Correlates of the Abnormal Involuntary Movement Scale in Veterans With Tardive Dyskinesia.

Purpose/background: To add to limited evidence on the Abnormal Involuntary Movement Scale (AIMS) as a measure of tardive dyskinesia (TD) in clinical practice settings, the characteristics and correlates of AIMS scores were assessed.
Methods/procedures: Veterans with schizophrenia/schizoaffective, bipolar, or major depressive disorders receiving antipsychotics and at least 1 AIMS score during October 1, 2014, to September 30, 2015, were identified. Tardive dyskinesia was determined by the International Classification of Diseases, Ninth Revision, Clinical Modification, codes. Correlates of AIMS scores were examined using χ or t tests. Odds ratios and β parameters with 95% confidence intervals for categorical and continuous variables associated with AIMS scores were derived from a multivariate logistic and linear regression, respectively.
Findings/results: Among 7985 veterans receiving antipsychotics, only 4706 (58.9%) had at least 1 AIMS examination. Of these, 229 (4.9%) were diagnosed with possible TD. The mean total AIMS scores and AIMS awareness/incapacitation scores were significantly higher for patients with TD (both P < 0.0001). Comparing diagnostic threshold criteria of AIMS ratings, only 17.5% to 37.1% of veterans with TD were successfully identified. Among TD patients, 21.4% had a total score of moderate-severe and 15.3% had ratings of at least mild movements in 2 or more body regions. In the regression analyses, being older, African-American, having schizophrenia/schizoaffective disorder, and receiving antipsychotics or benztropine significantly increased the severity of AIMS scores. Higher AIMS scores were not predictive of outcomes other than marital status in socioeconomic or healthcare domains.
Implications/conclusions: Although the AIMS is essential for TD research, its value in clinical practice without training and oversight remains unclear. Efforts to adapt screening procedures to clinical needs may be worthwhile.