CBT for Anxiety Over Video: What Happens to Thought Records, Experiments, and Homework
Last reviewed: 09/08/2026
Reviewed by: Dr. Kiesa Kelly
Most articles about online therapy answer a question you have probably already settled. Does it work? Broadly, yes. The harder question, and the one people sit with before a first appointment, is narrower: what happens to the work of CBT when it moves onto a screen?
CBT is not a conversation. It is a set of tools: you catch a thought and test it, you run small experiments to find out whether the feared thing happens, and you practice between sessions where the anxiety lives. Those tools are the treatment — so the useful question is what they look like on a screen. Our guide to how many CBT sessions anxiety usually takes covers the shape of a course; this one covers the mechanics.
In this article, you'll learn:
Why the delivery format matters less than the tools themselves
What changes about thought records and worksheets on video
How behavioral experiments and between-session practice work remotely
What your clinician can see less of over video, stated plainly
How to decide whether in-person is the better fit for you
The short answer: the format is not what carries the work
A 2024 systematic review pooled 54 randomized trials and 5,463 patients comparing therapist-guided remote CBT with in-person CBT. It found moderate-certainty evidence of little to no difference on primary outcomes, with a standardized mean difference of −0.02 (95% CI −0.11 to 0.07) across 51 trials [1]. That held regardless of condition, length of follow-up, and whether therapy was individual or group.
Two caveats belong with that number, because they are usually dropped. "Remote" pooled telephone, computerized programs, internet modules and videoconference, and only a minority were live video. In 29 of the 54 trials the therapist supported modules people worked through alone. So it tells you remote CBT is broadly comparable, not that video matches an office.
For that, the closest evidence is a multisite trial of 148 adults with generalized anxiety disorder who received a 15-session protocol either by videoconference or face to face. Video was statistically non-inferior at the end of treatment and again at 6 and 12 months [2]. What matters is which factor predicted who improved. Change in intolerance of uncertainty — the mechanism the therapy targets — carried a standardized beta of 0.56 (p < 0.001), while assignment to video or to the office did not significantly predict outcome at all (p = 0.118) [3].
That is the argument in one result: what predicted improvement was the mechanism, not the room. If you are weighing CBT for anxiety in Nashville against a telehealth appointment, that finding should carry the most weight.
Key takeaway: 🧰 CBT gets its results from what you practice, not from where the session happens.
Three things people get told about online CBT that aren't quite right
"The connection with your therapist won't be as strong." The evidence does not support this as stated. A 2024 meta-analysis of 18 studies comparing working alliance in videoconference versus in-person psychotherapy found no statistically significant difference in ratings by patients (SMD −0.09, 95% CI −0.26 to 0.07) or therapists (SMD 0.04, 95% CI −0.17 to 0.25) [4]. Two qualifications: a non-significant result is not proof of equivalence, and that interval still leaves room for a small disadvantage.
"Online means an app or a self-guided program." These get flattened together constantly. A self-guided digital program and a live session with a clinician who knows your history are not the same treatment, and much of what is reported as "online CBT" research is the former [1]. This article is about the latter.
"You can't do the real techniques remotely." Exposure work and behavioral experiments are routinely built into video protocols, and we cover the format question for exposure work specifically elsewhere. The accurate version of this concern is not that techniques cannot be delivered — it is that your clinician sees less of the room you are in.
What a thought record looks like on a screen
A thought record is the workhorse of CBT for anxiety. You capture the situation, the thought that arrived, what you felt and how strongly, the evidence for and against, then a more balanced read. In an office it happens on paper while your clinician writes.
On video it happens one of three ways: your clinician screen-shares a worksheet you can both see; you work in a shared document; or you each keep your own copy and compare. Ask which before session one — it determines what you need open.
Worth being straight about: we found no study directly comparing a screen-shared thought record with a paper one. What follows is clinical observation from doing this work both ways, offered as that and not as evidence.
What tends to go better on a screen is that you keep the record. In an office the worksheet stays in a folder, or goes home in a bag and is never seen again. A shared document is still there on Tuesday at 11 p.m. when the same thought arrives, and over a course of therapy you can scroll back and see the same distortion in four different weeks. What needs watching is that typing is slower than talking, and a session can quietly become a documentation exercise.
Key takeaway: 📝 Ask before session one how worksheets will be shared and where they will live.
Behavioral experiments when your therapist isn't in the room
A behavioral experiment is not exposure, though the two are related. You make a specific prediction — "if I ask the question in the meeting, people will think I'm slow and I won't recover" — then find out. The point is to collect data on a belief. Our checklist on preparing for CBT for anxiety covers the earlier steps.
They are built into structured video CBT protocols. In a 2025 randomized trial of videoconference CBT for people whose panic disorder remained symptomatic after medication, behavioral experiments were run during sessions 7 through 10 and homework was emailed after every session; that group improved substantially more than those continuing usual care alone [5]. The comparison was against usual care, not in-person CBT, and only 30 people enrolled.
The underrated advantage is that your clinician is looking at the environment where your anxiety actually operates, so an experiment can be designed against the real thing rather than a description recalled in an office ten miles away.
Consider what that looks like. You have been avoiding your work inbox first thing, telling yourself you will get to it once you feel steadier, and by ten o'clock there are forty messages and the dread has its own momentum. In an office you would describe this. On video your clinician can ask you to open it now, and you both watch what your anxiety predicted against what arrives. You predicted a catastrophe. What arrives is two invoices and a scheduling question.
Or: you stopped driving on the interstate after a panic attack on the ramp two years ago, and you have built a route that adds nineteen minutes to every trip. The experiment is not to get on the interstate this week. It is to sit in your parked car, notice what your body does at the thought of the on-ramp, rate it, and find out whether the number climbs forever or peaks and comes down.
Key takeaway: 🔬 An experiment tests a prediction — and at home it gets tested where it lives.
Homework is where much of the change happens
Between-session practice is not a bolt-on in CBT, and it is the part video changes least in importance and most in logistics.
A meta-analysis restricted to controlled studies contrasting the same therapy with and without homework found a pooled effect of d = 0.48 favoring homework [6]. A separate meta-analysis of 17 CBT studies with 2,312 clients found a relationship between how much homework people completed and how they did at post-treatment (quantity: Hedges' g = 0.79, 95% CI 0.57 to 1.02, across 15 studies) [7]. That second set is correlational — completing more homework tends to go with doing better, which is not the same as causing it. No study we could find isolates whether homework matters more in telehealth.
What changes on video is the handoff. There is no moment at the door where a sheet of paper is put in your hand. Practice gets emailed or lives in a shared document, so it cannot be lost — and can sit unread alongside everything else.
For scale: the UK's NICE guideline says high-intensity CBT for generalized anxiety disorder should usually consist of 12 to 15 weekly sessions, each lasting an hour — fewer if the person recovers sooner, more if clinically required [8]. The other several hundred waking hours are where the practice happens.
What your clinician can see less of
This part is usually skipped and deserves saying directly.
In a survey of 113 clinicians who specialize in OCD, therapists rated how well they could detect various things over video compared with in person. The gaps were substantial and consistent: environmental distractions, non-verbal communication, fidgeting, self-reassurance seeking, and cognitive avoidance were all rated markedly harder to pick up on camera. The same clinicians rated telehealth as progressively less feasible as symptom severity rose [9].
Read that carefully. It measured what clinicians believed about their own perception, not whether patients got better. And it studied OCD, not generalized anxiety, panic, or social anxiety — extending it across anxiety conditions is my clinical inference, not the study's finding.
Even so it points at something actionable. A camera frames your face, not your hands or the person in the next room. Subtle safety behaviors are exactly what a CBT clinician watches for, because they are how anxiety maintains itself while treatment appears to progress. The response is not to abandon video but to narrate: saying "I've been picking at my thumbnail for five minutes" gives your clinician what the camera did not.
Key takeaway: 👀 Video narrows what your clinician can observe. Saying out loud what they would have seen closes most of that gap.
What genuinely needs more planning at your end
Two things move from your clinician's responsibility to a shared one.
A private space that is actually private
APA's telepsychology guidelines, approved in August 2024, encourage psychologists to assess the remote environment for its effect on effectiveness, privacy and safety — including the home context, potential distractions, and risks of privacy breaches [10]. A private and quiet environment is listed among the responsibilities discussed with you at consent.
This is the most common thing that goes wrong, and rarely dramatically. It goes wrong as a slow narrowing: you do not say the thing you meant to say, because someone might be in the hallway. A car in a parking lot with headphones works, as does an hour when the house is empty.
The plan for a hard session
Anxiety work can bring up more than you expected, and sometimes the honest answer at the end of a session is that you are not steady yet. The 2024 APA guidance encourages psychologists to build a safety plan with you before services begin, to hold your location, phone number and an emergency contact, and to plan for technological disruption — continuing by phone, for example [10].
These are professional guidelines rather than law, written as encouragements rather than requirements. But they describe what careful telehealth practice looks like, and if none of it has come up by the end of your first session, it is fair to ask. A self-check such as the GAD-7 is worth bringing to that conversation.
Key takeaway: 🗺️ Ask what happens if a session ends badly or the connection drops. A prepared clinician answers immediately.
When in-person earns the drive
A heuristic you can apply now.
If your main obstacles are logistical — time, distance, childcare, an inconsistent schedule, the activation cost of getting out of the door — video is very likely the better choice. The obstacle that most often ends a course of therapy before it starts is logistical, not clinical. On attrition, a meta-analysis of 20 randomized trials found no significant difference between teletherapy and in-person therapy overall, and the video-versus-in-person comparison did not reach significance either (OR 1.25, 95% CI 0.97 to 1.61) [11].
If your anxiety is severe, if you have significant dissociation, if avoidance has been hard to pin down before, or if a previous course stalled without anyone being able to say why, weigh the in-person option more heavily. That is clinical judgment rather than a study finding, but it follows from the observational gap above. Because low mood and anxiety travel together often enough to change a plan, the PHQ-9 is worth completing alongside an anxiety screener.
If you are already several sessions in and nothing has shifted, the format is probably not the variable to change first; we wrote separately about what a plateau in anxiety therapy means. Many people end up doing both, starting on video and coming in for particular pieces of work.
Key takeaway: ⚖️ Logistics point toward video. Severity, dissociation, and hard-to-detect avoidance point toward some in-person work.
Questions worth asking before you start
These can be answered on a first call, and the answers tell you how the work will run.
How will we handle worksheets and thought records — screen share, shared document, or my own copy? You are asking whether the tools have a home.
What does between-session practice look like, and how does it reach me? Vague answers often mean homework is treated as optional, which the homework evidence argues against.
What is your plan if I become very distressed, or if we lose the connection? A clinician practicing carefully will have an answer ready.
How will you know if I'm using safety behaviors you can't see on camera? This tests whether they have thought about the limits of the format.
What is your training in CBT for anxiety, and how will we track whether it is working? Structured measurement should be part of the answer.
Screeners are a starting point for that conversation rather than diagnostic tests; our other screening tools cover the rest.
Your next step
The format question is reasonably settled: therapist-guided remote CBT performs comparably to in-person CBT, and in the closest trial we have on generalized anxiety, what predicted improvement was change in the mechanism the therapy targets — not which arm someone was assigned to. That should change the question you carry into a first appointment. Not "will this work on a screen," but "what will we actually be doing, and how will I know it is working?" Those answers are how you tell a well-run course of CBT from a series of conversations, in either format.
Anxiety running the show?
Evidence-based therapy can turn the volume down on anxiety — a clinician can help you find the approach that fits your life rather than a one-size-fits-all plan.
Frequently Asked Questions
How does a therapist share a thought record during an online cbt session?
Clinicians generally do one of three things: screen-share a worksheet, use a shared document you can both type into, or ask you to keep your own copy while they fill one in alongside you. No research has compared screen-shared and paper thought records, so this reflects common practice rather than trial evidence. Ask your clinician which they use before session one, because it changes what you need open on your device.
Can behavioral experiments for anxiety be done over video?
Yes — they are built into structured video CBT protocols. In a small 30-participant randomized trial of videoconference CBT for people whose panic disorder was still symptomatic after medication, behavioral experiments were run during sessions 7 to 10 and homework was emailed after every session. That trial compared video CBT against continued usual care, not against in-person CBT, so it shows the method can be delivered and can help.
What happens if I get distressed during a telehealth therapy session?
This is the part worth settling before you need it. Current APA telepsychology guidance encourages psychologists to build a safety plan with you before services begin, and to hold your location, phone number and an emergency contact. It also encourages a plan for a dropped connection, such as continuing by phone. If your clinician has not raised any of this by the end of the first session, ask.
Does CBT homework matter more when therapy is online?
Homework matters in CBT generally, and we are not aware of a study that isolates whether it matters more on video. A meta-analysis of controlled trials comparing the same therapy with and without homework found a pooled effect of d = 0.48 favoring homework. Separate analyses linking how much homework people complete to how they do are correlational, not causal. What video changes is logistics, not importance: the practice happens in the same rooms your anxiety happens in.
Can my therapist see less about me over video?
Some things, yes, and it is worth naming rather than pretending otherwise. In a survey of 113 OCD specialists, clinicians rated themselves markedly less able to notice environmental distractions, non-verbal cues, fidgeting and self-reassurance over video than in person. That was clinician-rated confidence in an OCD sample, not a measure of how patients actually did. The practical response is to say out loud what a clinician in the room might have seen.
About the Author
Dr. Kiesa Kelly is a licensed clinical psychologist with more than 20 years of experience in psychological assessment and evidence-based treatment. Her work centers on cognitive behavioral approaches for anxiety and related conditions, and on the practical question of how structured treatment is delivered well — including by telehealth, which our practice uses as a primary delivery model rather than as a substitute arrangement. Our specialized therapy services are described in more detail on the practice site.
References
1. Zandieh S, Abdollahzadeh SM, Sadeghirad B, et al. Therapist-guided remote versus in-person cognitive behavioural therapy: a systematic review and meta-analysis of randomized controlled trials. CMAJ. 2024;196(10):E327–E340. https://doi.org/10.1503/cmaj.230274
2. Bouchard S, Dugas MJ, Belleville G, et al. A multisite non-inferiority randomized controlled trial of the efficacy of cognitive-behavior therapy for generalized anxiety disorder delivered by videoconference. J Clin Med. 2022;11(19):5924. https://doi.org/10.3390/jcm11195924
3. Bouchard S, Dugas MJ, Belleville G, et al. Correction: A multisite non-inferiority randomized controlled trial of the efficacy of cognitive-behavior therapy for generalized anxiety disorder delivered by videoconference. J Clin Med. 2025;14(1):226. https://doi.org/10.3390/jcm14010226
4. Seuling PD, Fendel JC, Spille L, Göritz AS, Schmidt S. Therapeutic alliance in videoconferencing psychotherapy compared to psychotherapy in person: a systematic review and meta-analysis. J Telemed Telecare. 2024;30(10):1521–1531. https://doi.org/10.1177/1357633X231161774
5. Seki Y, Takemura R, Sutoh C, et al. Videoconference-delivered cognitive behavioral therapy in patients with symptomatic panic disorder following primary pharmacotherapy: a randomized, assessor-blinded, controlled trial. BMC Psychiatry. 2025;25(1):861. https://doi.org/10.1186/s12888-025-07320-2
6. Kazantzis N, Whittington CJ, Dattilio FM. Meta-analysis of homework effects in cognitive and behavioral therapy: a replication and extension. Clin Psychol Sci Pract. 2010;17(2):144–156. https://doi.org/10.1111/j.1468-2850.2010.01204.x
7. Kazantzis N, Whittington C, Zelencich L, Kyrios M, Norton PJ, Hofmann SG. Quantity and quality of homework compliance: a meta-analysis of relations with outcome in cognitive behavior therapy. Behav Ther. 2016;47(5):755–772. https://doi.org/10.1016/j.beth.2016.05.002
8. National Institute for Health and Care Excellence. Generalised anxiety disorder and panic disorder in adults: management. NICE guideline CG113. Recommendation 1.2.19. https://www.nice.org.uk/guidance/cg113
9. Wiese AD, Drummond KN, Fuselier MN, et al. Provider perceptions of telehealth and in-person exposure and response prevention for obsessive-compulsive disorder. Psychiatry Res. 2022;313:114610. https://doi.org/10.1016/j.psychres.2022.114610
10. American Psychological Association. APA guidelines for the practice of telepsychology. Approved by the APA Council of Representatives, August 2024. https://www.apa.org/about/policy/telepsychology-revisions
11. Lin T, Heckman TG, Anderson T. The efficacy of synchronous teletherapy versus in-person therapy: a meta-analysis of randomized clinical trials. Clin Psychol Sci Pract. 2022;29(2):167–178. https://doi.org/10.1037/cps0000056
Disclaimer
This article is for informational purposes only and is not a substitute for individualized clinical advice, diagnosis, or treatment. Reading it does not create a clinician–patient relationship. If you are concerned about anxiety or another mental health condition, speak with a licensed clinician about your specific situation. If you are in crisis or thinking about harming yourself, call or text 988 to reach the Suicide and Crisis Lifeline, or call 911.