Host review edition · October 16, 2026

Outcomes and analysis

Primary endpoint: within each module or live session, change in first-attempt correct responses on three matched clinical concepts among learners with usable baseline and first post responses. Both formats use parallel case variants. Report live and enduring results separately, and record prior exposure because the enduring questions adapt the same live item bank. Do not count repeated exposure as independent evidence. Preserve first attempts; report retries and eventual passing separately.

Report item-level and participant-level paired scores, sample sizes, percentage-point changes and 95% confidence intervals. The analysis plan prespecifies participant-level resampling for the paired score change and appropriate paired item summaries. Missing responses are reported, not converted into incorrect answers or silently dropped from participation denominators. Retry success is reported separately from learning.

Planning success targets are at least 15 percentage-point matched improvement, 70% usable paired participation per live session, 70% enduring completion among starts and 30% response to delivered follow-up invitations. These are provisional targets, not predicted results or a power calculation. Pilot data and partner benchmarks should determine whether they are feasible.

At 60 days, invite consenting eligible participants to the three parallel follow-up application cases corresponding to each completed module/session and brief selections about encounters, actions and barriers. Nine reserved follow-up items cover the full series; these are not used as formative exercises. Coordinate invitations and prior exposure across live/enduring participation. Report retained competence separately from self-reported practice, with delivered-invitation denominators and bounce counts. No patient outcome or verified-performance claim is planned.

Use a stable pseudonymous identifier for analysis, with the identity linkage retained securely by the host. Deduplicate people across sessions and the companion. Provide course-specific results first; a series total must not double-count people or costs. Suppress small subgroup cells under the approved privacy policy and treat exploratory comparisons as such.

The proposed Moore level 4 assessment requires independent review of each parallel case, a defensible scoring key, pilot evidence and a prespecified analysis before release. The prepared items are not yet a validated instrument. Evaluate the live and enduring formats separately and interpret any 60-day application score within its response and equivalence limitations.