Distributing data skills in medicineHealthcare is a team sportService Overview1. Conditioning 2. Warm-ups: 3. Drills: 4. Strengthening: 5. Cool Down: A common languageLearning by doingImplementation
Distributing data skills in medicine
Technology is increasingly at the heart of healthcare and its future. Data is opening up new opportunities for improved care through life-changing science, monitoring devices, better diagnostics, and optimised clinical pathways.
These developments create new challenges for medical professionals. In addition to the essential knowledge and experience of medical practice, new skills are required to understand, analyze, interpret, and utilize data sets, as well as to respond critically and creatively to new advancements.
In our research into current medical training, it is clear that formal training is not keeping up with rapidly changing technological advancements, leaving young professionals lacking in these essential skills. In addition, developing greater multi-disciplinary collaboration across medical teams will support the use of data and improve effectiveness.
The big question:
How might we provide every qualified medical professional with the critical, technical, creative and centered skills to understand and work with data as part of their future practice?
Healthcare is a team sport
Every person involved in delivering care has an important role to play, a specific area of expertise, and their collective skills are the secret sauce for innovating in care. Still, and especially when working with technology, professionals are siloed and not on the same page when collaborating due to a lack of a common language.
We should not expect all physicians to become data scientists, the same way you shouldn’t expect your goalie to be a great striker. However, you should expect them to excel at their position while still understanding how to work together with other players towards your goal.
To have bright players working together in a brilliant team, you need an effective training programme that leverages their strengths and gets them ready for the uncertain realities of the playing field. With that in mind, we created Datagym.
Healthcare is a team sport, and Datagym is the coach
Datagym is an experiential educational platform that teaches clinicians and technologists how to use data as a building block for improvement and innovation.
It provides support and a safe space for learning, collaboration, and experimentation through real world projects in clinical organizations.
Datagym has five stages: Conditioning, Warm-ups, Drills, Strengthening, and Cool-down
Ensures a common baseline of technical knowledge through an online foundations course and a data skills self-assessment to match multidisciplinary teams together according to their complementary skills.
Online foundations course
The online foundational course, works to ensure everyone has a common understanding of terminology, concepts, and limitations around different technologies.
Why online courses?
- This concept was gladly taken from MIT’s critical data hackathons in which participants had better outcomes once they were required to take a free online course before attending the event.
- The method of delivery for courses should be kept online, to enable self-paced learning, and also reduces overhead cost, by recycling already existing courses — whether from MOOCs, or from the many university lectures that moved online due to covid.
Self-Assessment matching with multi-disciplinary teams
We provide self assessments to match people to teams that have skills that complement each other’s, as it can be hard to identify who has which skills in the working environment
Teaches learners how to identify opportunities and problem areas from within their own context through a guided design thinking process — specifically a discovery phase — guaranteeing a holistic understanding of people, systems, and processes.
Why design thinking?
Design methods excel at finding the right problems to solve & communicating effectively across disciplines
Clinicians prefer following frameworks, and design thinking is already well established.
It is crucial to have challenges that are not too broad or too narrow, with specific deadlines and goals to reach
“Doctors don’t deal well with uncertainty, they need a framework” Dr. B.K., Medical School Professor, design thinking in healthcare advocate.
How would Datagym facilitate the design process? Via a Challenge Shaping Consultancy
The challenge shaping consultancy works as a hybrid of an educational offering and a design studio; providing training, process set up, recruitment and facilitation throughout the process. This is done as a way to ensure learners make the most out of their time.
- The facilitation aspect is also important as our research shows that when clinicians partake in a design process, they feel the most uncertainty when defining a problem (or writing a how might we), and benefit from others with more experience
Equips students to generate and communicate solutions through ideating and low-fi prototyping and identify the most feasible solutions to develop.
Over a weekend, all teams and relevant stakeholders meet in person along with experts to not just define their problems but also to ideate and create rough prototypes to be developed further.
Having all stakeholders and key decision makers in the room: the magic behind GV Design Sprints
- Decisions are made much faster when all involved can bounce ideas off of each other
- By condensing this process, learners avoid realizing late in the development process that their idea might not be as feasible as they expected
Encourages students to bring projects as close to reality as possible through iterating via an agile development framework with the support of a coach.
Peers also have check ins with other groups to communicate in progress work and learn from each other
Why agile development
- Agile is the new norm for the NHS, as their systems are incredibly complex and require iteration as unexpected issues along the way will arise.
Development should be safe to fail, not fail-safe.
“When systems are so complex, development needs to be iterative and safe to fail” T.F. Medical School Professor, & NHS Digital
Coaching (and guidance)
Responsible for running weekly standups, and solving issues that may arise
Coaches work on a freelance basis, and have experience working with the topic being explored.
Peer to peer learning: the village of mentors
- Learners from different teams will be connected to each other via a slack channel, to ask questions quickly and encourage asynchronous collaboration
- Monthly show and tells work to teach students to share in progress work with others to receive feedback and learn from each other's experiences
5. Cool Down:
Participants document and share their learnings while building on the knowledge of others through presentations, case study writing, or even publishing code on GitHub.
Successful projects can apply for internal funding and further development if the teams are interested.
The importance of documentation and sharing
- It is a form of making collaboration work beyond the time the project had been worked on. It also ensures accountability and reproducibility in projects, as peers can review your work.
Documenting how you handled data ensures others can use the same code or APIs for their own projects. It is also important for the maintenance of the project itself, avoiding work being done twice.
- Even documenting how you have cleaned up data can have massive impact, as others might be facing the same challenges.
Analogue: Open Source Software
- OS projects have a much larger community maintaining and improving the code over time
The open source community has been one of the most important sources of innovation in tech, with much proprietary software using parts of open source software at their core
- Example: Much of AI development we see in use is based on TensorFlow, an open source platform developed initially by Google, supported by a large community of developers
Analogue: The Turing Way
- Initiative by the Alan Turing Institute to ensure that Data Science projects and research are reproducible and ethical.
- The manual grows over time based on community support, and is completely open sourced.
- By opening up access, projects can be reviewed by a global network and improved upon over time
Why document failures?
- Other healthcare institutions may have similar challenges they might face and try the same approach, which is bound to fail. When failures are documented and shared, others can avoid the same mistakes.
- It is also an important practice for understanding what might have gone wrong and what could be improved upon.
Why not make all projects live?
- The reality is that most projects won't be production ready and will need further testing and clinical validation
Datagym aims to provide a low pressure, safe space for experimentation
- The goal is to open up learners' minds to possibilities; to change the way they approach digital transformation in the long run
Clinicians in the NHS are already under immense pressure, and before they start live projects they must build their skills to avoid wasting resources
“I've worked with lots of clinicians and non clinicians who were trying to do digital within their organizations, [...] they are very much thrown in at the deep end, it’s like the blind leading the blind. They have no idea how to solve the problems, So they are very much learning on the job, making lots of mistakes, growing from the mistakes, and getting gradually moving to the next step. [...] And there are no formal methods, learning methodology or framework or anything” — Y.K., Head of Innovation at a large healthcare agency
A common language
Being on the same page is crucial for working across disciplines and a critical issue in digital healthcare.
We have observed that frequently clinicians and technologists have different words for the same thing. For example, collaboration means working with people for clinicians, while for developers, it means publishing code on GitHub.
We teach our learners the terminology, basic theory and use cases for current technologies and their limitations through our online foundational courses, but that is only one part of the equation.
Design Thinking completes the puzzle. Through workshops and sprints, we ensure all learners have a shared understanding of the right problem to solve while taking into account all the people and moving parts that might be involved in the process.
Learning by doing
Learning is done best by getting your hands dirty. Many of the advancements we see are only reproducible in research labs, or are just marketing malarkey. Still, the reality of digital transformation is one of complex systems with multiple moving parts — reality is not a controlled experiment, and technology can often fail.
When pushing for innovation or improvement, professionals need to equip themselves with a growth mindset; to be comfortable with being uncomfortable and stay calm in uncertainty.
Datagym learners develop their ideas further into MVPs through an agile framework with the support of their peers and a mentor, all whilst being in a safe space to fail and experiment.
Students learn to think in an iterative mindset of continuous improvement and the value of sharing unfinished ideas and failures with their peers through weekly stand-ups and occasional show-and-tells.
Learning by doing also strengthens collaboration and communications skills in learners; something you cannot learn by reading a book, but can practice and improve upon
Why real-world projects in multi-disciplinary teams?
Parallel world: Hackathons
- Hackathons have the goal to solve open problems together within a set period of time
- Great for teaching collaboration; also keeps the stakes low, which relieves pressure from doing live projects.
- Of course just putting people in a room and hoping for magic to happen does not work — hackathons rarely have long term effects.
Datagym provides the right framework to make sure learners make the best out of their time, much like many fellowship programmes
Datagym is provided as a B2B service for large hospitals, clinics and trusts to train staff while working on issues their institutions face.
Not only would these organizations benefit from capacity building, home-grown solutions to local problems and better collaborations, but they also prove themselves to be the ideal point of intervention: Their staff already have the many permissions necessary for handling health data, an in-depth understanding of the challenges of their practice; Their staff are also under pressure to up-skill but may not have the time to take a post-graduate course.
Healthcare is a challenging field to innovate in. For the program to be successful, it needs to be associated with a large institution, ideally a teaching hospital or university; brand recognition is crucial for educational initiatives in healthcare.