Skip to content
If you are in crisis in the United States, call or text 988. This site cannot help with an emergency.
A Feel Good Thing

Research

Messaging therapy vs weekly live video: what the clinical trials show

Facts checked November 14, 2026. Operated by Alpha Origin LLC. Not medical advice.

Editorial illustration of one person texting on a phone and another on a laptop video call with a counselor. Decorative only—not staff or patients of this site.

Two different products, one label

“Online therapy” covers at least three formats, and the research on one does not automatically carry over to another. Asynchronous messaging means you write when you want to, and the therapist replies on an agreed schedule. Live video is a booked session, usually 30 to 45 minutes. Live chat is a booked session done by typing in real time.

The platforms on this site mix these formats differently. Talkspace sells a messaging-only cash tier at $69 a week. Calmerry has a messaging promo at about $49.50 a week. BetterHelp’s typical plan is one live session a week plus messaging. Online-Therapy.com Standard is one 45-minute live session a week. Prices are from each company’s pricing page as of October 15, 2026. So the useful question is not “does online therapy work?” It is whether a plan that is mostly messages works as well as one built around a weekly video session.

This page summarizes published trials. It is not medical advice, and it does not tell you which format is right for you. If you are in crisis in the United States, call or text 988. Messaging therapy is not crisis care.

Video has the deepest evidence

Fernandez and colleagues pooled the video literature in a 2021 meta-analysis in Clinical Psychology & Psychotherapy. It covered 56 within-group studies (1,681 participants) and 47 studies comparing video with something else (3,564 participants). Video therapy did far better than a wait list. Its results differed negligibly from in-person therapy. The effect was largest for CBT aimed at anxiety, depression, and PTSD.

Guided internet CBT is a separate thing again: structured written lessons with a therapist checking in. Carlbring and colleagues reviewed 20 trials (1,418 participants) that compared it directly with face-to-face CBT. The pooled difference was close to zero (Hedges g = 0.05). The authors added that few individual conditions had been tested this way. That is evidence for structured, written CBT with a therapist involved. It is not evidence for open-ended chat with no structure.

Typed CBT, delivered live

The best-known trial of typed therapy is older. Kessler and colleagues, in The Lancet in 2009, randomized 297 primary-care patients with depression in England. One group got usual care plus CBT typed live with a therapist. The other got usual care while on a waiting list. At four months, 38% of the typed-CBT group had recovered, against 24% of controls. At eight months it was 42% against 26%.

Note the comparison group: a waiting list, not video and not an office. The trial shows that typed CBT works better than waiting. It does not show that typing equals talking.

Head to head: asynchronous messages vs weekly video

Three randomized trials have compared the two formats directly. All three ran on Talkspace’s platform, were led by researchers at the University of Washington or UC Berkeley, and list Talkspace employees among the co-authors. Keep that in mind, and also notice that the results point the same way.

  • Song et al., Journal of Medical Internet Research, 2023. 83 adults with at least moderate depression symptoms (PHQ-9 of 10 or higher) got either scheduled messaging or a 45-minute video session each week. The messaging group improved substantially in depression, anxiety, and day-to-day functioning. Those gains were noninferior to video. Ratings of the therapist relationship and engagement did not differ significantly.
  • Areán et al., Psychiatric Services, 2024. 215 adults with depression. The messaging group stayed engaged with their therapist longer: 7.8 weeks on average, against 4.9 for video. There were no meaningful differences in depression, anxiety, or disability at 6 or 12 weeks, and no adverse events in either group.
  • Pullmann et al., JAMA Network Open, 2025. 850 adults with depression; average starting PHQ-9 was 15. At 12 weeks, 47.5% in messaging and 47.2% in video had responded to treatment. Remission was 31.4% and 30.3%. By week 5, more people had dropped out of video (21.3%) than messaging (13.2%). Among people who were not improving, video built a stronger therapist relationship early. Participants also recommended video more often (97.2% against 87.5%).

How to read the numbers

Three terms carry most of the weight in these papers. Noninferior means the new format did not do meaningfully worse than the comparison, within a margin the researchers set in advance. It is a narrower claim than “better” or even “identical.”

Response and remission are thresholds on a symptom scale. In depression research, response usually means the score fell by at least half. Remission usually means it dropped into the minimal range, which on the PHQ-9 is under 5. When about 47% respond and about 31% remit in both arms, as in the 2025 trial, both formats helped roughly half the people. Neither helped everyone.

Dropout is an outcome too. A format people quit early cannot work for them, however good it looks on paper. Two of the three trials found people stayed with messaging longer.

What the trials do not tell you

  • They studied adults with depression over about 12 weeks. They do not cover anxiety disorders as a main condition, trauma-focused therapy, couples work, eating disorders, or severe illness.
  • Each trial set the messaging schedule in advance. A retail plan where replies come “within one business day” may be a looser product than what was tested.
  • Outcomes were mostly self-reported symptom scales such as the PHQ-9. That is the standard measure, but it is a questionnaire, not a clinician’s diagnosis.
  • All three were run on one company’s platform, with its therapists and its software. Results on another platform may differ.
  • Nothing here supports messaging for someone who may be at risk of self-harm. Crisis support means a phone call or an in-person service, starting with 988 in the U.S.

How to use this when you pick a plan

  • Messaging is a reasonable first choice if scheduling is what keeps you out of therapy, you write easily, and you do your thinking between sessions. In the largest trial, people stayed with it at least as well as with video.
  • Video makes more sense if the relationship with the therapist is what you are paying for. In the 2025 trial, people who were not improving formed a stronger early bond on video. Video also fits structured CBT exercises where real-time back-and-forth matters.
  • You can combine them. In the JAMA trial, people still not responding at week 6 were reassigned to messaging plus video, either weekly or monthly. Adding weekly video did no better than adding monthly video. Several platforms sell a messaging plan with a monthly live session.
  • Set a checkpoint. Note how you feel at the start, then check again around week 6. If nothing has moved, change the format or the therapist rather than renewing on autopilot.
  • Check the money side. Messaging is usually the cheaper cash tier, but it is the hardest format to claim out of network. The superbill guide explains why.

Next: Live video and messaging by platform · What each format costs per month · Superbills and out-of-network claims

Related providers

Overviews stay on this site. “View provider” leaves to the company’s public site (or an affiliate path only when a tracking URL is installed).

Sources

See also How we compare and Compensation disclosure. For crisis help in the United States, call or text 988.