Abstract
Introduction: Mindfulness-based programs (MBPs) are commonly organized around eight weeks, yet direct experimental evidence on the consequences of varying this architecture is sparse. In a contextual inventory of 39 established programs, 11 of the 15 with an unqualified fixed-week structure used eight weeks.Methods: This rapid review examined direct evidence on assigned program length, session intensity, pacing, outcome and mechanism trajectories, and postprogram maintenance. APA PsycINFO was searched through Ovid from inception to September 3, 2026. The review used one database and one reviewer, and no protocol was prospectively registered. All 301 records were screened and 81 full texts were assessed. Risk of bias was appraised with tools matched to study design, and certainty was assessed for eight synthesis claims using an adapted Grading of Recommendations Assessment, Development and Evaluation approach.Results: Fifty-two reports representing 47 studies were included. Five reports directly compared assigned architectures that varied different dimensions; the remainder mainly described trajectories under one assigned program. Randomized length comparisons did not demonstrate reliable between-length differences, but none was designed to establish equivalence. In the largest controlled trajectory, Mindfulness-Based Stress Reduction first showed statistically detected separation from usual care at session 2 for sleep, 4 for physical function, 6 for pain intensity, 7 for pain interference, and 8 for depression; these were power-dependent statistical crossing points, not minimum effective durations. The session of first detected separation also differed across mechanisms, mechanism-outcome coupling changed across sessions, and postprogram maintenance varied by outcome. No included estimand was judged at low overall risk of bias, and of the eight synthesis claims, two were of low certainty and six of very low certainty.Discussion: The literature supports neither universal abbreviation nor preservation of eight weeks as an empirically optimized standard. The main conceptual advance is to separate evidence about change during a program from evidence about the consequences of changing the program: calendar length, contact, spacing, outcome timing, mechanism timing, individual heterogeneity, and durability are distinct design questions. Well-powered randomized architecture comparisons with aligned repeated measurement and prespecified noninferiority or optimization criteria are needed.