Psychometric properties, pragmatic screening cut-offs, and feasibility of two Clock Drawing Test formats in older psychiatric outpatients: Associations with MMSE and HDS-R.

Abstract

Aim: To compare the psychometric properties, pragmatic screening cut-offs, and feasibility of two Clock Drawing Test (CDT) formats in older psychiatric outpatients.

Methods: We retrospectively analyzed 447 outpatients aged ≥65 years who completed the Mini-Mental State Examination (MMSE), the Hasegawa Dementia Scale-Revised (HDS-R), and two CDT formats scored using the Freedman method: a free-drawn clock (CDT1) and an examiner-provided clock face (CDT2). Pearson correlations, forced-entry multiple regression models, exploratory factor analyses, and receiver operating characteristic (ROC) analyses were used. Possible cognitive impairment was operationally defined as MMSE ≤ 23 and/or HDS-R ≤ 20. Administration time was summarized in a convenience subsample (n = 20).

Results: Mean (SD) age was 80.20 (5.15) years, and mean education was 11.74 (2.36) years. CDT totals correlated positively with MMSE and HDS-R totals (r = 0.47-0.58, all p < 0.01). Education, functional disability, CDT1, and CDT2 predicted MMSE total, while age additionally predicted HDS-R total. Both CDT formats showed interpretable two-factor structures with good internal consistency (Cronbach's α 0.82-0.90). ROC analyses identified the same pragmatic cut-off (10/11) for both formats (CDT1 area under the curve [AUC] 0.72, 95% CI 0.67-0.77; CDT2 AUC 0.71, 95% CI 0.66-0.76). Mean CDT administration time was 2.13 min.

Conclusion: Both CDT formats showed clinically useful psychometric properties and similar pragmatic cut-offs in older psychiatric outpatients. CDT1 favored sensitivity, and CDT2 favored specificity at the same threshold. The CDT was highly feasible in routine practice, requiring approximately 2 min to administer.