Embedding Data and Clinical Decision-Making Within The Digital Prosthetic Socket Fitting Process
Interview with Joshua Steer of Radii Devices
In this interview, Joshua Steer of Radii Devices discusses the collaboration between Radii Devices and HP to enhance the prosthetic socket fitting process through digital and data-driven methods. Steer outlines how this partnership, initially led by the Office for Veteran Affairs’ Office of Advanced Manufacturing, focuses on integrating data and clinical decision-making to streamline the design and production of 3D-printed prosthetic sockets, and how the use of machine learning in improving fitting processes and shares insights on overcoming challenges in embedding data-driven decision-making in clinical settings.
What led to the collaboration between HP and Radii Devices, and how has this partnership influenced the development of your prosthetic socket fitting process?
The original collaboration came through a project led by the Office for Veteran Affairs’ Office of Advanced Manufacturing. Their aim is to create a process to provide 3DP prosthetic sockets at scale to their Veteran population.
This process aims to ensure that clinicians can focus on their clinical decision making, supported by Radii’s data, whilst the structural design aspects of the device are automated ready for 3DP on HP’s MJF printers.
It was quickly clear that Radii’s and HP’s core capabilities and aims are complimentary, and we’ve been exploring how that can benefit both companies’ user bases.
Your presentation at CDFAM NYC focuses on integrating data and clinical decision-making into the digital prosthetic socket fitting process. How does this digital approach improve upon traditional methods?
Digital has been around for several decades in prosthetics, the aim being that you can remove the need for manual, plaster-based fabrication techniques. However, the rich data that is captured within a digital process has not been well leveraged and analyzed historically.
Working with HP, we can use this data to generate population-based models so that during their design process they quantitatively know the design window for the population, leading to better comfort and socket attachment, as well as better placement of features and metamaterial properties within the socket.
How does Radii Devices leverage machine learning to enhance the fitting process for prosthetics, and what specific data are you using to inform these decisions?
When training our model, we require three key pieces of data: the original shape of the limb, the modified interface of the prosthetic socket, and a whole host of outcome measures. We can then extrapolate industry standard fit modifications and apply them to incoming scans instantaneously vs. requiring fully manual input each time. The clinician still has full control over tuning the outcome, but they have a strong data-backed starting point.
What feedback have you received from clinicians and patients using this new fitting process, and how has it shaped the ongoing development of your technology?
Key feedback has been around adapting the model to the individual clinicians, and their preferences. It only takes a handful of examples of their work to tune the model.
Could you share an example of how your technology has improved outcomes for prosthetic users?
In one of our earliest trials in the UK, a participant described their prosthesis as “the most comfy socket I’ve ever tried”. They are an experienced prosthetic user and to be able to provide them with a socket which they are still comfortably wearing to date was such a great early validation of the work we are doing.
What are the most significant challenges you’ve encountered in embedding data-driven decision-making in clinical settings (let me guess, human reluctance to change), and how have you overcome them?
One of the main challenges has been how to keep the clinician as the driver of the design. Creating a model to predict optimal socket design can be done in several different ways.
The key challenge is how to create a model that the clinician can interact with, understand, and take ownership of. Ultimately, they have the patient in front of them, so they have to be able to adjust the output based upon their expert judgment.
Finally, what key insights or lessons do you hope the audience will take away from your presentation and what do you hope to gain from attending CDFAM NYC?
How data can be used across the provision of these body fitting devices, but that data needs to be tailored for the individual process and person, whether they are the clinician or the designer of the product.
For those who are creating the structural design of the product for 3DP, they need information on the whole population, so they can understand the necessary design window. Whilst for the clinician who is fitting the device to the patient in front of them, the data needs to be presented so they can integrate it within their clinical decision making processes.
