Temporal reliability of personality in psychiatric patients

The Personality Assessment Schedule, an interview schedule specifically designed for assessing personality disorder, was administered twice to 28 psychiatric patients, with a mean interval of 2.9 years between each assessment. The first assessment was made by a psychiatrist and the second by a medical student who had no prior knowledge of the patients. The reliability of the 2 assessments was measured using 4 different techniques. Although the reliability of individual personality traits was inconsistent over time, the categorical diagnosis of personality disorder was good (Kw = 0.64), giving some support to the validity of the schedule. Reasons for discordance in the assessments were examined and appeared to be due mainly to confusion between clinical symptoms and personality traits, retrospective errors in recording past personality in chronic patients, and special difficulties in determining the primary abnormality in severe personality disorder.