We’re benchmarking two Bluetooth goniometers against a Vicon system and a manual universal goniometer this month in our outpatient lab. For those who’ve brought them into trials or pragmatic clinic use, what accuracy and reliability targets (e.g., ±3° error, ICC ≥.90) were sufficient to change practice, and did the workflow/data export (REDCap or otherwise) hold up in real sessions?
We switched to digital for knees/elbows once we cleared mean absolute error ≤3° vs Vicon on sagittal moves and ICC ≥.90 across 3 raters; forearm rotation still wandered 5–7°, so we kept the manual there. REDCap export held up via CSV to the API (Resources – REDCap), but iOS Bluetooth drops killed flow (), so we added a 30‑sec 2‑point calibration at session start and locked device names to cut drift. If you’re aiming for “±3°” this month, track plane‑specific MAE, not just overall ICC — are you pushing live to REDCap or batching offline?
In our outpatient pilot, we required MAE ≤3° vs Vicon and inter-session ICC ≥.85, and we did a quick “zero-check” on a 90° jig at start and end — if drift topped 2°, we benched it for the day. Building on @timD_76, forearm rotation stabilized when we cleared nearby metal/magnets and kept the phone in line of sight for Bluetooth; REDCap was fine with CSV + timestamps once we set a simple field map (like telling the goniometer not to sit next to the fridge).