Most digital health companies collect clinical evidence once, to get a foot in the door. Sword Health treated it as a strategy that changed as the company grew. Over ten years it put clinical research to three different uses in sequence: first to overcome clinician skepticism, then to establish digital care as a new standard, and now to prove cost-effectiveness so payers will reimburse it.
That progression was the throughline of “From Hype to Health,” a Castor session with Fernando Correia, Chief Medical Officer at Sword Health and part of its founding team, and Derk Arts, CEO of Castor. The stakes are large: musculoskeletal conditions affect roughly one in two US adults,[1] and low back pain is the leading cause of disability worldwide.[2] It lands now because the funding correction that hit digital health after 2021 raised the bar: the companies that survived have to show clinical outcomes and cost savings, not one or the other.
What the session covered
Fernando laid out the three waves plainly:
“The first thing that we used our clinical trials for was to reduce friction towards adoption. The second wave of clinical trials, the ones where we used Castor, were geared towards defining our solution as a new standard of care. And finally, the third goal is to prove cost effectiveness in order to unblock financial reimbursement.”
Fernando Correia, Sword Health
Underneath that strategy is a deliberate mix of real-world evidence at scale and focused randomized trials. From the start, Sword decided which clinical data to collect prospectively on its members, so small early samples would compound into a large real-world dataset. Sword’s own low back pain cohort now exceeds 200,000 people, which Fernando described as one of the largest real-world samples in digital care. His point was that the dataset is not a byproduct, it is built on purpose:
“This takes years to build. Early on, nothing happens… then all of a sudden you have really, really big samples. But you only have that data if you consciously build that machine.”
Fernando Correia, Sword Health
Building that machine is an organizational problem as much as a data one. Fernando described his job as brokering a standing tension between clinicians who want to measure everything, product teams worried about survey fatigue and dropouts, and a commercial team that needs data to win clients. The resolution he keeps landing on is to start from what a buyer needs to see, then pick the shortest validated scale that still captures it.
The randomized trials came later. Sword’s earliest studies in 2016 ran on pen and paper. Its two pandemic-era RCTs comparing digital and in-person care ran on Castor, and one of them, a chronic low back pain trial with Emory University, was published in npj Digital Medicine in 2023 and reported outcomes comparable to in-person physical therapy.[3] Derk summed up where Castor fits in one line: a platform to generate clinical evidence and run a clinical trial. The 200,000-patient real-world dataset was Sword’s own. The controlled trials behind the published evidence are where Castor sat.
What’s in the full recording
The conversation goes further than the strategy:
- Fernando ranks what clinical research actually delivered, in the order Sword lived it, in answer to an attendee question (46:58).
- The candid mechanics of choosing a patient-reported outcome scale by length, usage, and licensing fees, so it fits the member experience without driving dropouts (48:27).
- Why Sword became a care provider itself after partners treated its technology as merely incremental, and his contrarian case that sometimes you should ship, even in healthcare (38:36).
- Sword’s publication record, an internal count of 40 clinical papers, and, in Fernando’s account, a UCSF study on recovery after ovarian cancer where Sword supplies the digital care arm at cost (26:05).
Watch the full conversation
Fernando and Derk get specific about how to sequence evidence and design trials that hold up in the real world. Watch the session on demand. If you are building the evidence base for a digital health or device program, see how Castor supports clinical trials, real-world evidence, and medical device studies.
Frequently asked questions
What did Sword Health do differently to prove its digital care works?
Sword used clinical evidence as a sequenced strategy rather than a one-time study: early trials to overcome clinician skepticism, randomized trials to establish digital care as a standard, and cost-effectiveness studies to unblock payer reimbursement. It ran this alongside a large real-world evidence dataset built from prospectively collected member data.
How does real-world evidence work alongside randomized trials in digital health?
Real-world evidence compounds over years from data collected prospectively on every member, which is how Sword built a low back pain cohort of more than 200,000 people. Randomized trials run in parallel to test digital care against in-person care directly. Sword ran both, using the real-world data for scale and the trials for controlled comparison.
What was Castor’s role in Sword Health’s clinical evidence?
Castor was the clinical data platform for Sword’s two pandemic-era randomized trials comparing digital and in-person care, one of which was published in npj Digital Medicine in 2023. Sword’s 200,000-patient real-world dataset was collected on Sword’s own systems, not on Castor.
References
- United States Bone and Joint Initiative (BMUS): musculoskeletal conditions affect roughly one in two US adults. https://www.sciencedaily.com/releases/2016/03/160301114116.htm
- Global Burden of Disease 2021 / Lancet Rheumatology (2023): low back pain is the leading cause of years lived with disability worldwide. https://www.healthdata.org/news-events/newsroom/news-releases/lancet-new-study-shows-low-back-pain-leading-cause-disability
- Correia FD, et al. Digital versus conventional rehabilitation for chronic low back pain: randomized controlled trial. npj Digital Medicine 2023;6:121. https://swordhealth.com/articles/nature-study