Chat-based support reduced smoking relapse
Adding mobile chat counseling to usual care improved 6-month biochemically confirmed abstinence in recent quitters compared with low-intensity text messages.
*Randomized clinical trial; Level 1b (OCEBM).

Citation

Luk TT, Su X, Wong V, et al. Mobile Chat Messaging for Smoking Relapse Prevention: A Randomized Clinical Trial. JAMA Internal Medicine. 2026;186(3):294-302. doi:10.1001/jamainternmed.2025.7439.

Background

Most people who quit smoking return to smoking, often within the first weeks. Scalable relapse-prevention support is limited, and mobile messaging could extend help beyond clinic visits.

Patients

590 adults in Hong Kong cessation services who had stopped smoking for 3–30 days and previously smoked daily. Excluded: participation in other cessation studies, diagnosed mental health condition, or regular use of psychiatric medicines; required a smartphone with the messaging app installed.

Intervention

Usual cessation treatment plus 3 months of chat-based relapse support: live counselor messaging (work hours) and access to a supportive chatbot.

Control

Usual cessation treatment plus 8 generic text messages over 3 months.

Outcome

Primary: biochemically validated abstinence at 6 months (exhaled carbon monoxide <5 parts per million or negative saliva cotinine test). Secondary: self-reported prolonged abstinence, 7-day abstinence, and relapse (7 consecutive smoking days) at 6 months.

Follow-up Period

6 months.

Results

6-month outcome Intervention Control Relative effect (95% CI) NNT
Biochemically validated abstinence (primary) 45.9% (135/294) 35.5% (105/296) Relative risk 1.29 (1.06–1.58) 10
Self-reported prolonged abstinence 57.5% (169/294) 47.6% (141/296) Relative risk 1.21 (1.03–1.41) 10
Self-reported 7-day abstinence 65.6% (193/294) 54.7% (162/296) Relative risk 1.20 (1.05–1.37) 9
Relapse 33.0% (97/294) 44.9% (133/296) Relative risk 0.73 (0.60–0.90) 9
Analyses were intention-to-treat, with missing outcomes counted as relapse.

Limitations

Participants and counselors could not be blinded. Results may not generalize well beyond mostly male, clinic-linked recent quitters, or to people with mental health conditions (excluded). A local tobacco tax increase may have raised quit rates in both groups. One counselor delivered the intervention, and tests may miss exclusive use of some non-smoked nicotine products.

Funding

Hong Kong government Health and Medical Research Fund; no funder role reported.

Clinical Application

Add counselor-guided chat relapse support for recent quitters already in cessation care; expect about 1 extra confirmed quitter per 10 treated versus minimal texting.