Clinimetric evaluation of the Simpson-Angus Scale in older adults with schizophrenia

Parkinsonism (or Parkinson's syndrome [PS]) remains common in patients exposed to antipsychotic drugs. One clinical tool used in its detection and follow-up, the Simpson-Angus Scale (SAS), has been under revision lately. We further examined the discriminative power of the SAS to detect PS and its efficacy as a measure of PS intensity in chronic schizophrenia.

Fifty-six outpatients between 50 and 75 years of age, under stable antipsychotic drug therapy, provided consent to undergo an evaluation along the SAS and Unified Parkinson's Disease Rating Scale III motor subsection, split according to the presence or absence of PS defined in the UK Parkinson's Disease Society Brain Bank (UKPDSBB) criteria or Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition criteria.

The identification rate for PS was 39.3% based on UKPDSBB criteria applied to the Unified Parkinson's Disease Rating Scale III, compared with 62.5% and 87.5% according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition criteria and SAS cutoff value greater than 0.3, respectively. Median SAS scores for PS and PS-free participants were comparable. The SAS yielded high sensitivity (90.9%) but low specificity (17.7%). κ Values generally revealed only slight agreement between the group allocation provided by the SAS and the UKPDSBB criteria. Receiver operating characteristic curve for screening performance of the SAS provided poor prediction of subject status.

The SAS lacks specificity and constitutes an imperfect detection and measurement tool for PS in older adults. Raising the cutoff score would avoid inflation in PS identification. The scale is probably best used as a measure of change relative to baseline score following an intervention, but results should be interpreted with caution.