ABSTRACT

D espite earlier m isconceptions that all personality disorders might ‘burn out’ in late life, an initial meta-analysis (Abrams & Horowitz, 1996) o f 11 studies indicated an overall prevalence rate o f 10%. In these studies, the m ost com m only diagnosed personality disorders were obsessive-com pulsive, dependent and ‘not otherwise specified’. W ith the addition o f five studies, a revised meta-analysis suggested an overall prevalence rate for late-life personality disorders o f 20% (Abrams & Horowitz, 1999). Across the included studies, the m ost frequently diagnosed disorders were paranoid, selfdefeating, and schizoid. Other studies not included in the meta-analysis or published after it was com pleted support the initial conclusion that cluster C personality disorders are m ost com m only diagnosed in late life (Kenan et al., 2000; Kunik et al., 1994; Vine & Steingart, 1994) as well as the previous finding o f relatively high rates o f personality disorder ‘not otherwise specified’ compared with other individual personality disorder diagnoses (Kenan et al.,

2000; Kunik et al., 1994). Therefore, reasonable conclusions from the meta-analyses and other studies are that the overall prevalence rate is between 10 and 20%, that the highest prevalence may be in clusters A and/or C, and that personality dysfunction in older adults may underestimated by focusing on those who m eet full criteria for any one disorder.