Therapeutic alliance in online therapy: the evidence says it holds — the weak point is somewhere else

What the therapeutic alliance is (and why it matters so much)

The therapeutic alliance — or working alliance, in Bordin's (1979) classic formulation — has three components: agreement on the goals of treatment, agreement on the tasks for reaching them, and the affective bond between client and therapist.

Its relevance is among the most robust findings psychotherapy research has: the quality of the alliance predicts therapeutic outcome consistently, across approaches, populations and diagnoses. Often better than the particular technique used.

Which is why the question of what happens to the alliance in a remote format is not a methodological curiosity. It is the question.

What the evidence says about alliance at a distance

The most cited meta-analysis on this question finds no difference. Seuling and colleagues reviewed 18 publications directly comparing alliance in videoconference and in-person psychotherapy.

This runs in the same direction as the systematic review by Simpson and Reid (2014), covering 23 studies, in which clients rated the alliance at least as highly as in the in-person format, across different diagnostic groups.

With one caveat worth not hiding

Norwood and colleagues (2018) analysed four non-inferiority trials, all of cognitive behavioural therapy, and found a difference unfavourable to videoconference (SMD = −0.30) whose confidence interval crossed the established non-inferiority margin. That is: they could not conclude that the online alliance was equivalent in that particular subset.

It is a result drawn from few studies, tightly bound to one approach, and the later and broader Seuling meta-analysis points elsewhere. But it is there, and an honest article mentions it.

And clinical efficacy follows

Comparative meta-analyses — Carlbring et al. (2018) among others — have been showing equivalent effects between internet-delivered and in-person CBT for anxiety, mild to moderate depression, phobias and stress.

So where is the real weak point?

This is where the conversation gets interesting, because the remote format does have fragilities. Just not the ones we usually name.

1. Adherence, not the bond

What suffers online is not the quality of the relationship but its continuity over time. Dropout rates tend to be higher, especially among people with low initial motivation, limited digital literacy or difficulties with emotional regulation. Cancelling a video call has almost no social cost; cancelling an appointment you were going to travel to does.

An excellent alliance in session 4 is not much use if there is no session 9.

2. The ruptures that go unnoticed

Repairing alliance ruptures is one of the moments with the most therapeutic potential in a process. But to repair a rupture you first have to detect it, and detection rests largely on subtle signals: posture, a change of pace, a gaze that drifts away. On screen you have a framed face and little else.

The risk is not that ruptures are more frequent remotely. It is that you find out later.

3. The setting stops being under your control

In the consulting room you design the setting: space, privacy, freedom from interruption. Remotely, the client sets it, and sometimes they set it in the car, with a flatmate on the other side of the wall. That directly conditions what can be spoken about.

4. Risk management and crises

This is the most serious limitation and the one that least tolerates improvisation. Faced with decompensation, unstabilised suicidal ideation, severe dissociation or active violence in the home, physical distance is a real constraint. It calls for a protocol in advance, not a reaction. We work through it in the guide to online therapy.

5. The gap between sessions

And this one is, in our experience, the most underestimated. In person, the therapeutic process is held up by an infrastructure of context: the same room, the same journey, the ritual of arriving. Online, all that scaffolding disappears and the weight of continuity falls entirely on the clinical record and on the therapist's memory.

When someone comes back after six weeks and you have had twenty sessions that week, all over video, recovering exactly where you left off — what hypotheses you were holding, what task you agreed, which sentence broke in them last time — stops being an act of memory and becomes a systems problem.

Five practices for holding the alliance remotely

1. An explicit digital setting in the first session. Camera on, a space with privacy, what happens if the connection drops, the cancellation policy. Turning the implicit into an agreement reduces friction later.
2. Measure the alliance instead of sensing it. Brief instruments such as the WAI-SR or the Session Rating Scale take a minute at the end of a session and replace an impression with a number. Remotely, where non-verbal signals arrive clipped, that is worth double.
3. Say out loud what the room would have shown you. "You seem quieter today than last week — does that land?" Explicit verbalisation makes up for part of the lost channel, and it is an intervention in itself.
4. Have a written risk protocol before you need it. The client's physical location in each session, an emergency contact, the resources in their area, criteria for referring to in-person care. Written, not memorised.
5. Armour the continuity between sessions. A clinical record that gives you back the thread in thirty seconds is worth more, in a remote format, than any additional technique.

What clinical software can do (and what it cannot)

Let us be precise, because it is easy to over-promise here: no software creates a therapeutic alliance. You do that, with your listening and your judgement.

What a good clinical tool can do is take away what erodes it. If the evidence says the bond survives distance but continuity suffers, the ground where technology contributes something is exactly that:

• Longitudinal clinical memory: arriving at the session knowing where you left off, without reconstructing it at the last minute.
• Structured, consistent notes: so the record is comparable across sessions and lets you see the arc, not just the episode.
• Less administrative load after the session: every hour you do not spend documenting is an hour of attention you can give to the person in front of you.
• Tracking indicators, alliance scores included if you collect them, so a downward trend shows up sooner.

Alchely is built on that idea, and on the conviction that administrative load is not a minor inconvenience of clinical practice but one of the things that wears therapists down most. With servers in the European Union, no model training on client data and a formal data processing agreement, because in therapy trust is not negotiable at the technical layer either.

Frequently asked questions

Sources

• Seuling, P. D., Fendel, J. C., Spille, L., Göritz, A. S. and Schmidt, S. (2024). Therapeutic alliance in videoconferencing psychotherapy compared to psychotherapy in person: A systematic review and meta-analysis. Journal of Telemedicine and Telecare, 30(10), 1521–1531.
• Simpson, S. G. and Reid, C. L. (2014). Therapeutic alliance in videoconferencing psychotherapy: A review. Australian Journal of Rural Health.
• Norwood, C., Moghaddam, N. G., Malins, S. and Sabin-Farrell, R. (2018). Working alliance and outcome effectiveness in videoconferencing psychotherapy: A systematic review and non-inferiority meta-analysis. Clinical Psychology & Psychotherapy.
• Carlbring, P., Andersson, G., Cuijpers, P., Riper, H. and Hedman-Lagerlöf, E. (2018). Internet-based vs. face-to-face cognitive behavior therapy: A systematic review and meta-analysis. Cognitive Behaviour Therapy.
• Bordin, E. S. (1979). The generalizability of the psychoanalytic concept of the working alliance. Psychotherapy: Theory, Research & Practice.