Merikangas and Weissman (1991) pointed out the hazards of using treated rates for drawing inferences about true prevalence of personality disorders. They identified a number of factors that could lead to bias in treated samples: (1) differences in the availability of treatment; (2) the role of cultural factors in help-seeking behavior; (3) differences in the severity of the disorder; (4) the potential influence of other comorbid psychiatric disorders; and (5) differences among the personality disorders in the likelihood of seeking treatment. However, the importance of personality disorders in clinical practice makes their prevalence in such settings valuable information in its own right.
Dahl (1986) reported results of the systematic assessment of DSM-III personality disorders in 231 consecutive admissions to a psychiatric inpatient unit in Norway. Chronic patients and those with organic disorders were excluded. Approximately 45% of the sample received a personality disorder diagnosis (40% of females and 49% of males). Forty-four percent of those with a personality disorder received one diagnosis, 36% had two diagnoses, 15% had three diagnoses, and 5% had four diagnoses. Schizotypal, histrionic, antisocial, and borderline personality disorders were each present in approximately 20% of the sample. Avoidant personality disorder was diagnosed in 9% of the sample and the remaining personality disorders were diagnosed much less frequently. In a series of 100 patients admitted with major psychiatric disorders (affective disorders, schizophrenia, and other functional psychosis), Cutting et al. (1986) found that 44% had an ''abnormal personality'' based on informant interviews about the period preceding the acute episode.
Oldham and Skodol (1991) investigated the prevalence of personality disorders among the 129,268 patients treated in New York state mental health facilities in 1988. Using the system's centralized database, they found that 10.8% received a personality disorder diagnosis. The personality disorder diagnosis was the primary diagnosis of 1.2% of patients and only 0.2% of patients received more than one personality disorder diagnosis. The most common diagnosis was borderline personality disorder; 17.2% of the patients with any personality disorder received a diagnosis of borderline personality disorder. These authors concluded that the standard record-keeping procedures underestimated the prevalence of personality disorders and that personality disorders were not being systematically assessed.
The prevalence of personality disorders has also been examined in nonpsychi-atric medical populations. Casey and her coworkers reported on the prevalence of personality disorders in British clinical settings using the ICD classification (Casey et al., 1984; Casey et al., 1985). They found a 34% prevalence of personality disorder in primary care settings; anxiety states and alcohol abuse were the conditions most commonly associated with personality disorders. In a rural general practice, they found a prevalence rate for personality disorders of 20%. Reich et al. (1989) found a significant positive association between the presence of personality disorder and the probability of being hospitalized for a nonpsychiatric medical illness.
As is the case for both axis I and axis II disorders, treated rates shed little light on the prevalence of disorders in untreated samples. However, because personality disorders have important implications for service provision, it is useful to consider their frequency in various clinical populations. Obviously, the nature of the setting in which the disorders are studied and the method by which they are studied will influence the findings. The treated rates of the individual personality disorders will be included in the next section. [Data on treated rates are drawn primarily from a review by Widiger (1991)].
Was this article helpful?