Can People With Borderline Personality Disorder Change?

New research shows that borderline personality disorder can fluctuate over time.

Posted Jul 24, 2018

Source: Stockfour/Shutterstock

The idea that personality is stable over adulthood is no longer considered accurate in life-span psychology, even though the average person often assumes that change is impossible. When it comes to borderline personality disorder (BPD), many people therefore also commonly believe this is a permanent condition. You may know someone whom you believe fits the diagnostic criteria — they may demonstrate extreme fluctuations in their sense of self, ability to maintain relationships, impulsiveness, feelings of emptiness, and fear of abandonment. It seems to you that this person will be that way forever, and for better or worse, you need to figure out how you’re going to manage to keep your own life on an even keel.

According to College of William and Mary’s Christopher Conway and colleagues (2018), borderline personality has “a reputation among practitioners as an intractable disorder” (p. 1). Longitudinal studies that follow people over time, in contrast, suggests that the disorder does remit in the large majority of cases (85 percent), with only a minority of people experiencing the occasional recurrence of symptoms (30 percent). Conway et al. believe that rather than debate the issue of whether BPD is stable or not, the more relevant question pertains to which components of the disorder change and which are stable. This “hybrid model,” as they propose, suggests that there may be trait-like (stable) and state-like (dynamic) elements to BPD. Their research was intended to follow individuals with BPD over a long enough period to assess the relative degrees of stability and dynamic change in the features of this disorder.

Conway and his associates tested the hybrid model of BPD using a novel statistical method that tracks stability and change in personality. Known as “Trait-State-Occasion” (TSO) modeling, this approach allows personality researchers to separate the components of personality that are stable from those that might change. The research team obtained data from five testing occasions, collected over a 10-year period on a sample that initially consisted of 668 individuals tested at the New York State Psychiatric Institute. The participants completed an interview-based personality disorder measure, as well as a standard test assessing Five Factor Model personality traits (conscientiousness, extraversion, openness to experience, agreeableness, and neuroticism). Additional measures obtained at baseline provided Conway et al. with data on possible correlates of state- versus trait-based scores, including a questionnaire assessing childhood experiences relevant to BPD (e.g., neglect, abuse, witnessing of violence in the home), measures of temperament and inhibition, and an overall scale to assess their level of functioning.

The statistical approach the authors used allowed them to tease apart the stable from the variable components of personality. The traits that showed invariance using this model represented “the stable core of borderline proneness that does not fluctuate over the follow-up interval” (p. 4). Conversely, the “occasion factors represent causes of borderline PD that vary over time” (pp. 4-5). In correlating the time-invariant borderline proneness factor with the time-varying personality factors, Conway et al. reported that the borderline proneness factor was most highly correlated with neuroticism, and least strongly correlated with openness to experience. Changes over time in borderline proneness were also highly correlated with neuroticism, minimally with conscientiousness, agreeableness, and extraversion, and least correlated with openness. Overall, about half of the variance associated with changes in scores over the 10 years of the study appeared to be stable over time. Borderline symptoms appeared to be more stable over time (a .81 time-invariant correlation). There was also a strong relationship between those components of borderline pathology that varied over time and the five factor traits of neuroticism, openness, and conscientiousness.

Individuals who reported higher scores on the childhood maltreatment questionnaire had, as might be expected, a greater tendency toward borderline proneness. The time-invariant component scores within the model were highly related to scores on the negative temperament measure, but the time-varying component was far less so. There were also weak relationships between borderline proneness and positive temperament scores, and no relationships at all for the global measure of functioning.

These findings support, in the words of the authors, the view of BPD as having fixed and situational elements: “borderline PD can be decomposed into a borderline proneness element that is stable over time and a dynamic element that fluctuates from year to year” (p. 7). On the diagnostic side, this finding means that individuals showing variations in their levels of BPD symptoms might still fit the diagnosis even though their symptoms change in magnitude.

On the experiential side, the Conway et al. study suggests that people with BPD can change over time. It would seem that the people in the best position to modulate their symptoms are the ones who had a more positive childhood environment. However, even for individuals whose BPD seemed most related to childhood experiences, the correlation was not perfect with borderline proneness.

BPD, then, does not need to be conceptualized as a “totally unyielding condition” (p. 10). The component of BPD that is sensitive to change via social and environmental factors should be amenable to some combination of professional intervention and supportive relationships. If there are no such influences, the trait component of BPD may instead take over. Even a change in housing, as the authors suggest, could provide the stimulus for change in a positive or negative direction. Other researchers using the TSO approach have noted similar contributions of stable and variable influences on non-clinical aspects of functioning, such as self-esteem (e.g., Donnellen et al., 2012).

To sum up, an individual’s level of BPD at any one time reflects some proportional mixture of those elements that will remain constant and those that can either improve or worsen. With the right environmental inputs, it may be possible to change the course of the individual’s functioning over time, particularly if you recognize that such change is indeed possible.


Conway, C. C., Hopwood, C. J., Morey, L. C., & Skodol, A. E. (2018). Borderline personality disorder is equally trait-like and state-like over ten years in adult psychiatric patients. Journal of Abnormal Psychology, doi:10.1037/abn0000364

Donnellan, M. B., Kenny, D. A., Trzesniewski, K. H., Lucas, R. E., & Conger, R. D. (2012). Using trait-state models to evaluate the longitudinal consistency of global self-esteem from adolescence to adulthood. Journal of Research in Personality, 46(6), 634–645.