ABSTRACT
Background
Misdiagnosis is common in bipolar disorder (BD). Currently used screening tools are brief and cost-effective, but there is a lack of understanding of their reliability over time, as well as whether responses are influenced by demographic or clinical factors.
Aims
To examine the cross-sectional and longitudinal reliability and validity of two commonly used screening tools, as well as their associations with current mood states and other participant characteristics.
Methods
A total of 331 adult patients with a diagnosis of BD completed the Mood Disorder Questionnaire (MDQ) and the Hypomania Symptom Checklist-32 (HCL-32) as well as measures of mood online at baseline and three months later.
Results
The MDQ was found to have low reliability with poor internal consistency (α = 0.531; α = 0.647 at respective timepoints) and test–retest reliability (r
s
= 0.582). The HCL-32 was more reliable with good internal consistency (α = 0.815 at both timepoints) and test–retest reliability (r
s
= 0.725). MDQ scores were significantly positively associated with comorbidities and current symptoms of activation/mania, and negatively correlated with age and years since last hospitalisation. The HCL-32 was also significantly positively correlated with activation/mania, and negatively correlated with age, years since diagnosis and years since last hospitalisation. BD-I patients scored higher than other BD subtypes. Mood state was the only variable with a significant change over time.
Conclusions
The HCL-32 appeared more reliable than the MDQ, although all participants needed to exceed an MDQ threshold at baseline to participate in the study, which may have biased our results.