BSMS > Research > Primary care and public health > Health Data Science > ASTRODEM > News-and-publications

News and publications

Presentations and talks, publications and collaborations relating to the Astrodem project.

BACKGROUND IMAGE FOR PANEL

Latest news

Astrodem team investigate implementation of clinical decision support systems in general practice

The Astrodem project was centred on developing algorithms to aid GPs in identifying or recording dementia sooner than they otherwise would. However, much research has shown that algorithms which aid clinical decision making or recording (clinical decision support systems, CDSS) are poorly taken up by GPs. We interviewed GPs to understand what factors make a good CDSS which is likely to be used, and what features make CDSSs work poorly in the clinic. Based on our findings we make a number of recommendations for CDSS developers when designing a new system. Read our findings online in the journal BMC Medical Informatics and Decision Making below.

Read the findings here >

Astrodem team publishes findings in BMJ Open

UK statistics suggest only two-thirds of patients with dementia get a diagnosis recorded in primary care. GPs report barriers to formally diagnosing dementia, so some patients may be known by GPs to have dementia but may be missing a diagnosis in their patient record. This study aimed to produce a method to identify these 'known but unlabelled' patients with dementia using data from primary care patient records. The team were able to train machine learning models to correctly identify patients who had dementia but no recorded diagnosis with an accuracy of 94%, interestingly models were improved substantially by including information from clinical notes and letters. Usually this type of unstructured patient data is unavailable to researchers.

'Automated detection of patients with dementia whose symptoms have been identified in primary care but have no formal diagnosis: a retrospective case–control study using electronic primary care records' is open access here >

Astrodem team publish findings in Wellcome Open Research

The Astrodem team has published their latest findings in Wellcome Open Research. Using 77 coded concepts which may precede a dementia diagnosis, the team investigated how early a machine learning model could accurately identify dementia, and if these models were better when predicting Alzheimer’s disease or Vascular dementia. We found that models only made accurate classifications in the year before dementia diagnosis, when GPs were already screening and referring patients for dementia symptoms. The full results can be read via the link below.

Read the Wellcome Open Research results in full >

Astrophysics and AI may offer key to early dementia diagnosis

Crucial early diagnosis of dementia in general practice could improve thanks to a computer model designed in a collaboration between BSMS and astrophysicists at the University of Sussex.

Currently, only two-thirds of people with dementia in the UK receive a formal diagnosis, and many receive it late in the disease process, meaning that a large number are missing out on the care that could help them achieve a good quality of life.

The team, led by Dr Elizabeth Ford, Senior Lecturer in Primary Care Research at BSMS, used data from GP patient records to create a list of 70 indicators related to the onset of dementia and recorded in the five years before diagnosis.

Read how Astrophysics and AI can help early Diagnosis >

Blue graphical image showing the outline of a head and brain overplayed over some graphical representations of computer circuit boards

Student Research Associates work on Astrodem over the summer
The Astrodem team was pleased to host two students over the summer period (2019). Medical student Rish Jain and physics student Paula Ramirez worked with Liz Ford and Richard Hoile, looking for indications of why some people get a dementia diagnosis recorded and others don’t. Using a small number of primary care records of patients who seem to have many indications of dementia (e.g. symptoms, family involvement in care), half of whom had a diagnosis for dementia recorded and half of whom didn’t, we used data driven methods to try to explain the lack of diagnosis in the patients without one. We found mixed results which we are hoping to explain with input from our clinical colleagues; further work will look more holistically at the social and family circumstances of the patients to understand if those give us a clue to the diagnostic process.

Headshots of Rish Jain and Paula Ramirez side by side

Astrodem student intern secures PhD place
Joanne Sheppard, a University of Sussex Physics undergraduate student who worked with the Astrodem team as a Junior Research Associate in the summer of 2018, has gone on to successfully secure a funded PhD place at UCL in the newly formed UKRI Centre for Doctoral Training in AI-enabled Healthcare. Joanne worked with Liz programming machine learning models using a big dataset of primary care patient records, to try to find patients with dementia who were not yet given a formal diagnosis and her work will shortly be submitted for publication. We wish Joanne every success in her future studies!

Joanne Sheppard headshot

Presentations and talks

13 January 2021
Liz Ford gave a presentation entitled Can we use digital technology to detect or predict dementia, and if we can, should we?, at an online seminar discussing Computing Applications in Digital Health, hosted by the University of Brighton Research Centre for Secure, Intelligent and Usable Systems.

15 October 2020
Liz Ford described the Astrodem study at a seminar called Applications of Machine Learning to Improve Healthcare Delivery for Older Adults, which was hosted by the US National Institute on Aging (NIA) Division of Behavioral and Social Research (BSR), part of the US National Institutes of Health. The meeting focuses on identifying research gaps and opportunities of how researchers can apply machine learning approaches to improve health care delivery for older adults.

11 February 2020
Liz Ford gave a presentation at the Brighton and Hove meeting of the University of the Third Age. Liz’s talk spanned describing electronic patients records, how they are used for research, the ethics of this, and how the Astrodem team have used these records for research on dementia.

23 October 2019
Liz Ford gave a presentation about research using electronic health records, including results from Astrodem, at the Brighton Digital Festival Event called DATA HEALTH AND THE ARTS: Creating space, bridging boundaries.

7 October 2019
Liz Ford gave a talk to about research using electronic health records, including results from Astrodem, at the Brighton and Hove Ageing Well Festival at the event Age-Friendly Coffee and Converation – Health and Social Care.

2 October 2019
Liz Ford gave a talk at an event run by Zinc, a social enterprise dedicated to addressing some of society’s most pressing challenges by fostering innovation. Zinc has a mission-led approach and Liz presented on results of the Astrodem study to spark ideas for the “Five healthy extra years in later life” theme.

Liz Ford standing in front of a powerpoint slide, doing a talk at the Department of Health and Social Care

16 September 2019
Liz Ford gave a presentation at the Department of Health and Social Care at a workshop organised by the Open Innovation Team at the Cabinet Office on methods and ethics of Population Segmentation. Liz presented on the ethics of using patients’ NHS electronic patient data for population based research and on data quality issues with this data.

28 March 2019
Dr Liz Ford and colleague Dr Lamiece Hassan from University of Manchester and HDR-UK ran a public engagement event at the Alan Turing Institute on developing trustworthy data governance standard for the use of medical free text in research. See Lamiece’s blog about the data govenance event here >

24 April 2019
Dr Liz Ford recently chaired a panel session at the 2019 Healthcare Text Analytics Conference in Cardiff, entitled 'Ethics and Governance in text mining for trustworthy health research: progress and opportunities.'

Dr Liz Ford standing with a laptop and microphone in front of some slides at the Healtex Healthcare Text Analytics Conference

19 October 2018
Astrodem team member Johannes Starlinger has written an article about AI and precision medicine on the website dataconomy.com.

6th September 2018
Liz Ford presented the ASTRODEM project at the Team East Sussex Ageing Society Meeting held at the University of Sussex. Team East Sussex (TES) is the local federated board for the South East Local Enterprise Partnership (SELEP). Membership of Team East Sussex is made up of business, local authority, social enterprise and education representatives from across the county.Presentations and discussions at the meeting were captured by a cartoonist, and ASTRODEM was captured as shown.

Cartoon of the proceedings capturing presentations and discussions at the ASTRODERM meeting

19 June 2018
Liz Ford was invited to speak at Sussex Impact Day 'Breakfast with Impact - Spark your Curiosity'. Among other speakers from the University of Sussex, and introduced by the Vice-Chancellor Prof Adam Tickell, Liz spoke about the Astrodem project in a talk entitled 'Using astrophysics to close the diagnosis gap for dementia'.

24-26 April 2018
Several members of the Astrodem team attended the Medical Informatics Europe Conference 2018, Gothenburg, Sweden, 24-26 April. Dr Phil Rooney presented his paper entitled 'Modelling uncertainty in diagnoses in UK primary care records: development of a method using probabilistic programming with Bayesian priors'. Dr Liz Ford presented her paper entitled 'Development of a model for finding unlabeled cases of rheumatoid arthritis in UK primary care patient records'.

18-19 April 2018
Several members of the Astrodem team attended the UK Healthcare Text Analytics Conference (HealTAC), Manchester 18-19 April. On Wednesday 18 April, Shanu Sadhwani presented her paper entitled 'Understanding underdiagnosis: The potential contribution of free-text to dementia research'. On Thursday 19 April, Dr Liz Ford chaired a panel discussion entitled 'Understanding how to gain public trust in healthcare text analytics'.

30 January 2018
On 30 January 2018, we held a public event to discuss people’s opinions about sharing their medical records for research. This well attended event was run in conjunction with Healtex and attendees discussed the perceived privacy risks of sharing the 'free text' portions of their medical record as well as the coded information.

7 November 2017
Liz Ford was invited to speak at the Science Match Future Medicine conference in Berlin. In a three-minute talk, Liz presented the aims, objectives and preliminary results of the Astrodem project.

7 September 2017
Dr Liz Ford contributed to the British Science Festival event 'Improving the quality of life for people with dementia' together with Prof Sube Banerjee (BSMS) and Dr Penny Dodds (University of Brighton). From astrophysicists using theories on galaxy discovery to help GPs spot people at risk of dementia, through to PARO, a robotic baby seal designed to help those with it, this event explored what living well with dementia looks like and how society and services can contribute.

6 September 2017
Shanu Sadhwani chaired a symposium at the Division of Health Psychology Annual Conference, Cardiff, entitled 'Helping carers to care: Supporting the formal and informal carers of people with dementia', and presented her first PhD study 'Stigma, Nihilism and Uncertainty: GP barriers to diagnosing dementia. A qualitative study'.

19 July 2017
Dr Philip Rooney and Dr Liz Ford visited CPRD to present two talks, entitled 'Under-diagnosis of mental health conditions in general practice' and 'Understanding the True Correlation Between Conditions Using Electronic Health Records and Probabilistic Programming'.

12 July 2017
Shanu Sadhwani presented at the Society of Academic Primary Care Annual Conference in Coventry, results from her first PhD study, entitled 'What shapes GPs decisions to pursue a diagnosis of dementia? A qualitative study'.

11 July 2017
Liz Ford gave a presentation to the Institut für Informatik at Humboldt-Universität zu Berlin, Germany, detailing the aims and preliminary results of the Astrodem project. Liz spent two days in Germany working with Dr Johannes Starlinger, to set up a collaboration in which Dr Starlinger will help the Astrodem team to define longitudinal predictors of dementia, by clustering patients who have similar patient pathways. During this time, Drs Ford and Starlinger also visited Doc-Cirrus, a tech company developing and providing software to GPs in Germany, to learn more about how to develop our predictive models in a way in which they can be quickly implemented in GPs’ clinical workflow.

9 May 2017
Sussex Data Analysis Forum. Dr Elizabeth Ford, "Forming collaborations in health data science and understanding health data". Liz talked about the history of the Astrodem project and what data analysts need to know if they want to work with routinely collected health data in mental health.

28 April 2017
NIHR day at University of Sussex: What can the NIHR do for you? With chief scientist Chris Witty. Dr Elizabeth Ford presented on "Health Informatics Research at Brighton and Sussex Medical School", including presenting the aims of the Astrodem project.

25 April 2017
Informatics for Health 2017, Manchester UK. Conference of the Farr Institute and the European Federation for Medical Informatics. Dr Elizabeth Ford, "Which variables are useful for phenotyping dementia in primary care records? A meta-analysis". Presented in the session "Electronic Phenotyping."

24 April 2017
Informatics for Health 2017, Manchester UK. Conference of the Farr Institute and the European Federation for Medical Informatics. Dr Philip Rooney, "Understanding the True Correlation Between Conditions Using Electronic Health Records and Probabilistic Programming" presented in the session "Machine Learning."

28 March 2017
Sussex Data Analysis Forum. Dr Philip Rooney, "Applying Probabilistic Programming Techniques to Electronic Health Records."

30 March 2017
MQ Mental Health Research Charity Data Science Event, Manchester, UK. Dr Elizabeth Ford was invited to speak on "Challenges and Opportunities using Primary Care Data for Mental Health Informatics" and presented some preliminary findings from the Astrodem project.

BACKGROUND IMAGE FOR PANEL

Publications

Ford, E., Edelman, N., Somers, L. et al. (2021) Barriers and facilitators to the adoption of electronic clinical decision support systems: a qualitative interview study with UK general practitioners. BMC Med Inform Decis Mak 21, 193.

Ford E, Starlinger J, Rooney P et al. Could dementia be detected from UK primary care patients’ records by simple automated methods earlier than by the treating physician? A retrospective case-control study [version 1; peer review: 2 approved]. Wellcome Open Res 2020, 5:120.

Ford E, Rooney P, Hurley P, Oliver S, Bremner S and Cassell J (2020) Can the Use of Bayesian Analysis Methods Correct for Incompleteness in Electronic Health Records Diagnosis Data? Development of a Novel Method Using Simulated and Real-Life Clinical Data. Frontiers in Public Health 8:54. doi: 10.3389/fpubh.2020.00054.

Ford, E., Rooney, P., Oliver, S. et al. (2019) Identifying undetected dementia in UK primary care patients: a retrospective case-control study comparing machine-learning and standard epidemiological approaches. BMC Med Inform Decis Mak 19, 248.

Hoile, R., Tabet, N., Smith, H., Bremner, S., Cassell, J. & Ford, E (2019): Are symptoms of insomnia in primary care associated with subsequent onset of dementia? A matched retrospective case-control study, Aging & Mental Health.

Ford, E., Boyd, A., Bowles, J. et al., (2019). Our data, our society, our health: A vision for inclusive and transparent health data science in the United Kingdom and beyond. Learning Health Systems 2019;e10191.

Ford, E., Greenslade, N., Paudyal, P., Bremner, S., Smith, H.E., Banerjee, S., Sadhwani, S., Rooney, P., Oliver, S. and Cassell, J. (2018). Predicting dementia from primary care records: a systematic review and meta-analysis. PLoS ONE 13(3): e0194735.

Stockdale, J., Cassell, J., and Ford, E (2018) “Giving something back”: A systematic review and ethical enquiry of public opinions on the use of patient data for research in the United Kingdom and the Republic of Ireland Wellcome Open Res 2018, 3:6.

Invited attendances

22-24 June 2018
Liz was invited to attend a science camp at Google X campus in California. For the past 13 years Digital Science, Google, O'Reilly Media and Nature have organised an annual Science Foo Camp ("SciFoo" for short). Each year, the event brings together more than 250 people from around the world, who undertake groundbreaking work in diverse areas of science and technology. The event is invitation only and operates under Chatham House rules. Liz co-hosted a discussion session entitled 'From innovation to implementation: how to engage end users in the scientific process and pipeline for real world impact', which was well attended. She enjoyed attending sessions on diverse topics such as understanding what 'intelligence' is in light of the development of artificial intelligence, on designing human-machine systems for public trust and the social good, and what community structure we need for data-intensive health research. Liz made some great connections with other international academics who work with health big data, as well as industry representatives.

17 April 2018
On Tuesday 17 April, Dr Liz Ford presented a seminar in the Centre for Primary Care, University of Manchester seminar series, on the latest results from the Astrodem Study.

22 September 2017
Philip Rooney attended the stakeholder consultation workshop on machine learning and data mining based research using electronic healthcare record data at Clinical Practice Research Datalink (CPRD). This workshop aimed to inform the CPRD policy and scientific review process for research using machine learning approaches.

19 June 2017
Liz Ford participated in the Wellcome Trust meeting 'Future Technologies and the Use of Patient Data'. The aim of this meeting was to discuss partnership working between the NHS and technology companies or academics, to harness new and emerging technologies for analysing patient data at scale.

27 April 2017
Liz Ford participated in the Farr Institute Public Engagement Consensus Workshop, which aims to produce a consensus statement around best practice for public engagement for research using data linkage and data sharing.

6 April 2017
Liz Ford attended the launch of the Wellcome Trust Initiative 'Understanding Patient Data'.

13-16 March 2017
Farr Institute Future Leaders in Health Data Science – first cohort meeting, Manchester, UK. Liz Ford presented her research in health data science so far, as well as on the health informatics research happening in Brighton and Sussex.

14 November 2016
Healtex Launch Event. Liz Ford was invited to speak on 'Issues of Privacy and Governance when Using Medical Free Text for Research' at the launch event of the UK Healthcare Text Analytics Network, an EPSRC-funded UK-wide network for breaking down barriers to the use of healthcare text data for research.

BACKGROUND IMAGE FOR PANEL

Related initiatives and collaborations

University of Brighton Advanced Data Analytics Research Group

Healtex – UK Healthcare Text Analytics Network

Members of the Farr Institute of Health Informatics Research

Understanding Patient Data – a Wellcome Trust-led initiative to make language around use of patient data easier to understand for the public

Simpatix – a process-enhanced similarity search project by Dr Johannes Starlinger

BACKGROUND IMAGE FOR PANEL

Further grant funding

ASTRODEM is funded by a Wellcome Trust Seed Award (202133/Z/16/Z ASTRODEM – Using Astrophysics to close the “diagnosis gap” for dementia in general practice; £94,000; 1 October 2016-30 September 2018).

Additional funding for this project includes:

Awarded 2 July 2018 to 31 August 2018
University of Sussex Junior Research Associate Bursary: 'Development of an algorithm for detecting unlabelled patients with dementia in UK general practice patient records'. £1,800 awarded to Joanne Sheppard who is currently studying for a degree in Physics, and will work with the Astrodem team over the summer as an intern.

Awarded: 8 January 2018-7 January 2020
Alzheimer's Research UK Prevention and Risk Reduction Fund: 'Predicting risk of dementia using routine primary care records' (£245,000) PI Dr David Reeves, University of Manchester. Liz Ford is co-investigator.

Awarded: 30 June 2017
Santander University of Brighton Travel Grant. “Building synergistic modelling and governance approaches for health data: a collaborative project” £1662. Dr Liz Ford will travel to Berlin to work with Dr Johannes Starlinger at Humboldt-Universität zu Berlin. Using ASTRODEM data, they will work on clustering patients who can be matched on sequences of events occurring over time in their patient history. These clusters can then be used to predict patient outcomes. We hope these longitudinal methods will add value to the ASTRODEM project.

Awarded: 1 July 2016-30 September 2016
University of Sussex Research Development Fund, “A new Bayesian approach to modelling dementia risk using general practice patient records” £6,443 + £4000 matched from schools.