State of the Nation Survey: Data-Driven Health and Care

NHSX, in collaboration with the AHSN AI Initiative and other partners, and supported by experts across the system, are launching another State of the Nation Survey for Data-Driven Health and Care in 2019.

We are aware that there are some truly remarkable data-driven innovations, apps, clinical decision support tools supported by intelligent algorithms being developed, and that electronic health systems are being widely adopted. In parallel, we are seeing advancements in technology and, in particular, artificial intelligence (AI) techniques. Combining these developments with data-sharing across the NHS has the potential to improve diagnosis, treatment, experience of care, efficiency of the system and overall outcomes for the people at the heart of the NHS, public health and the wider health and care system.

Following on from the survey that was held last year to understand what AI technologies were being developed , we want to go one step further and understand where they are being developed, what problems they are solving and collect more tangible information on the data and regulatory landscape.

We hope that through these actions we can make the UK the best place to safely and ethically develop, deploy and use data-driven health and care technology.

You can take the survey here, but for those who want to know what we’re asking but do not want to take the survey, the questions are below.

1.Are you developing the solution or procuring the solution?

2.Which group of health and care system users is your data-driven solution for? Select as many as applicable.

• Person with long-term condition
• Parent/Carer
• Clinician (e.g. decision-support)
• Commissioner/System Manager (e.g. operational efficiency)
• Person with a physical disability
• Person with a cognitive or learning impairment
• Person with broad care needs
• Person interested in monitoring their health (e.g. fitbit)
• Person wishing to access ad-hoc services (e.g. video consultation)
• Person seeking mental health support
• Users for population screening purposes
• Other (please specify)

3. What category of outcome are you expecting to achieve for your identified ‘user’? Please select as many as applicable.

• Improved Quality of LIfe
• Improved independence/autonomy
• System efficiency
• Better experience of health services
• Better experience of care services
• Better access to health service
• Better access to care services
• Prevention of ill-health/improvement of health
• Faster diagnosis
• Faster identification of care need
• Other (please specify)

4. How do you classify your data-driven solution? Select as many as applicable.

• Diagnostic
• Therapeutic
• Population health
• Care-based
• Triage
• Self-care
• Health promotion
• Remote Monitoring
• Remote Consultation
• Other (please specify)

5. At which point of care do you expect your data-driven solution to be deployed? Select as many as applicable

• Primary Care
• Secondary Care
• Community Care
• Tertiary Care
• Individual Care of Self e.g. user’s home/office
• For the purposes of population screening
• Other (please specify)

6. What type of data are you developing/training your data-driven solution on? Select as many as applicable

• Completely synthetic data (e.g. algorithm generates data itself)
• Hybrid data (synthetic dataset modelled on real-world de-identified dataset)
• Personal data
• Non-personal data
• Non-patient data (e.g. data related to estates)
• Other (please specify)

7. Where did you get your training dataset from? (e.g. National data registry, NHS centrally-held data repository, Local Authority, specific hospital, care home, international dataset, other). Select as many as applicable.

• NHS Hospital (Acute) Trust
• NHS Mental Health Trust
• NHS Ambulance Services Trust
• NHS Community Health Trust
• NHS Digital
• NHS CCG
• NHS General Practice
• Care Home
• Private Company
• Charity
• University
• Local Authority
• LHCRE
• Health Data Research UK
• CSU
• Joint Controller
• Patient
• Staff
• Don’t Know
• Other (please specify)

8. How much did it cost you to get that training dataset? Provide a range or state ‘unaware’ if you do not know.

9. Who is the data controller of the training dataset you are using to develop your data-driven solution? Select as many as applicable.

• NHS Hospital (Acute) Trust
• NHS Mental Health Trust
• NHS Ambulance Services Trust
• NHS Community Health Trust
• NHS Digital
• NHS CCG
• NHS General Practice
• Care Home
• Private Company
• Charity
• University
• Local Authority
• LHCRE
• Health Data Research UK
• CSU
• Joint Controller
• Patient
• Staff
• Don’t Know
• Other (please specify)

10. What commercial arrangement do you have in place to access the dataset? Select as many as applicable.

• Consortium agreement
• Equity Share
• Grant Funding Model
• Golden Share
• IP Sharing
• Open-Access
• Profit-Sharing/Revenue-Sharing
• Royalty Payment
• Local Asset-Backed Vehicle
• Cost Recovery
• Don’t know
• Other (please specify)

11. Did you seek ethical approval/advice from any regulation (HRA, MHRA, CQC, ICO) before starting to develop your product or before beginning your research?

• yes
• no

12. Where in the process of medical device CE mark classification is your data-driven solution?

• Already classified
• In process of being classified/intend to seek classification
• Not seeking classification/not applicable to the solution

[13a. You have indicated that your data-driven solution is not classified/you are not seeking classification. What is the primary reason for this? ]

• I was not aware of the process for CE marking as a medical device
• Medical device classification is not applicable to me
• Other (please specify)

[13b. What is the classification you are seeking/intend to have for your data-driven solution?]

• Class I Medical Device
• Class IIa Medical Device
• Class IIb Medical Device
• Class III Medical Device
• IVD Medical Device — List A
• IVD Medical Device — List B
• IVD Medical Device for Self-Testing
• IVD Medical Device under Other/General Device
• IVD Medical Device under Class A under IVDMDR (New)
• IVD Medical Device under Class B under IVDMDR (New)
• IVD Medical Device under Class C under IVDMDR (New)
• IVD Medical Device under Class D under IVDMDR (New)
• Active Implantable Medical Device
• I don’t know

14. Are you developing your product/conducting your research in accordance with the behaviours set out in the Code of Conduct for data-driven health and care technologies?

• Yes
• No
• I am not sure

[15. You have indicated that you are currently not following/unsure whether you are following the behaviours set out in the Code of Conduct for data-driven health and care technologies to guide the development of your product/research process. What is the primary reason for this?]

• I was not aware it existed
• It is not clear how I would implement the behaviours
• My company has its own code of practice that I am following
• It is too prohibitive to the development process
• Complying is too expensive
• Other (please specify)

16. The Code of Conduct is designed to promote ethical and responsible development of data-driven health and care technologies. It is especially concerned with promoting Fairness, Accountability and Transparency. With this in mind, have you:

• Assessed possible issues of bias in your data samples? Yes/No
• [if above = “Yes”] Have you taken measures to rectify such issues in the pre-processing phase?
• Considered whether your algorithmic system is fair and non-discriminatory in its architecture, procedures, and outcomes? Yes/No
• [if above = “Yes”] Have measures been put into place to redress such issues?
• Incorporated the explainability of the system into its design? Yes/No
• Set up procedures to make the rationale of your outputs of your system understandable to all affected stakeholders? Yes/No
• Do you intend to seek separate datasets for training purposes? Yes/No
• Do you intend to seek access to separate datasets for validation purposes? Yes/NO

17. Taking into consideration the need to train, validate, evaluate and seek appropriate regulatory approval, how likely is it that your data-driven solution will be ready for development at scale within the next:

• 5 years (5 point likert scale)
• 3 years (5 point likert scale)
• 1 year (5 point likert scale)

18. Do you have any other thoughts, comments or blockers you would like to suggest to us concerning the Code of Conduct?

These questions help us to understand the make-up of the data-driven health and care ecosystem. The answers will be kept confidential and will not be used to identify you. Please only answer if you feel comfortable.

19. How would you classify your organisation/company? Select as many as applicable

• CCG
• Health or Social Provider
• Higher Education Institute
• Local Government
• Patient Group
• Private Company (not community interest company)
• Community Interest Company/ Social Enterprise
• Other Voluntary or Community Sector
• Mental Health Provider
• Other (please specify)

20. How many employees are in your organisation/company?

• Less than 10
• 10–49
• 50–259
• 250 and over

21. How would you classify your role in your organisation?

• CEO/Director
• Consultant
• Junior/Assistant
• Middle Management
• Senior Management
• Other (please specify)

22. Which statement best describes your current experience with data-driven health and care technologies?

• I use data-driven tech
• I evaluate data-driven tech
• I procure data-driven tech
• I regulate data-driven tech
• I develop data-driven tech
• Other (please specify)