FHT 5th Dec Meeting Outputs

1. Presentation slides & videos

2. themes discussed at the event

3. Summary of low hanging fruits

4. collectively Mapping future directions

5. how can i get involved with the fht consortium?

Presentation slides & videos

opening remarks

1. “Introduction: framing the context and purpose.” Speaker: Becky Inkster. Presentation slides.

2. “What are the approaches being used for financial data?” Speaker: Gareth McNab, Nationwide, Co-Leader of Open Banking for Good. Presentation slides with notes coming soon.

3. “What are the approaches used for health data?” Speaker: Stan Shepherd, CEO, Instant Access Medical. Presentation slides coming soon.

LIGHTNING TALKS

1. Monzo & Lyrical Kombat: Building financial literacy. Presentation slides with audio.

2. Toucan: Alliance banking (link to Report)

3. Ducit.ai: Deep financial assessment with Christians Against Poverty and Community Money Advice (slides coming soon)

4. Wysa: AI-enabled mental health chatbot exploring financial health. Video here.

themes discussed at the event

Everyone is potentially vulnerable and each person has unique complex needs:

• There was a recognition that everyone faces different issues and complexities and that we are all possibly vulnerable at different points in our lives;

• There was a strong sense that by knowing more about a person holistically it can allow us to see a fuller picture and also step away from treating individuals based on population measures.

Emphasis on consumer-driven / user-centred approaches:

• We need to design solutions that put the power back into consumers hands;

• How do we create value (benefit) for the consumer/user (e.g., ‘their data their benefit’)?

• Importance of involving users from the very onset of design.

Striving for standards & recognising specific responsibilities, roles & offerings:

• With a myriad of different tools are we creating problem we don’t need? See low hanging fruits section (below).

• Rather than unlocking data as a first step, can we unlock insights derived from data in order to create standardised practices across banks and create a new dialogue without needing to ‘look under the hood’?

• What was very apparent was that we must recognise that financial services, institutions, organisations, Fintech start-ups, digital health start-ups and academics etc. have different types and amounts of data that they can generate or can access (e.g., anonymous versus identifiable; finance or health; inside or outside conventional settings etc.) hence differing responsibilities and they will only be able to contribute toward addressing certain problems in certain ways (i.e., we need to understand the specific purposes and limitations of different types of data).

• We must avoid ‘solutions looking for problem’ style approaches

Discussion on the role of data science: opportunities and cautionary notes

• There is a huge gap in the evidence-based literature on the convergence of health and finance data, which data science could help address.

• How do globally trusted brands protect consumers and maintain trust, but balance possibilities of unlocking financial data for social good?

• “The quality of debt advice can be seen as poor; it’s a random process” and so how can data-driven approaches support human decision-making in offline health-relevant settings in real-time? How can we measure health outcomes in Fintech A/B style scenarios? and vice versa (as will be discussed below).

• The importance of knowing the limits of the appropriateness of using data: ‘just because you can it doesn’t mean you should’ (e.g., lend more money to cat owners, really?!?! stay away from crude proxy measures; huge push for advanced methodologies)

• Discussion on data sharing and ownership: data trusts, 4th parties/safe havens like the Turing, commercial incentives to make that easy, sharing through the consumer (consumer pivot points) etc. Noting that “…finding that trusted body is a real challenge, regulators under funded etc”.

• Tension about ‘looking under the hood’ if you are sitting on a lot of potentially revealing data (issues of safeguarding, duty of care etc.). Mixed messages from FCA, ICO, tech and healthcare (i.e., protect vulnerable consumers; move slowly and cautiously; move fast and break things; first do no harm).

• A suggested focus on health outcomes that only use financial data and don’t mix with health records prior to deep linkage/cross-phenotyping.

• Examining data donation options; Biobanks, longitudinal cohorts (e.g. ALSPAC) and analysing data with health and finance conversational pathway information from forums (Mumsnet)/social networks (Talklife)/chatbot (Wysa) (see low hanging fruits section).

Additional issues raised:

• Can data-driven insights help us prioritise different times of need? How might we consider redistributing resources according to intensity/severity of need (shifting away from routine appointments toward priority in-person crisis support?)

• How do we handle issue of actual versus potential detriment?

• How do we improve the publics understanding of the links and issues?

• How do we increase consumer trust?

• How do we build a sense of trustworthiness?

• What are the risks and benefits of creating converging finance health solutions that are detaching from formal healthcare systems? What are the dangers as we shift toward unchartered healthcare and vulnerability support?

summary of low hanging fruits

1. CapitalOne: Building a universal open source lexicon library platform to raise standards for vulnerable disclosures

   • CapitalOne announced a commitment to improve standardisation across financial services by creating an open source platform-based lexicon dictionary of terms to support vulnerable users through voice analytics.

   • Each bank shouldn’t have to re-invent the wheel in codifying lexicon dictionaries (i.e., working in silos)

   • This goal is good because its easy to see the benefit from multiple stakeholder perspectives; “Vulnerability is a non-competitive issue” (Tim Hawley). We need standards and consistency, the same treatment for consumers across services (one strategy may independently probably get you 80% of the way there to solving the problem, but by joining forces you can go that final mile to best serving consumers).

   • A founding principle: “you must put more in than you take out”

   • Could this library act as a way to encourage trust and openness amongst financial services? Would the notion of banking services working together to raise standards increase trust for consumers?

   • Action: how can we (FHTC) help move this cause forward alongside CapitalOne and partners?

2. Gambling blocks: scientifically evaluating the different approaches used by different financial services

   • The FinHealthTech Consortium will support developments being made in this space: can we evaluate and pool data-derived insights from each financial service that uses gambling blocks with the aim of identifying a common standardised approach to roll out the block across services. Different banks have different parameters for their blocks and it would be important to understand what is the 'best' approach to use for gambling blocks etc (e.g., why is 72 hours selected as a duration of blocking for one bank but this differs for another).

   • This consortium will strive to introduce scientific rigour into discussions that link with ongoing work in this space. There are currently a number of projects which are looking at the role that financial services firms can play in tackling problem gambling. Researchers at the University of Bristol, for example, are currently working with Gamble Aware on a three-year programme exploring how financial firms can reduce gambling-related harm. The first project within the programme considers the potential of spending controls to allow customers to block gambling transactions, working with financial firms to explore the data they currently hold in this area. A number of other complementary projects are also ongoing, including work by the Behavioural Insights Team, Money and Mental Health Policy Institute and GamCare.

3. Evaluating existing data sets containing conversational data about mental health and money (longitudinal cohorts, peer-peer mental health social network data, AI-driven chatbot, forum for maternal health etc). There are opportunities to develop research questions and possibly be involved with data analysis and publication.

collectively Mapping future directions

More details will be coming soon, but these are some initial DOTS & directions

1. Toucan – more details coming

2. Ducit.ai – Can we link this into clinical frameworks? E.g., Henrietta Bowden-Jones, Director, National Problem Gambling Clinic and National Centre for Gaming Disorders.

3. Jessica Lennard – Visa transactional data and vulnerability, links with Turing, DSSG etc

4. Monzo & LYRICAL KOMBAT – building financial literacy within hip-hop culture.

5. Alan Turley – How can we tackle unclaimed benefits (20b) – more details coming

7. More great work coming out of Nationwide Open Banking for Good (OB4G) – watch this space

8. Sebastian Vollmer, Turing Institute Data Science for Social Good (DSSG) programme opportunities – What are the lowest barriers to accessing data combined with organisations that can take action? DSSG as a Turing-based sandbox environment for data analyses looking for new challenges.

9. Stan Shepherd – Instant Access Medical are committed to working with partners on their platform to combine personalised insights across mental health, physical health, financial health.

10. Jonathan Dawson – Can a financial assessment be built into clinical evaluations? Would it be possible, for patients who wish to consent, for their doctor to send a notifications to banks about a patient’s diagnosis to close the loop and provide instant tailored financial services to their circumstances? Can we create a financial assessment tool designed for doctors and patients when health issues arise and financial circumstances are high risk. Can we use something similar to Ducit.ai at the front lines of clinical medicine, health data working inside NHS? If doctors were able to notify financial services these financial services wouldn’t have to make the inferences through voice analytics etc.

11. Squanderlust – Can we shift the focus around the idea of crisis points rather than routine appointments in order to relieve administrative burdens?

how can i get involved with the fhT consortium?

There are many ways to get involved: whether it be to stay up-to-date with the latest news, research and innovations, or perhaps to enquire about becoming a co-author for a chapter for the FHT book, or maybe wanting to be a speaker for the DIMH2020 FHT summer programme, maybe you’d like to offer your data skills or data sets for consideration…. or other ways, let us know…