Psychometric evaluation of a newly developed measure of emotionalism after stroke (TEARS-Q)

Abstract

Objective: To evaluate, psychometrically, a new measure of tearful emotionalism following stroke: Testing Emotionalism After Recent Stroke - Questionnaire (TEARS-Q).

Setting: Acute stroke units based in nine Scottish hospitals, in the context of a longitudinal cohort study of post-stroke emotionalism.

Subjects: A total of 224 clinically diagnosed stroke survivors recruited between October 1st 2015 and September 30th 2018, within 2 weeks of their stroke.

Measures: The measure was the self-report questionnaire TEARS-Q, constructed based on post-stroke tearful emotionalism diagnostic criteria: (i) increased tearfulness, (ii) crying comes on suddenly, with no warning (iii) crying not under usual social control and (iv) crying episodes occur at least once weekly. The reference standard was presence/absence of emotionalism on a diagnostic, semi-structured post-stroke emotionalism interview, administered at the same assessment point. Stroke, mood, cognition and functional outcome measures were also completed by the subjects.

Results: A total of 97 subjects were female, with a mean age 65.1 years. 205 subjects had sustained ischaemic stroke. 61 subjects were classified as mild stroke. TEARS-Q was internally consistent (Cronbach's alpha 0.87). TEARS-Q scores readily discriminated the two groups, with a mean difference of -7.18, 95% CI (-8.07 to -6.29). A cut off score of 2 on TEARS-Q correctly identified 53 of the 61 stroke survivors with tearful emotionalism and 140 of the 156 stroke survivors without tearful emotionalism. One factor accounted for 57% of the item response variance, and all eight TEARS-Q items acceptably discriminated underlying emotionalism.

Conclusion: TEARS-Q accurately diagnoses tearful emotionalism after stroke.

Keywords: Stroke; emotionalism; mental health.

Figure 1.

Receiver operating characteristic curve for testing emotionalism after recent stroke – questionnaire. This gives a graphical representation of the values of sensitivity and specificity for each potential cut point for the TEARS-Q measure. The Youden cut point maximises the sum of sensitivity and specificity and occurs at 1.5 yielding a specificity of 90.1% and a sensitivity of 88.7%.

Figure 2.

Scree plot for TEARS-Q indicating one dimension. The pronounced ‘elbow’ shows that most of the information comes from the first component of a principal component analysis: that is TEARS-Q is a unidimensional measure.