Associations of health literacy and eHealth literacy with cyberchondria: a systematic review and meta-analysis

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Abstract Background Health literacy and eHealth literacy are related but conceptually distinct constructs that may influence how individuals engage with online health information. However, observational studies have reported positive, negative, and null associations with cyberchondria. This systematic review and meta-analysis synthesized the available evidence and examined whether the association varied according to health-literacy construct and study population. Methods PubMed, Web of Science, Embase, CINAHL, PsycINFO, and the China National Knowledge Infrastructure were searched from database inception through March 23, 2026. Analytical observational studies assessing general health literacy, eHealth literacy, or digital health literacy in relation to cyberchondria were eligible for the systematic review. The strict primary meta-analysis included only unadjusted zero-order Pearson or Spearman correlations with corresponding sample sizes. Correlations were transformed to Fisher’s z values and pooled using restricted maximum-likelihood random-effects models with Hartung–Knapp confidence intervals and 95% prediction intervals. Moderator analyses examined health-literacy construct and study population. Exploratory univariable meta-regressions, Pearson-only and leave-one-out sensitivity analyses, eligibility-boundary analyses, funnel plots, and Egger’s regression were also conducted. Results Thirty reports involving 18,904 participants were included in the systematic review, all of which used cross-sectional designs. The strict primary meta-analysis included 19 reports contributing 21 independent effect estimates and 9,556 participants. The pooled correlation was close to zero and not statistically significant ( r = 0.041, 95% Hartung–Knapp CI − 0.105 to 0.185), with very high heterogeneity (I² = 97.4%) and a wide 95% prediction interval (− 0.563 to 0.616). Neither health-literacy construct ( P = 0.287) nor study population ( P = 0.214) significantly moderated the association. Exploratory meta-regressions showed no statistically significant associations with mean age or the percentage of female participants. Pearson-only, leave-one-out, and eligibility-boundary analyses did not materially alter the overall conclusion. Egger’s regression showed no statistical evidence of small-study effects ( P = 0.461), although its interpretation was limited by the modest number of effects and very high heterogeneity. Conclusions The average cross-sectional association between health literacy, including eHealth literacy, and cyberchondria was negligible, but the direction and magnitude of individual study effects varied substantially. Current evidence does not establish a consistent protective or harmful association or support causal conclusions. Future longitudinal and theory-driven studies should distinguish literacy constructs, control important confounders, and investigate potential mediating and moderating mechanisms rather than focusing only on the direction of the association. Trial registration PROSPERO CRD420261346255.