Waitlist Controls in Anxiety Trials: Why 'Better Than Nothing' Is a Low Bar

When a new therapy for anxiety is tested, it needs a comparison group. The simplest version of that comparison is a waitlist: one group receives the therapy now, the other waits and receives it later. If the therapy group improves more, the treatment “works.” This design is common, practical and deeply limited.

Why waitlist controls are popular

Psychotherapy research faces a genuine ethical constraint: randomly assigning anxious people to receive nothing at all is harder to justify than in drug trials, where the placebo group still receives a pill and clinical contact. A waitlist solves this by promising treatment to everyone — just not yet. Ethics committees approve it more readily, and recruitment is easier because participants know they will eventually get the intervention.

The result is that a large proportion of the evidence base for anxiety treatments rests on waitlist comparisons. A meta-analytic update of CBT for anxiety and depression found that the majority of included trials used waitlist or no-treatment controls [1].

How the design inflates effects

The core problem is that a waitlist control does not control for much. In a well-designed drug trial, the placebo group receives the same clinical contact, the same monitoring, the same expectation of possible improvement. The only thing withheld is the active ingredient. A waitlist withholds everything: the contact, the structure, the expectation and the ingredient.

This matters because the effect sizes look different depending on the comparator. When CBT for anxiety and depression was compared against waitlist, the pooled effect sizes were roughly twice as large as when the same therapies were compared against pill placebo or psychological placebo [1]. The treatment didn’t change — the yardstick did.

The nocebo problem

There is a subtler issue. Being told you are on a waitlist is not a neutral experience. You know you are not receiving help. You may feel disappointed, discouraged or stuck.

A network meta-analysis of psychotherapy trials for depression found evidence consistent with a nocebo interpretation: people on waitlists appeared to do worse than people receiving no treatment at all in a design where they didn’t know treatment was available [4]. If waiting for promised help is itself demoralizing, the waitlist group’s deterioration inflates the apparent benefit of treatment.

An exploratory randomized trial tested this directly. Participants were randomly assigned to receive an intervention, wait for the intervention, or receive assessment only (no mention of a future intervention). The waitlist group showed less improvement than the assessment-only group, consistent with a suppression effect [3]. The study was small and in a different population, but the direction aligns with the nocebo hypothesis.

What waitlist comparisons can and cannot tell you

They can show: a treatment is better than nothing over a defined period. This is useful early in a treatment’s development, when you need to know whether it has any effect at all.

They cannot show:

• Whether the treatment works because of its specific ingredients or because of non-specific factors like attention and expectation
• How the treatment compares with existing established treatments
• The real size of the specific treatment effect

A decision framework for choosing control conditions in behavioral trials recommends matching the control condition to the question being asked: waitlist for “does it do anything?”, attention control for “is it the specific content that matters?”, and active comparison for “is it better than what we already have?” [5].

Better designs and their costs

Active controls compare a new treatment with an established one. This answers the question patients actually care about, but requires larger samples (because you’re looking for a potentially small difference between two effective treatments) and costs more.

Psychological placebos provide time, attention and a credible rationale, but without the theorized active ingredients. Designing a credible psychological placebo that participants can’t distinguish from real therapy is genuinely difficult, and whether any psychological control is truly “inert” is debatable [2].

Dismantling designs take a treatment apart to see which components are necessary. For instance, testing full CBT versus CBT without the cognitive component. A systematic review of dismantling trials in mental health found that complete packages usually outperformed components, but often by small margins [6].

Each design answers a different question. The most informative evidence comes from a progression: initial waitlist-controlled trials, followed by active-controlled trials once basic efficacy is established.

How to read a waitlist-controlled result

When you encounter a study reporting that a treatment “works” based on a waitlist comparison:

1. Note the comparator. If it is waitlist-only, the effect size is likely inflated.
2. Ask whether there are also trials against active comparisons. If not, the evidence is early-stage.
3. Look at whether the treatment group improved or the waitlist group worsened, or both.
4. Check whether the study measured expectation or credibility, which would help separate specific from non-specific effects.

For how effect sizes differ across comparators and what that means practically, see understanding effect sizes.

What remains open

Whether waitlist controls systematically bias the literature is increasingly acknowledged, but the field has not converged on what should replace them as the default. Active comparisons are expensive. Psychological placebos are hard to design. And for digital interventions, which now make up a growing share of trials, a waitlist may be even more problematic, since the “doing nothing” condition is particularly stark. The question is not whether waitlist-controlled trials are bad science — they answer a real question — but whether we overweight them in treatment guidelines.

When to look for more support. If you are waiting for treatment and your symptoms are getting worse, don’t wait for a scheduled start date to seek help. Talk to your GP or call a helpline. In the US, call or text 988 for the Suicide and Crisis Lifeline. In the UK and Ireland, Samaritans are on 116 123.