Preliminary validation of a UK modified version of the BASIS-32

INTRODUCTION: The authors examined the reliability and validity of a UK-modified version of the Behavioral and Symptom Identification Scale (BASIS-32).

METHOD: Data from two samples of patients from acute psychiatric inpatient settings were used in the analyses (n=303, n=92). The factor structure of the scale differed from that of the original BASIS-32.

RESULTS: Five factors emerged: (i) depression and anxiety, (ii) lability, (iii) psychosis, (iv) substance misuse and (v) functioning. The full scale of the modified BASIS demonstrated high internal consistency (Cronbach's alpha=0.93). Internal consistency for the subscales ranged from 0.86 to 0.45. The depression and anxiety subscale discriminated patients with a diagnosis of unipolar depression from those with other diagnoses (median score=3.11 (Inter quautile range (IQR)=2.67,3.33) versus median score=2.44 (1.67,3.11), P < 0.001), The substance misuse subscale discriminated patients with a diagnosis of alcohol or opiate dependence from those with other diagnoses (median score=2.33 (2, 3) versus median score=1.33 (0.67, 2), P < 0.001). However, the psychosis subscale did not differentiate patients with a psychotic illness from those with a nonpsychotic diagnosis (median score=1.4 (0.6,2.4) versus median score=1.2, (0.6,2), P=0.16).

CONCLUSION: The total scale appeared to be at least as good as the Brief Symptom Inventory (BSI) in its responsiveness to change. The effect size for the BASIS=1.17 versus 0.91 for the BSI. Convergent validity was partially demonstrated between the modified BASIS and the BSI.