Call me out of the loop, but I’d never heard anyone refer to a “digital twin” before today. I’m glad stumbled across the concept, though, as there’s usually something interesting lurking in the unfamiliar. This case was no exception.
In the broadest sense, the term digital twin applies to a digital replica of physical assets, processes, people, places, systems and devices, according to the gods of Wikipedia. The same Wikipedia article notes that digital twins can be integrated with artificial intelligence, machine learning and software analytics with spatial network graphs to produce living digital simulations that update and change as the physical twin changes.
By this point, you probably see where I’m headed with this. One logical extension of the digital twin approach is to create a digital patient which can be populated with the data generated by a flesh-and-blood person. Rather than relying solely on existing data, digital patients absorb patient data in as close to real time as possible.
You’re also probably way ahead of me when it comes to how that digital patient will work, but I’ll spell it out anyway. According to an article by Henk van Houten, CTO of Royal Philips, a digital patient is “a lifelong, integrated, personalized model of the patient that is updated with each measurement, scan or exam, and that includes behavioral and genetic data as well.” This definition is as clear as any I’ve seen out there, so I’ll roll with it.
According to van Houten, a digital patient offers advantages above and beyond a static digital record and it “Integrates and analyzes every bit of information – like a smart assistant that accompanies patients and their caregivers along the patient journey,” he writes. “As it is updated over time, it provides intelligent advice to support medical decisions and to help patients manage their disease.”
Under van Houten’s version of the digital patient model, primary care physicians will have more control of their patients’ health, in part by having physiological measurements automatically integrated into the patient’s information, along with AI-generated care recommendations. “The strength of the digital twin paradigm is that it combines scientifically-proven knowledge with biophysical modeling insights derived from data,” van Houten says. This will involve a robust remote monitoring program which collects and shares ambulatory data flows, but these days that’s certainly possible.
Of course, the accuracy of these recommendations will depend greatly on how good the patient-generated health data is, but that’s another story. It won’t help much if the patient is measuring them sporadically using a device that is not very accurate. Not only that, if we don’t get better at interoperability and mining unstructured data, some if not all bets are off.
It’s also worth pointing out that few healthcare organizations are ready for more data than they already have in-house, making the idea of building a sophisticated digital patient somewhat remote at present.
Even so, it seems certain that over time we’re going to be integrating all the patient’s data we can get our hands on into their existing records. At least in theory, the digital patient model seems like an elegant and powerful way to understand individual patients more thoroughly, and that’s always a good thing.