Peer-supported online therapy modestly reduces anxiety

Peer-supported online cognitive behavioral therapy modestly improved anxiety after low-risk emergency chest pain.
*Randomized clinical trial; Level 2 (OCEBM).

Citation

Musey PI Jr, Kroenke K, Connors JN, et al. JAMA Internal Medicine. Published online September 28, 2026. doi:10.1001/jamainternmed.2026.4225

Background

Most emergency chest pain is not due to a heart attack, but anxiety is common, often missed, and may drive repeat visits. This trial tested practical anxiety treatments started after emergency department discharge.

Patients

375 adults discharged from 6 emergency departments with low-risk chest pain and at least moderate anxiety. Exclusions included high heart-risk scores, hospital admission, active psychosis or suicidal thoughts, unstable vital signs, homelessness, incarceration, out-of-town residence, and inability to speak English.

Intervention

Peer-supported online cognitive behavioral therapy, or therapist-delivered cognitive behavioral therapy by telehealth. All groups received anxiety education.

Control

Recommended primary care follow-up plus anxiety education.

Outcome

Primary: anxiety score using Generalized Anxiety Disorder-7 (0 to 21; higher is worse). Secondary: global anxiety improvement, depression, physical symptoms, and disability.

Follow-up Period

12 months.

Results

Outcome Significant result
Anxiety score, online therapy vs primary care referral (primary) 1.22-point greater improvement (95% CI 0.01 to 2.43)
Severe baseline anxiety subgroup Online therapy improved 2.8 points more than primary care referral (95% CI 1.0 to 4.6)
Any patient-reported anxiety improvement Online therapy: OR 3.1 (95% CI 1.7 to 5.8), NNT 6; therapist telehealth: OR 2.8 (95% CI 1.6 to 5.0), NNT 6
Moderate or much better anxiety Online therapy: OR 2.4 (95% CI 1.4 to 4.1), NNT 5; therapist telehealth: OR 2.3 (95% CI 1.4 to 3.8), NNT 5

OR = odds ratio; NNT = number needed to treat; CI = confidence interval.

No significant between-group differences were found for depression, physical symptoms, or disability. Analyses were intention-to-treat among participants with at least 1 follow-up assessment. The usual minimally important anxiety-score difference is about 2 to 3 points, so the overall average benefit was small; the severe-anxiety subgroup benefit was more clinically meaningful. The control was enhanced usual care, not no treatment.

Limitations

About one-third of therapy participants completed no sessions, and 20% had no follow-up assessment. The study occurred within one health system, limiting generalizability. The average anxiety benefit was statistically significant but below the usual threshold for a meaningful patient-perceived difference.

Funding

Patient-Centered Outcomes Research Institute; no funder role reported.

Clinical Application

Offer peer-supported online therapy mainly to emergency chest-pain patients with prominent anxiety, especially severe anxiety; it is feasible but not practice-changing for all.