Concordance between objective and subjective sleep measures of people with bipolar disorder: A systematic review and a meta-analysis.
Edwin Lok Yan Wong, Harvey Neal, Christopher Jones et al.
✦ AI-curated · Sources linked
In 30 seconds
This systematic review and meta-analysis examined agreement between objective and subjective sleep measures in people with bipolar disorder. Fifteen studies involving 572 participants contributed to five meta-analyses, comparing primarily actigraphy-based measures with sleep diaries and questionnaires. Total sleep time showed modest concordance: subjective estimates were higher but positively associated with objective measurements. Agreement was poor for sleep onset latency, sleep efficiency and wake after sleep onset, with subjective estimates respectively longer, higher and shorter. Analyses found no significant differences across bipolar presentations or diagnoses, but these subgroup results were underpowered and inconclusive.
Key findings
- Fifteen studies with 572 participants were included in five meta-analyses.
- Total sleep time showed modest concordance; subjective estimates were higher but positively associated with objective values.
- Concordance was poor for sleep onset latency, sleep efficiency and wake after sleep onset, with subjective estimates longer, higher and shorter, respectively.
- No significant differences across bipolar presentations or diagnoses were detected; subgroup analyses were underpowered and inconclusive.
Why it matters
Sleep assessment informs research and clinical management in bipolar disorder, but these findings indicate that subjective and objective measures are not interchangeable for several sleep parameters. Understanding where estimates diverge can help clinicians and researchers interpret sleep reports alongside measurement methods, rather than assuming that diaries or questionnaires closely reflect actigraphy-based results.
What to keep in mind
The evidence had high heterogeneity, small samples and inconsistent or inadequate statistical reporting. Few studies limited analyses of potential confounding factors, and subgroup analyses were underpowered.