Gender Bias in Clinical Diagnosis of Borderline Personality Disorder and Autism Spectrum Disorder: A Retrospective Cohort Study

Abstract

Autism spectrum disorder (ASD) and borderline personality disorder (BPD) share overlapping clinical features including social cognitive differences, interpersonal difficulties, emotional dysregulation, and alexithymia. The two conditions show different gender prevalence patterns, and it has been suggested that gender bias may exist in diagnostic patterns of ASD and BPD, with females being preferentially diagnosed with BPD instead of ASD. In this analysis of real-world clinical records from n = 417,752 adults receiving secondary mental health care, we use natural language processing methodology to assess the subsequent incidence of ASD in cohorts diagnosed with BPD and the incidence of BPD in cohorts diagnosed with ASD. To examine the possibility of gender-based diagnostic bias, we assess for an association between gender and the sequence and timing of diagnoses. Results showed substantial rates of concurrent ASD and BPD; the 5-year cumulative incidence of later ASD diagnosis among people with BPD was 3.02% [95% CI: 2.62, 3.43] and of later BPD diagnosis among people with ASD was 3.84% [3.08, 4.60]. Consistent with predictions of unidirectional gender-based diagnostic bias, the typical association between male gender and ASD diagnosis was reduced in cohorts first diagnosed with BPD, whereas the association between female gender and BPD remained in those first diagnosed with ASD. In addition, the interaction between gender and diagnostic sequence was associated with the length of delay between diagnoses, such that women waited longer for an ASD diagnosis (relative to a BPD diagnosis) compared with men. Together, these results provide a number of avenues for future research and highlight the need for clinicians to consider neurodevelopmental assessment for patients diagnosed with BPD, particularly women, in order to investigate and address potential gender-based diagnostic bias.