can ask Palantir and NHSE experts any questions.
There are also online learning materials on the FDP home page >?FDP CoE >?Academy >?Platform users >?Select your role and a syll
Keep your eyes out for continual improvements to the FDP CoE.?
To get access to the FDP,?please see the guide here:? NHS FDP Training - NHS National Data Platform - FutureNHS Collaboration Platform.?
Also sign up to the FDP newsletter here:? Get the latest from Federated data platform (FDP) (confirmsubscription.com)
Process mining community
The launch of our AnalystX process mining community of practice was taken place via our AnalystX huddles, if you have missed the session, you can watch the recording Process Mining in the NHS
For more information about process mining and how to apply in your organisation, we have created a page with introduction, list of open and completed projects, learning materials and a list of contacts for questions and queries.
Nhs.pycom
Nhs.pycom and Zella King presented a topic on- ?python resource for modelling of short-term demand for emergency beds.
To revisit this session, please go to NHS.pycom webinar
Generative AI working group.
Our GenAI working group collaborated with NHS England Artificial intelligence Quality (AIQ) community of practice, sharing information about who they are and what they do.
The recording is available in our Generative AI - AnalystX - area.
To learn more about our Gen AI use cases and example, or if you would like to join our team, please visit Generative AI - AnalystX - FutureNHS Collaboration Platform
To sign up for future events for the Gen AI working group, please use this form.
Strategic partners events and updates:
Cambridge spark
Registration for the last data & AI apprenticeship cohorts of 2024 will be closing on 25th October. For more information visit Data Apprenticeships for the NHS and Healthcare Professionals (cambridgespark.com)
ESRI UK
Facts and Dimensions:
We are privileged to work with Facts and Dimensions to provide basic sql training to all our members, for course material , how to register and future courses, please visit the Facts and dimensions - AnalystX - FutureNHS Collaboration Platform ?
HDRUK
Online Webinar: 9 October 11:00am
This free webinar is for anyone who wants to receive an overview of good coding practices, writing code and ensuring good quality of their code.
Interactive Workshop: starting 30 October 1:00pm
In health research, we are increasingly reliant on research software. Many of us are developing data analysis scripts, code, or even entire software packages. This course is designed for those who write code and want to ensure good quality of this code.
Snowflake
Join Snowflake, GMC and NHE England for an insightful session that delves into the experiences of women working in data within healthcare and the NHS. Our panelists Emmi Poteliakhoff MBE Mridula Sori Camellia Williamson Sarah Blundell will share their unique data journeys, give valuable insights into the challenges they’ve overcome, the advice they’d give to their younger selves, and how women can shape the future of data in healthcare.
This event is part of Snowflake World Tour 10th October Excel Centre
South East Digital Skills Development Network
The October edition is out now - October Newsletter Out Now! | LinkedIn
Microsoft
Ever wanted to learn more about Gen AI. there are some courses available from Microsoft.
Public Policy Project
Public Policy Projects (PPP) have recently launched an impactful report covering the findings from the Integrated Care Delivery Forum – Spring conference in Birmingham. If you have an interest in up-to-date insights on the sector, check out the insights from PPP's Spring Forum.?
The next Integrated Care Delivery Forum – Autumn event will feature over 50 expert speakers, including ICS leaders, NHS stakeholders, and community representatives. Join us to foster collaboration and innovation in integrated care.?
Learning and development material:
Below is a list of strategic partner academies with free learning for NHS colleagues.
Snowflake
?
?
Microsoft
?
Databricks
Celonis
?
?
Oracle Academy
Special thanks to all our AnalystX supporters:
Ming Tang Ayub Bhayat Emmi Poteliakhoff MBE Ed Kendall Sukhmeet Panesar Alex Cheung Janet Broome Benjamin Tucker Rowan Bradnum Danny Silk Lauretta Anderson Peter Rodgers Steve Adams Aslam Atcha Tilly Allen Abigail Leachman Hugh Neylan Sarah Culkin Ian Townend Shabs (Shabbir) Lorgat Steve Faro Sarah BlundellSamantha Menot Paul Reid Paul Parker Hassan Chaudhury Alex Lawrence Robert Walker Elisa Anastasi Dr Mustafa Ghafouri Tristi Tanaka Leontina Postelnicu James Aitman Simon Wellesley-Miller Charlotte Lewis Angela McFarlane Rony Arafin Rosalind Way Libby Duane Adams Sukhmeet Panesar Ruth Holland Steven Totman Lauren Chadwick Filipe Santos Filipe Gracio, PhD Filipe McManus Mark Bailey Tiffany Young Phillip Akers Chris Howarth Christine Banks Chris Habberley Karina Gajewska Kwesi Afful Mark Shill David Bibbs Chris Love Zo? Turner Chris Mainey Emma Karoune nicola troup Liam Mahon MSc North West Skills Development Network Skills Development Network Yorkshire & Humber Branch Skills Development Network North East and North Cumbria
?
?
Data Visualiser for the NHS | Co-lead UK Healthcare Tableau User Group | 2024 Tableau Ambassador
5 个月Thanks for including us AnalystX!
Fractional CxO | Driving Innovation in Strategy, Digital, Data, and Analytics | Focus: All Things Health |Simplifying Complexity, Delivering Transformative Solutions | Working in the Grey Spaces between People and Tech
5 个月Alex Cheung superb work as always!
Tableau Ambassador 23/25. Helping time-starved professionals Excel @ Tableau. Tableau User Group Co-Leader
5 个月Another brilliant newsletter! And thanks for mentioning the Tableau Healthcare User Group happening virtually THIS week.
Ramadhan posts in Ramadhan
5 个月A super update that highlights just how lively this community of communities continues to be!