Telehealth vs In-Person AI Scribing | Scriber Wiki
Last updated
AI scribing works for both modalities, but the workflow and audio quality differ in ways that affect transcription speed and note accuracy.
In telehealth sessions (especially Doxy.me with Scriber's direct integration), audio flows directly from the telehealth platform to the transcription engine, eliminating intermediaries and cutting transcription latency to ~3 seconds. In-person sessions require a microphone or room recorder, which introduces placement and background-noise variables that can add 1–2 seconds to transcription time and occasionally require a retranscription pass.
The real difference is logistical, not clinical. Understanding where each modality excels helps you set realistic expectations and prevent the most common frustration: expecting telehealth-speed transcription from in-person sessions.
Telehealth Scribing (Doxy.me Direct Integration)
When you use Scriber's direct Doxy.me integration, the telehealth platform handles audio capture natively. You start your session in Doxy, enable Scriber from within your Doxy room, and Scriber receives the audio stream directly — no separate app, no recording device, no manual setup.
Advantages:
- 3-second transcription window. xAI processes the audio while you're reviewing the session transcript on-screen. By the time the patient leaves the call, the transcript is usually complete.
- Minimal setup. Click one button; Scriber starts. No fumbling with room mics or separate devices.
- Speaker clarity. Telehealth audio is typically cleaner than room-recorded in-person audio because both sides are coming through a microphone designed for voice.
- Back-to-back scheduling. You can finish notes between sessions because transcription is near-instantaneous.
Considerations:
- Bandwidth dependency. Scriber requires stable internet to stream audio. On a shaky connection, the transcription may pause or retry. This is rare, but it's a single point of failure that in-person recording doesn't have.
- Patient audio only. Telehealth captures the patient's voice clearly; your own voice is captured through your microphone. Sometimes this creates a slight tonal imbalance, but clinical content is captured equally.
- Patient notification. Because Scriber integrates directly with Doxy, a "Scriber is recording" indicator appears in the patient's room. You don't need a separate recording device notification, but the fact of recording is explicit.
Most clinicians using Doxy + Scriber start 2–3 minutes before the scheduled session time to confirm audio levels and connectivity. That's it.
In-Person Scribing
In-person sessions require a microphone or array to capture audio in a physical space. You have several options: a single condenser mic on your desk, a lavalier mic clipped to your collar, a room microphone mounted on the wall, or a multi-mic array.
Advantages:
- No connectivity requirement. Audio is recorded locally to Scriber's app on your device. If the internet drops, the recording continues.
- Flexibility. You can move around, step away, or change room setup without affecting the audio stream.
- Patient agency. Some clinicians prefer having a separate, visible recording device that the patient can see (or not see) rather than a software integration that feels less tangible.
Considerations:
- Audio quality variability. Room acoustics, background noise (traffic, HVAC, building sounds), and mic placement all affect transcription quality. A well-placed condenser mic yields audio as clean as telehealth; a phone mic on a desk can introduce echo and ambient noise.
- Transcription latency. xAI usually completes in 5–10 seconds; Speechmatics in 8–15 seconds. This is fast enough to finalize notes between sessions, but it's not "ready while the patient walks out" like telehealth.
- Equipment setup. You need to position a microphone, test it, and sometimes adjust placement if audio is poor. This is a one-time setup but requires initial troubleshooting.
For in-person recording, the best practice is a dedicated desk microphone (like a Blue Yeti or similar) positioned 2–3 feet away, angled toward both you and the patient. Lavalier mics work well if you move around during sessions. Phone or computer-built-in mics introduce too much room noise and echo.
Noise and Accuracy
Both modalities handle background noise similarly: transcription engines are trained to filter it out, but heavy background noise (traffic, loud HVAC) can reduce accuracy by 2–5%. The difference is that telehealth noise (Zoom beep, notification) is usually speaker-isolated, while in-person noise is omnidirectional.
If your in-person space is noisy (open office, shared clinic space), consider:
- A noise-canceling headset with a built-in mic (your own voice is captured clearly; room noise is suppressed)
- Positioning the patient closer to the microphone
- Scheduling sessions during quieter hours when possible
Hybrid: When You Mix Both
Some clinicians see some patients in-person and some on telehealth, sometimes switching modalities mid-week. This works fine. Your muscle memory for reviewing notes stays consistent; the only variable is transcription latency (telehealth ~3s, in-person ~8–12s).
One caution: if you're switching a regular telehealth patient to in-person, test your in-person setup with that patient first. Explain that you're trying a new recording method and want to ensure audio is clear. Most clinicians do this for 1–2 sessions before confirming the setup works.
Practical Workflow Differences
Telehealth (Doxy integration):
- Start session in Doxy, enable Scriber, session proceeds normally.
- Transcription complete by end of session.
- Note ready for review immediately after disconnect.
- Finish note in 2–3 minutes; move to next patient.
In-person:
- Mic is already positioned; start Scriber app before patient arrives.
- Session proceeds; app records continuously.
- Transcription begins when you stop recording (usually immediately after patient leaves).
- Note is ready 10–20 seconds later.
- Finish note in 3–5 minutes; schedule allows a longer buffer between patients if needed.
Neither is objectively better — they're logistics, not clinical quality. The deciding factor is usually your existing platform (if you're already on Doxy, the integration is a no-brainer; if you're in-person, a desk mic is all you need).
Related questions
See it in action with Scriber
Start with a 7-day trial — HIPAA-compliant, no audio retention, BAA included at signup.