Using nationwide registry data, we show that personality disorder diagnoses are widely used in specialist health care, reflecting the prevalent nature of these mental health conditions. Age trends show an absence of diagnoses in childhood, a sudden rise in incidence in late adolescence with peaks in early adulthood, and declines in incidence through adulthood toward zero incidence in later life. Of note, personality disorder diagnosis became more incident in the Norwegian population from 2010 to 2022. Borderline and avoidant personality disorder were markedly more incident than the other specific diagnoses, and they also increased the most in incidence over the years of follow-up. Personality disorder diagnosis in general was more incident, and increased more in incidence, among women than among men.

The observed relative distribution of specific personality disorders differs markedly from estimates obtained using diagnostic interviews in community samples [11], even when sampled from a comparable population (Norway in the 1990s) [22]. For example, the incidence of borderline personality disorder stands out as higher than other diagnoses, while its prevalence in community samples does not [11, 22]. This underscores the complementary relationship between diagnostic interview studies and registry studies. Crucially, secondary health care diagnosis typically is the end point of a causal chain where functional impairment plays a substantial role [14]. Therefore, these differences between studies with different methodologies are likely to reflect differences in functional impairment between the personality disorder diagnoses. Further, some of these differences might be due to the pragmatic nature of clinical practice – rather than focusing solely on results from diagnostic interviews, clinicians often consider the utility of selecting one particular diagnosis over the other. Borderline personality disorder diagnosis could be assumed particularly useful by clinicians because it has become associated with relatively clearly defined treatment courses. Other diagnoses, like histrionic personality disorder, could be associated with stigma and therefore with less utility [23]. Finally, we note that avoidant personality disorder is more widespread, and dissocial personality disorder less widespread, in Norway than in other countries for which data is available [11, 14, 22].

On the other hand, our estimates of incidences and lifetime risks for personality disorders largely converge with estimates in studies using similar methodologies in comparable demographic contexts (Denmark) [15, 16]. The most recent of these studies [16] found lifetime risks for personality disorder diagnosis at 7.1% among women and 3.1% among men, comparable to 6.7% among women and 4.1% among men in the Norwegian population. This convergence suggests that Danish and Norwegian mental health professionals share a similar understanding of personality disorder diagnosis. Further, the lifetime risk of dissocial personality disorder diagnosis among women was 0.08% in Denmark and 0.09% in Norway, and among men it was 0.41% in Denmark and 0.39% in Norway. Further, the lifetime risk of borderline personality disorder was 2.95% in Danish and 2.85% in Norwegian women. The only divergence was between Danish and Norwegian men in borderline personality disorder risk, at 0.35% and 0.73%, respectively. The risks for specific personality disorders other than borderline and dissocial were not estimated in the Danish studies.

We find that the incidence of personality disorder diagnosis is low through childhood, though increasing from mid adolescence, peaking in early adulthood, then declining through later life. This pattern is common among mental illnesses. However, there is reason to believe that diagnostic practices shift this curve to the right. The ICD-10 diagnostic manual, still used by Norwegian clinicians through the period of study, explicitly states it is “unlikely that the diagnosis of personality disorder will be appropriate before the age of 16 or 17 years” [24]. In addition, clinicians might be reluctant to diagnose adolescents with personality disorder due to concerns about the stability of personality at that age and/or fear of stigma [5, 25,26,27]. Such practices contrast with the evidence that personality disorder, especially borderline personality disorder, is prevalent, measurable, and treatable in adolescents [5, 25, 28,29,30].

Personality disorder diagnosis is more common in women than in men, with the exceptions of dissocial, schizoid, and paranoid personality disorder. This pattern is found also in other populations [16]. The higher incidence of diagnosis in women is likely influenced by a combination of gender differences in health care-seeking behavior [31], clinician bias [32], and gender differences in the exposure to etiological factors [33].

The incidence of personality disorder diagnosis increased through the study period, though the increase was due entirely to a subset of specific diagnoses: borderline, avoidant, anankastic, and mixed personality disoder. The increase in diagnosis could in part reflect increases in the underlying symptomatology in the population. The period of follow-up coincides with a historical period where the Norwegian population appears to have become more burdened by general psychiatric symptoms [34]. This secular trend has been more pronounced in young women than in other demographics [34], potentially contributing to the gender differences we observe in the increase in personality disorder diagnosis. The increase in diagnosis could also in part reflect conscious efforts by Norwegian health authorities and researchers to standardize and improve diagnostic measurement. In 2012, a national advisory unit on personality disorders was established, aiming to educate health services and the public on personality disorders, diagnostic assessment, and management and understanding of lived experience (user) perspectives [35]. Further, the increase in diagnosis could also in part reflect more widespread administration of psychotherapeutic treatments for personality disorder or an increased awareness of their efficacy [6, 7]. For example, evidence-based structured treatments for borderline personality disorder with corresponding therapist training courses were increasingly available in Norway from 2010. Thus, disentangling the different factors contributing to increases in personality disorder diagnosis should be a focus of future research.

This study describes phenomena which are particular to Norway and which, though largely converging with epidemiological results from comparable countries (Denmark) [16], cannot immediately be generalized across national contexts. We also note that we study registered diagnoses, and that the results depend on the recording of diagnosis in Norwegian specialist health care. A further limitation is the potential for insufficient censoring of prevalent cases at the start of follow-up, since our washout period lasts only 4 years. By extending the washout period, we show that central epidemiological parameter estimates do not change in a manner consistent with considerable bias, but we cannot completely eliminate the potential for false incidences due to left truncation.