Greater adherence to healthful dietary patterns is associated with lower insomnia risk in the Women's Health Initiative Observational Study.
Level 3 - non-randomized controlled study
Prospective observational cohort study evaluating baseline exposure and longitudinal outcomes over 3 years.
PubMed 41065709 · doi:10.1093/sleep/zsaf316
What was done
In the Women's Health Initiative Observational Study (WHI-OS), researchers evaluated the prospective relationship between baseline diet quality and insomnia over a 3-year follow-up. Diet was assessed via a baseline Food Frequency Questionnaire to calculate Alternate Mediterranean (aMed) and DASH diet scores, which were dichotomized into good versus poor adherence using a data-driven approach. Insomnia was measured at baseline and Year 3 with the WHI Insomnia Rating Scale. Multivariable logistic regression adjusted for sociodemographic, lifestyle, and health factors to assess incident insomnia among women without baseline insomnia (n = 50,644) and longitudinal insomnia status (stable or new-onset insomnia vs. stable absence or remission, n = 74,513).
What was found
Among women without insomnia at baseline (n = 50,644), good versus poor diet quality was associated with lower odds of incident insomnia at 3 years: - aMed: OR 0.925 (95% CI: 0.879–0.974, p = .003) - DASH: OR 0.937 (95% CI: 0.891–0.985, p = .01) In longitudinal analyses across 74,513 women, baseline adherence was associated with lower odds of stable or new-onset insomnia: - aMed: 6.3% lower odds (OR 0.937 [95% CI: 0.903–0.971], p < .005) - DASH: 8.5% lower odds (OR 0.915 [95% CI: 0.883–0.948], p < .005)
Why it matters
This study provides large-scale prospective evidence that healthful dietary patterns (Mediterranean and DASH) are modestly associated with reduced risk and persistence of insomnia in postmenopausal women, identifying diet as a potential modifiable target for sleep health.
Limits
The observational design cannot establish causality and remains subject to residual confounding. Dietary intake relied on self-reported Food Frequency Questionnaires at baseline, introducing potential recall bias, and scores were simplified into dichotomized categories. Sleep was evaluated via a self-reported rating scale at only two time points, without objective sleep measurements (such as actigraphy or polysomnography) or tracking of interim fluctuations. Findings from this postmenopausal female cohort may not generalize to men or younger populations.