Nurse-led Intervention for Cognitive Reserve and Stress Management for Teachers Approaching Retirement: a Cluster-randomized Controlled Trial

This interventional trial evaluated a nurse-led, workplace-based approach to supporting cognitive and psychological health during the pre-retirement period - a life stage in which changes to daily structure, occupational complexity, social contact, and role identity may influence later cognitive trajectories and stress. The rationale drew on cognitive-reserve theory together with evidence that persistent perceived stress can constrain the behavioral expression of reserve; the two prespecified outcomes were therefore evaluated concurrently rather than as independent targets. Development followed Medical Research Council guidance for complex interventions, and the intervention was specified using the Template for Intervention Description and Replication (TIDieR).

The school served as the unit of randomization, while the individual teacher was the unit of recruitment, delivery, and measurement. To limit selection bias, all eligible teachers within each participating school were invited and individual recruitment was completed before the allocation sequence was generated. A researcher responsible for the statistical analysis then generated the school-level sequence by computer, assigned schools 1:1 to conditions, and informed schools of their allocation afterward. Delivering the program at the level of the whole school was intended to reduce exchange of intervention materials between conditions (contamination). Because the program was educational and behavioral, participants and the nurse interventionists could not be masked; masking of outcome assessment and analysis was not undertaken, and both outcomes were self-reported.

The intervention was developed by the research team and reviewed by faculty experts in community-health, psychiatric mental-health, and gerontological nursing for content, cultural suitability, sequence, and feasibility. Standardized materials - an illustrated Arabic booklet, visual aids, activity materials, brief instructional videos, and session checklists supported consistent delivery. Two nurse-researchers rehearsed manualized session scripts and standardized timing, activity administration, attendance recording, and documentation of protocol deviations before implementation. Facilitators used content checklists to confirm delivery of planned objectives and to log any deviations in content or duration.

The primary analysis used complete-case linear mixed-effects models with fixed effects for group, time (as a categorical factor), and their interaction, plus a participant-level random intercept, estimated by restricted maximum likelihood. Because the archived analysis file did not contain the school identifier, these models accounted for repeated measures within participants but not for clustering of teachers within schools; with only four clusters, cluster-level variance and conventional asymptotic standard errors may be unstable. Inferential results are therefore interpreted as exploratory and are supported by two post hoc sensitivity analyses: a cluster-precision analysis applying a design-effect adjustment across a range of plausible intracluster correlation coefficients, and a series of progressively adverse missing-outcome scenarios. No a priori cluster-adjusted sample-size calculation was performed; census-style recruitment was used, and interpretation emphasizes effect estimates and their uncertainty rather than post hoc power.