Choose a Teletherapy Platform on Session Mechanics, Not the Demo

Every teletherapy platform looks fine in the sales demo, because a sales demo is the easiest session the platform will ever carry: two adults, two good computers, office bandwidth, and nobody trying to grab the mouse. Your Tuesday afternoon looks different. It is a four-year-old with apraxia joining from a parent’s phone, an adult post-stroke client whose spouse handles the technology, and a home exercise review where you need to watch someone move through a full squat without the video smearing. The platforms that survive that afternoon are not reliably the ones with the longest feature lists, and the difference does not show up in a demo — it shows up in a trial you design to be hard. This article is that trial: the compliance gate to clear first, the session mechanics to test by discipline, the low-bandwidth behavior to provoke on purpose, and the scheduling integration that quietly decides your no-show rate.

Before any demo, settle the compliance question, because it eliminates candidates faster than any feature comparison. Under HIPAA, a video platform vendor that creates, receives, or transmits protected health information on your behalf is a business associate, and you need a signed business associate agreement (BAA) with that vendor before it carries a session. The pandemic-era shortcut is gone: the HHS Office for Civil Rights announced that its COVID-19 telehealth enforcement discretion — the policy that let providers use everyday consumer video apps in good faith without a BAA — expired at 11:59 p.m. on May 11, 2023, and the 90-day transition period it granted ended August 9, 2023. Since that date, the ordinary HIPAA rules apply to telehealth in full.

Two practical consequences follow. First, the consumer video apps many practices adopted in 2020 were only ever permissible under that now-expired discretion; a vendor that will not sign a BAA is out of consideration regardless of its video quality or its price, and that includes free tiers of paid products, which often exclude the BAA that the paid tier includes. Second, a vendor’s claim of being “HIPAA compliant” is marketing language, not a verdict: OCR has been explicit that it has not reviewed the BAAs vendors offer and that no vendor list constitutes an endorsement. The document to get is the vendor’s actual BAA, read before signature — what it covers, which products it covers, and whether your subscription tier is inside it.

General-purpose video tools were built for conversations between adults who want to be in the meeting. A therapy session is a different activity: it has materials, movement, a client who may be four years old, and a clinical need to see and hear things a meeting never requires. This is where platforms genuinely differ, and it is where your evaluation should spend most of its time.

For speech-language pathology, the session lives and dies on audio fidelity and interaction. Aggressive noise suppression tuned for offices can filter exactly the speech signal you are listening for — a distorted /r/, a weak plosive, breathiness — so you need to test whether the platform lets you adjust or disable audio processing. Interaction matters just as much: a young child cannot sustain attention on a static screen share, so look for whether the client can click, drag, and draw on what you share — remote control, annotation, or a true shared activity space — and whether you can revoke that control instantly when a five-year-old discovers the scribble tool.

For occupational and physical therapy, the camera problem dominates. The client is often across the room, on the floor, or mid-transfer, which means you need the video to stay legible when the subject fills the frame and moves quickly — and you need to direct a caregiver to reposition the camera without breaking the session. Test whether video quality holds when someone walks away from the laptop, whether you can pin and enlarge the client’s video, and whether the platform stays usable when the client’s device is a phone propped against a water bottle. For pediatric work in any discipline, add the caregiver dimension: can a parent join the same session from a second device, and can you bring them in for the last five minutes without a new link?

Teletherapy candidates almost always fall into one of three classes: a general-purpose video tool used under a BAA, a teletherapy-specific platform built around therapy activities, or the telehealth module inside your EHR or practice management system. None of the three wins outright — each is strong exactly where the others are weak — so the useful move is not asking which class is best but running the same tests against a finalist from each. The table below is that test plan. Every cell is something to verify yourself in a trial, not a fact to assume about the class.

Read your results by caseload, not by total score. A practice that is mostly adult neuro rehab can live with a thin activity library but cannot live with fragile video on movement; a pediatric SLP practice is the reverse. The rows where your caseload cannot tolerate a failure are gates, like the BAA — a candidate that fails one is out, however well it does elsewhere.

A trial where you open the platform, admire the interface, and close it again tells you nothing. Recruit a colleague or family member as your test client, put them on the worst realistic device — a mid-range phone, not your spare laptop — and run the three sessions below before you look at pricing. The whole protocol fits in an afternoon.

1. 01

   Set up like a family, not like a clinician

   Send your test client the session link the way a family would get it, and have them join from a phone using only what the link provides. Count every step between tapping the link and seeing your face: downloads, account creation, permission prompts, waiting rooms. Each step is a place a real family fails at 3:58 for a 4:00 session — and the count is a number you can compare across platforms.

2. 02

   Run a pediatric interactive session

   Share a real activity you use — not the platform’s demo content — and hand your test client control. Have them click, drag, and draw the way a distractible child would, then revoke control mid-scribble. Watch for the lag between their action and what you see; more than a beat of delay kills turn-taking activities in real sessions.

3. 03

   Run a movement session from across the room

   Have your test client prop the phone up, step back until their whole body is in frame, and move — squat, transfer, floor-to-stand. Direct a camera repositioning by voice. Then run the same session again with them on a cellular hotspot instead of wifi, and note what degrades first: frame rate, resolution, or audio. That failure order is what your clinical eye has to work around all year.

The most common teletherapy failure is not a frozen frame — it is a family at the appointment time with no idea where the link is. That is a scheduling problem, and it is why the integration between the video platform and the system that runs your schedule belongs in the platform decision rather than after it. Trace the full lifecycle of one recurring appointment: where the session link is created, whether the reminder that goes to the family contains it, what happens to the link when the appointment is rescheduled, and what a family does when they lose the email. If the answer to any of these is “someone at the front desk fixes it manually,” you have found a recurring cost that will outweigh most feature differences.

Recurring caseloads sharpen the requirement. A weekly therapy client is not one appointment but forty, and platforms differ on whether a recurring series carries one stable link or generates a new one per visit, whether a reschedule updates the link in the family’s reminder automatically, and whether a clinician’s whole day is joinable from one place or from ten separate calendar entries. This is also the strongest practical argument for the EHR-integrated class: when the schedule itself generates the session link and puts it in the reminder, an entire category of missed sessions disappears — provided the video holds up under the tests above, which is exactly why the video tests come first.

The platform is necessary but not sufficient — three adjacent questions decide whether the session it carries is one you may conduct and bill. First, licensure follows the client. ASHA’s guidance reflects the general rule across the therapy professions: it is the client’s location during the session that determines where you must be licensed, and you should verify requirements in both your state and the client’s state before initiating services. The interstate licensure compacts — the PT Compact, the Occupational Therapy Licensure Compact, and the Audiology and Speech-Language Pathology Interstate Compact (ASLP-IC) — are changing this picture by letting eligible clinicians practice across member states through compact privileges, but membership and implementation status differ by state and profession, so check the compact’s own site for where things stand before you count on it.

Second, consent. Informed consent for telepractice should be documented and should cover the services being provided and the technology used to provide them — and several states attach their own telehealth-specific consent requirements on top. Where the rules of your state and the client’s state differ, ASHA’s practical advice is to follow the more restrictive one. Build the consent into your intake paperwork once, rather than improvising it per family.

Third, coverage. Medicare’s authority for PT, OT, and SLP telehealth services has run on statutory deadlines since the pandemic, lapsing and being restored as Congress acts; under the Consolidated Appropriations Act, 2026, CMS states that therapists may furnish Medicare telehealth services through December 31, 2027. Commercial payers and Medicaid programs each set their own telehealth rules. None of this affects which platform is best, but it decides whether the sessions are reimbursable — so verify your payer mix against current policy before committing to a teletherapy service line, not after.