Changes in Health Literacy and Subsequent Presenteeism Among Workers: A Repeated-Measures Longitudinal Database Study

OBJECTIVES: To examine associations of changes in health literacy with subsequent changes in presenteeism and incident presenteeism among Japanese workers. METHODS: We analyzed 8 semiannual in-app surveys (June 2021-December 2024) from the DeSC database. Health literacy (HL) was measured with the 5-item Communicative and Critical Health Literacy scale (1-5), and presenteeism (Pres) with the WPAI-presenteeism item (0%-100%; higher indicating greater impairment; employed respondents only). Each observation used 3 consecutive waves: ΔHL(t-1)=HL(t-1)-HL(t-2) and subsequent ΔPres(t)=Pres(t)-Pres(t-1). Linear mixed models adjusted for prior Pres, earlier HL, and demographic, occupational, health-status, and lifestyle covariates; sensitivity analysis modeled prior Pres categorically. Additional analyses examined observations with prior presenteeism (Pres(t-1)>0), without prior presenteeism (Pres(t-1)=0) and categorical HL changes. RESULTS: Among 19,594 participants (60,203 observations), a 1-point more positive ΔHL was associated with more favorable subsequent ΔPres (β -1.24 percentage points; 95% confidence interval [CI], -1.52 to -0.97), with consistent sensitivity and prior-presenteeism findings; the incident odds ratio was 0.89 (0.84-0.93). Compared with unchanged HL, decreases were associated with less favorable changes (0.87; 0.46-1.27), whereas increases had a CI including zero (-0.32; -0.75 to 0.10). Increases of at least 0.6 points showed favorable exploratory associations. The individual-level association was modest (approximately -0.83 percentage points per standard deviation of ΔHL). CONCLUSIONS: More positive HL changes were associated with modestly more favorable presenteeism outcomes. Direction-based associations were clear for decreases but not increases; larger increases showed favorable exploratory associations. Intervention effects require further study.