Beliefs related to diagnosing borderline personality disorder and complex post‐traumatic stress disorder

Abstract

Objectives

There has been speculation that Borderline personality disorder (BPD) and Complex post‐traumatic stress disorder (cPTSD) are the same disorder or represent separate disorders on a broader trauma spectrum. BPD and cPTSD share several symptoms but differ in their presentation. Because the two disorders present so similarly, it could be that clinicians have difficulty differentiating them, leading to misdiagnosis. BPD is a commonly stigmatized condition, and stigma can be a source of diagnostic error, contributing to misdiagnosis. This study examined clinicians' diagnostic decision‐making and stigma towards BPD and Complex PTSD.

Methods

Participants (n = 426) completed an online survey where they diagnosed a written case vignette and completed measures of their familiarity with each disorder, stigma and symptom importance for their diagnosis.

Results

Results indicated that clinicians were significantly less familiar with BPD than with cPTSD; however, only familiarity with cPTSD predicted diagnostic decisions. Higher familiarity with BPD was associated with lower stigma towards BPD, whereas familiarity with cPTSD predicted lower stigma for both disorders. Stigma levels were significantly higher for BPD than for cPTSD, but only stigma towards cPTSD predicted diagnosis. Symptom ranking strongly influenced diagnosis, with clinicians more likely to assign a diagnosis based on symptoms they associated with BPD or cPTSD. BPD stigma influenced rankings of BPD symptoms, whereas cPTSD stigma affected rankings across all symptom categories.

Conclusion

These findings suggest that familiarity and stigma play crucial roles in shaping diagnostic decisions and highlight the need for improved clinician training to mitigate biases and enhance diagnostic accuracy.

Practitioner points.

• BPD and cPTSD share many similar symptoms and can be difficult for clinicians to differentiate, with stigma playing a role in diagnostic errors.

• Clinicians reported lower familiarity and higher stigma for BPD compared to cPTSD, yet only familiarity and stigma related to cPTSD predicted diagnostic decisions. Symptom rankings strongly influenced diagnosis, with BPD stigma shaping rankings of BPD symptoms and cPTSD stigma affecting rankings across all symptom categories.

• The difficulty of obtaining honest responses about stigma towards mental health disorders represents a key limitation of this study.

• Future research should extend this research to psychiatrists who, in some jurisdictions, are primarily responsible for diagnostic decisions.

INTRODUCTION

Accurate diagnoses in mental health are critical for providing clients with appropriate explanations of their conditions and informing treatment plans (Holmboe & Durning, 2014). Diagnostic decision‐making is the process in which clinicians gather information needed to make a preliminary diagnosis (Thammasitboon & Cutrer, 2013). Misdiagnosis can occur due to the clinician's lack of knowledge, inherent biases, or even an intentional decision (Kirk & Kutchins, 1988). Diagnostic errors, whether misdiagnosis or reluctance, are cognitive in nature and often result from biases and stigma from the assessor (Norman & Eva, 2010). Stigma can be expressed with reduced empathy, improper or reduced treatment and feelings of irritation, anger or difficulty, which impact a clinician's judgement (Masland et al., 2023).

Borderline personality disorder (BPD) is often a complex disorder for clinicians to accurately diagnose (Biskin & Paris, 2012). Misdiagnosis of BPD may be due, in part, to the relatedness of this disorder to other similar disorders, such as bipolar disorder (BD), post‐traumatic stress disorder (PTSD) and complex post‐traumatic stress disorder (cPTSD; Frost et al., 2020). The overlap in BPD symptoms with other personality disorders can also make it difficult to parse (Bach et al., 2016). Other comorbid disorders, like eating disorders and substance use disorders, add to the complexity, making BPD one of the most comorbid disorders (Eaton et al., 2011). Additionally, due to the societal stigma that accompanies BPD, clinicians may be reluctant to assign this diagnostic label.

cPTSD

Complex Posttraumatic Stress Disorder (cPTSD) was added to the newest edition of the International Classification of Diseases and Related Health Problems (11th Revision; ICD‐11; WHO, 2024). It has a prevalence rate of 3.8% (Cloitre et al., 2019) and was the disorder that clinicians most frequently recommended for addition to the ICD‐11 worldwide (Robles et al., 2014). cPTSD was proposed as a diagnosis to capture individuals who experience long and recurring trauma with symptoms that go beyond the traditional definition of PTSD (Herman, 1992). cPTSD contains all symptoms of PTSD, including re‐experiencing trauma, which can be in the form of flashbacks, deliberate avoidance of triggering events and hypervigilance (WHO, 2020). In addition to traditional PTSD symptoms, cPTSD is characterized by three new symptom clusters collectively called Disturbances in Self‐Organization (DSO), which include affect dysregulation, interpersonal disturbances and negative self‐concept. All of the core symptoms of PTSD must be met along with a symptom from each DSO cluster.

Attempts to distinguish BPD and cPTSD

The existing literature suggests that BPD and DSO may not be mutually exclusive, and symptoms are often presented differently between the two disorders. First, a diagnosis of BPD does not require the occurrence of a traumatic event, whereas a traumatic stressor is needed for the diagnosis of cPTSD. Emotion dysregulation is often present in both disorders but is expressed as emotional sensitivity and anger in cPTSD (Cloitre et al., 2014). Also, in cPTSD, fear of abandonment is often not present, and self‐identity does not often fluctuate; instead, it stays steadily negative (Cloitre et al., 2014).

Knefel et al. (2016) concluded that the BPD symptom network was not closely related to the PTSD or DSO networks, suggesting that cPTSD is unique and fundamentally different from BPD. However, minor connections were found between BPD and DSO. Three DSO symptoms—anger, reckless behaviour and feeling distant or cut off—grouped with one BPD symptom, identity disturbance. Another group consisted of BPD's unstable interpersonal relationships and DSO's difficulties with feeling close to others and avoidance of relationships. PTSD also showed connections: distressing dreams linked to BPD's chronic feelings of emptiness and PTSD's internal avoidance linked to BPD's identity disturbances. Owczarek et al. (2023), using the same analytic method, found that BPD and cPTSD are largely separate diagnoses, with affective dysregulation being the only cPTSD symptom related to BPD.

Jowett et al. (2020) found that cPTSD has stronger associations than BPD with early and recurring childhood trauma, along with greater functional impairment. BPD symptoms were elevated in relationship impairment. When comparing the severity of symptoms, impulsivity and inappropriate anger were the most common factors differentiating moderate severity cPTSD and BPD with low severity cPTSD and BPD. Cloitre et al. (2014) investigated the difference in construct validity between cPTSD and BPD in women with histories of childhood abuse. Using latent class analysis (LCA), they found four distinct classes: a Low Symptom class (low symptoms endorsement of all disorders), PTSD class (high PTSD with low cPTSD and BPD), a cPTSD class (high PTSD and cPTSD) and a BPD class (BPD only symptoms). Of the BPD symptoms, there were four that increased the odds of a diagnosis of BPD compared to cPTSD: fear of abandonment, identity disturbances, difficulty in interpersonal relationships and impulsivity.

Diagnostic decision‐making

Given the similarity between BPD and cPTSD, it is important to consider the processes which clinicians might use to differentiate the conditions when making a diagnosis. Clinicians collect data to make diagnoses in a variety of ways. Unlike other specialties, psychological diagnostic decision‐making has no gold standard for judging accuracy (Graber et al., 2005). Clinicians are unable to run a test to determine if their diagnostic hypothesis is correct. Instead, symptoms are usually collected through client self‐report and observation during the clinical interview (Singh & Weingart, 2009). Further, clinicians often use assessment tools such as single construct scales or broadband measures to assist them in the diagnostic process. Even with assessment tools, clinicians often use heuristics, or mental shortcuts, to aid in forming a diagnosis (Benishek et al., 2015). Clinicians often pay attention to the more salient symptoms while overlooking less salient ones.

Factors influencing diagnostic decisions

One factor that can influence the diagnostic decision‐making process is stigma. Mental health professionals have played a significant role in the stigmatization of mental health in the clients with whom they work (Ociskova et al., 2017). Bias and stigma can present in diagnostic criteria, therapeutic interviews, psychological tests and clinical judgement, which is especially true for personality disorders (Garb, 2021). In diagnosing BPD, mental health professionals report feeling reluctant to make a diagnosis because of fear that the diagnostic label will lead to further rejection in the healthcare system (Paris, 2007).

Many clinicians across specialties report lacking the knowledge or experience necessary to diagnose BPD. Saunders et al. (2015) found that psychiatrists were able to successfully identify BPD symptom criteria but felt uncertain with diagnostic differentiation. They also found that many mental health professionals also do not feel comfortable with the diagnosis of BPD, and many do not wish to disclose to their patients if they receive a diagnosis. Sisti et al. (2016) found that about 37% of psychiatrists have failed to document a diagnosis of BPD, and a minority (42%) were willing to disclose any form of personality diagnosis to their patients. They also reported that many psychiatrists agreed that they felt a lack of education about BPD or felt stigma towards their clients that prevented them from disclosing their diagnoses.

Beyond issues of diagnostic transparency, 85% of clinicians admitted using defensive practice, such as prolonged hospitalization, even when they admit it was not in their client's best interest, to protect themselves from repercussions with individuals diagnosed with BPD (Krawitz & Batcheler, 2006). These clinicians also reported perceiving a lack of support for more tolerant approaches to care and attributed their behaviour to anxiety and concerns about the treatment system (e.g., senior managers and clinical directors), external agencies (e.g., the media, national government) and the patient's broader social context (e.g., family and friends). McKenzie et al. (2022) found that mental health professionals reported that individuals with BPD evoked greater negative emotions and were more disliked than other individuals with mental health conditions. These clinicians also admitted to being less likely to give advice, think about, or spend extra treatment time with individuals who have BPD.

CURRENT STUDY

The aim of the present research is to examine factors, such as familiarity and stigma, on clinicians' diagnostic decisions when diagnosing BPD and cPTSD. By presenting clinicians with a case vignette with symptoms of both cPTSD and BPD, the study expected that clinicians with higher levels of stigma towards BPD would be less likely to diagnose BPD and more likely to diagnose cPTSD. The study also aimed to examine clinicians' ability to identify symptoms associated with BPD and cPTSD. We hypothesized that clinicians with greater stigma towards BPD would be more likely to identify symptoms in the case vignettes as solely cPTSD and less likely to classify symptoms as belonging to both BPD and cPTSD or only BPD. Through further exploration of how clinicians diagnose BPD and cPTSD, the research aimed to improve diagnostic practices in the future, such as through potential educational programs and training to reduce bias and increase diagnostic accuracy.

METHOD

Participants

The study recruited a sample of potential participants through state licensing boards. Many licensing boards for mental health professionals are publicly available. We recruited a broad sample of licensed psychologists, licensed professional counsellors (LPC), social workers and marriage and family therapists (MFT) by obtaining licensee lists from Florida and Rhode Island. Our decision to recruit from these specific U.S. states and specific mental health professions was in part due to the availability of obtaining participant email addresses. Of note, contact information was not available for psychiatrists. In order to take part in the study, participants must be currently or formerly licensed to provide mental health diagnoses and able to complete the survey in one sitting.

The total potential participant pool was 53,840. Previous studies using this recruitment strategy have obtained approximately a 1% response rate (Chung, 2021; Keeley et al., 2013; Kleva, 2023; Webb & Keeley, 2017). Of the 53,840 emails sent, 1304 participants clicked the survey link (2.42%), 881 participants (1.63%) reached the end of the survey and 426 participants answered 80% or more of the questions (.79%). It is important to note that our analytic plan does not entail investigating differences between states or professions. See Tables 1 and 2 for information about participants who completed the study.

TABLE 1.

Participant demographics.

	ƒ (%) or M (SD)
Age	46.47 (12.26)
Gender
Man	59 (13.8)
Woman	350 (82.3)
Nonbinary	8 (1.9)
Self‐identify	4 (.7)
Sex assignment
Cisgender	402 (94.4)
Different than sex assigned at birth	9 (2.1)
Race/Ethnicity
White/Caucasian	362 (85.0)
Hispanic/Latinx	49 (11.5)
Black/African American	17 (4.0)
Native American	2 (.5)
East Asian/Pacific Islander	5 (1.2)
Profession
Psychologist	100 (23.9)
Professional counsellor	188 (44.9)
Marriage and family therapist	62 (14.8)
Social worker	69 (16.5)
Highest degree earned
Master's	272 (63.8)
Doctoral	144 (33.8)
Other	7 (.7)
Conducts supervision
Yes	175 (41.1)
No	236 (55.4)
Weekly responsibilities
Psychotherapy	63.39 (31.80)
Assessment	9.90 (16.57)
Clinical supervision	6.22 (12.04)
Clients with BPD worked with/supervised per week	339 (79.57)
Clients with CPTSD worked with/supervised per week	377 (88.50%)

TABLE 2.

Participant training and current work.

	ƒ (%)
Work settings
Outpatient	46.50
Private practice	62.0
Telehealth	40.10
Hospital	3.76
Caseload
1–5	6.68
6–10	11.69
11–15	13.84
16–20	19.33
21–25	18.38
26–30	13.84
30+	11.00
Age of clients
11 and under	10.56
12–17	30.52
18–24	65.26
25–39	85.21
40–60	70.66
Over 60	33.33
Types of training for BPD and CPTSD
Didactic coursework	54.00
Clinical experience/practicum	60.56
Continuing education credits	56.81
Lived experience	36.38
Self‐study	70.19
Theoretical orientation
CBT	21.58
Psychodynamic	7.11
Family systems	1.58
Humanistic	8.16
Integrative/Eclectic	50.26
Other	11.32

Note: Clinicians indicated types of training they received for both BPD and Complex PTSD.

Procedure

Following approval from Virginia Commonwealth University's institutional review board, participants received an email link directing them to Qualtrics, an online survey platform. Participants reviewed informed consent information and indicated their willingness to participate. Participants were prompted to read a vignette about a mock client arriving for their initial appointment. Then, participants were presented with an option to diagnose the patient with BPD, cPTSD, both disorders, or they could write in other diagnoses that they believed were appropriate. Participants had access to both the vignette and the diagnostic criteria from the DSM for BPD and ICD‐11 for cPTSD when making their diagnostic decisions. Participants then viewed a list of symptoms from the vignette and selected to which disorder they believed the symptoms belonged. Participants also ranked the symptoms based on their perceived importance in making their diagnostic decision. Participants rated the certainty of their diagnostic rating on a sliding scale of 0–100. Next, participants answered questions on the Borderline Patients‐Emotional Attitudes Inventory (BPD‐EA) related to their stigma towards BPD and cPTSD. Last, participants answered a demographic questionnaire. They were then thanked for their participation and dismissed from the study.

Measures

Clinical vignette

The vignette was developed by Woodward et al. (2009). We adapted it to include diagnostic criteria for cPTSD. The client presented with difficulty in their relationship, frequently changing jobs and a history of a traumatic event. All symptoms were developed to be consistent with the diagnostic guidelines of BPD and cPTSD. A panel of content experts reviewed the vignette and confirmed the presence and absence of each symptom. In any instances where disagreement occurred, we revised the vignette to ensure that all symptoms were clearly presented and recognizable. The vignette met all diagnostic requirements for both BPD and cPTSD.

Diagnostic questions

Participants were asked to select a diagnosis for the client of cPTSD, BPD, or both, or they may provide a write‐in answer. Participants read a list of symptoms from the vignette and attributed each symptom listed to BPD, cPTSD, both, or neither disorder. They also rated their familiarity with BPD and cPTSD on a scale from 0 (not at all familiar) to 100 (very familiar).

Rank order

Participants were presented with a separate page where they were asked to rank the importance of the symptoms in their diagnostic decision. The page provided all of the symptoms included in the vignette. Participants ranked them starting with the most important.

Borderline patients‐ emotional attitudes inventory (BPD‐EA)

Bodner et al. (2011) developed the BPD‐EA to measure staff attitudes towards individuals with BPD. The BPD‐EA is a 20‐question measure with three subscales: negative emotions, experienced treatment difficulties and empathy towards the patient. The sub‐scales have demonstrated fair‐to‐good internal reliability in a previous sample of clinicians (α = .84, .75 and .60, respectively). Participants responded using a 5‐point Likert scale with 1 indicating “strongly disagree” and 5 indicating “strongly agree”. For the present study, items were modified to include first‐person language. The BPD‐EA was also used to assess participants' attitudes towards cPTSD; thus, the scale was modified by replacing “BPD” with “Complex PTSD” but otherwise keeping the item wording consistent. The order in which participants completed the BPD and cPTSD bias questions was randomized to control the possibility of order effects.

Demographic information

Participants were asked their age, gender identity, racial identity, profession, type of licensure, years of experience, if they see clients and if they supervise other clinicians. See Tables 1 and 2 for participant demographic breakdown.

Statistical analysis plan

SPSS 29.0 was used for all statistical analyses. Descriptive statistics (means, standard deviations and frequencies) were calculated to verify that data met the assumptions of the planned analyses.

A series of analyses were conducted to examine factors influencing clinicians' diagnostic decisions. Six multinomial logistic regressions were used to test whether (1) familiarity with the disorder predicted the diagnosis given; (2) clinician bias, measured via BPD‐EA scores for BPD and cPTSD, predicted diagnostic choice; (3) symptom rankings predicted the diagnosis assigned; (4) clinician stigma predicted symptom rankings; (5) symptom belongingness (i.e., the diagnosis to which participants attributed each symptom) predicted diagnostic decisions; and (6) clinician stigma predicted symptom belongingness. A paired‐samples t‐test was used to compare familiarity with BPD versus cPTSD. Finally, a forward stepwise regression identified which symptoms participants rated as most important in forming their diagnostic decisions.

RESULTS

Diagnostic decisions

Overall, clinicians diagnosed cPTSD (43.19%) more than BPD (12.68%), comorbid cPTSD and BPD (34.08%) and other diagnoses (10.09%). Clinicians who provided a diagnosis of “other” most often chose PTSD, substance use, or indicated that there was not enough information to make a diagnosis at this time. Confidence level ratings fell between 70% and 80% for each disorder with no significant difference between diagnoses, F(3, 418) = 1.65, p = .177, η ² = .012.

Familiarity with BPD and cPTSD

Clinicians indicated significantly less familiarity with BPD (M = 77.31, SD = 21.56) than cPTSD (M = 80.63, SD = 20.12), t(406) = −2.90, p = .004, d = 23.13. A multinomial logistic regression revealed that familiarity towards BPD did not significantly predict diagnostic decisions. However, clinicians who had higher familiarity with cPTSD were significantly less likely to diagnose BPD (χ ²(3) = 16.19, p < .001, Nagelkerke R ² = .042).

Familiarity with BPD predicted BPD stigma, F(1, 406) = 26.58, p < .001, R ² = .061 but not cPTSD stigma. Clinicians who were more familiar with BPD had lower BPD stigma. The more familiar clinicians were with cPTSD, the less stigma they had towards BPD (F(1, 405) = 15.61, p < .001, R ² = .037) and cPTSD (F(1, 406) = 57.29, p < .001, R ² = .124).

Stigma and diagnosis

Clinicians reported a higher level of stigma towards BPD (M = 2.24, SD = .57) than cPTSD (M = 1.92, SD = .46), t(419) = 14.63, p < .001, d = .71. Participants reported higher levels of stigma towards negative emotions associated with BPD (M = 1.86, SD = .66; cPTSD M = 1.54, SD = .60, t(419) = 12.89, p < .001, d = .63), greater feelings of treatment difficulty towards BPD (M = 2.62, SD = .74; cPTSD M = 2.16, SD = .60, t(418) = 14.15, p = .01, d = .69) and reduced empathy towards BPD (M = 2.49, SD = .52; cPTSD M = 2.32, SD = .52, t(416) = 7.66, p < .001, d = .38).

Stigma towards BPD did not significantly predict diagnostic decisions. In contrast, stigma towards cPTSD showed a significant overall model fit (χ ²(153) = 192.34, p = .017), explaining 40.1% of the variance (Nagelkerke R ² = .401). Individuals who gave a diagnosis of BPD only reported the highest stigma towards cPTSD (M = 2.02, SD = .49) compared to participants who gave other diagnoses: cPTSD (M = 1.91, SD = .43), comorbid (M = 1.93, SD = .48) and other (M = 1.84, SD = .46).

Rank order and belongingness

Table 3 presents the results of the rank ordering of symptoms based on diagnoses. In total, 401 (94.0%) participants engaged in this task by ranking at least one question. Across all diagnoses, the symptom that was most frequently ranked first (i.e., most important) was “when Kelly was ten, she was sexually abused by a next‐door neighbour” (n = 282; 66.20%). This symptom was ranked first by 79.9% of participants who diagnosed cPTSD, 24.1% of those who diagnosed BPD, 63.4% of those who diagnosed comorbid disorders and 63.4% of those who provided another diagnostic option.

TABLE 3.

Symptom rank order.

Symptom	CPTSD M (SD)	BPD M (SD)	Comorbid M (SD)	Other M (SD)	Total M (SD)	Frequency of 1st rank (%)
When Kelly was 10, she was sexually abused by a next‐door neighbour (CPTSD)	1.70a (2.70)	10.76ab (7.82)	3.08b (4.77)	3.80b (6.80)	3.51 (5.56)	70.32
John failed to call, and when he did the next day, Kelly met him with accusations of betrayal and screamed and swore at him (BPD)	12.08a (5.21)	3.34b (2.91)	5.78ab (5.31)	10.71a (6.45)	8.28 (6.14)	10.70
Kelly consistently avoids sexual activity with John as it makes her feel “icky” (CPTSD)	4.80a (3.29)	11.50c (6.16)	5.91ab (4.41)	6.10ab (4.70)	6.14 (4.74)	4.74
Kelly experienced feelings of worthlessness, shame, and disgust (CPTSD)	6.30a (3.30)	8.22b (5.58)	7.11ab (3.94)	5.50a (3.06)	6.74 (3.91)	3.49
After the relationship ended, Kelly reverted to her old pattern of drinking heavily and indiscriminately picking up men in bars. (BPD)	10.44a (4.54)	5.06c (3.42)	7.53ab (4.34)	9.00ab (5.05)	8.63 (4.76)	3.24
Kelly has woken up from a nightmare in which she was brutally attacked (CPTSD)	5.51a (4.33)	12.94c (5.99)	7.28b (4.57)	6.85ab (4.86)	7.18 (5.22)	1.75
Kelly said that her relationship with John was not the first time a relationship had ended “this way” (BPD)	19.66a (5.68)	10.16b (7.48)	16.26c (6.49)	19.58a (4.62)	17.30 (6.65)	1.50
Kelly states that she has never told anyone in the family about the abuse because she feels “too ashamed” (CPTSD)	7.74a (6.74)	16.26b (7.80)	11.53c (8.26)	8.03a (6.93)	10.13 (7.96)	1.50
Kelly, an intelligent and articulate woman who was an outstanding student and athlete in high school, says she has felt adrift for years (comorbid)	14.39a (5.52)	14.60ab (5.40)	14.36abc (5.42)	14.27abc (5.36)	14.27 (5.45)	1.25
The arguments between Kelly and John increased in frequency and intensified (BPD)	12.49a (4.69)	5.26b (3.64)	9.57c (5.23)	12.93a (5.88)	10.64 (5.46)	1.24
Kelly has always felt wary of her neighbours, and whenever she moves, she immediately has additional locks installed on the doors (CPTSD)	6.40a (5.30)	17.30c (6.50)	10.44ab (7.80)	9.23ab (7.45)	9.42 (7.52)	1.00
Two of the jobs Kelly quit were because embarrassment over angry outbursts directed at her supervisors (comorbid)	14.82a (4.06)	9.98b (5.28)	12.69c (4.64)	14.93a (4.64)	13.50 (4.77)	.50
Kelly reports experiencing frequent insomnia (CPTSD)	8.41a (4.42)	14.54b (5.60)	10.61c (5.05)	7.33a (4.17)	9.82 (5.21)	.50
John said that the only times Kelly looked animated had been when she was fighting with him (BPD)	20.13a (5.32)	15.60c (6.94)	19.55ab (5.30)	19.90ab (5.21)	19.34 (5.69)	.50
Kelly says she doesn't have much hope that she will stick with therapy and wonders if a person with problems such as hers is even capable of being helped (BPD)	18.26a (5.86)	16.44ab (6.05)	19.35ac (4.92)	17.43abc (5.36)	18.32 (5.60)	.50
Currently, Kelly lives with her parents but avoids speaking with them most days of the week (CPTSD)	14.25a (7.50)	18.30b (7.23)	16.85bc (7.98)	15.67ac (7.92)	15.67 (7.80)	.50
Kelly has quit three jobs in the past year (comorbid)	15.27a (4.39)	11.06b (5.70)	12.65bc (4.55)	14.15ac (4.53)	13.74 (4.87)	.25
Kelly became frightened of what might happen to her if she continued to have insomnia (comorbid)	12.90a (4.89)	16.16b (5.04)	15.53bc (4.84)	13.65ac (5.92)	14.28 (5.17)	0
Although she had been accepted for admission by several prestigious colleges, Kelly chose instead to work as a data entry operator and to pick up classes at a local community college (Neither)	16.56a (4.25)	15.76ab (5.24)	16.68abc (5.14)	16.25abc (4.89)	16.47 (4.75)	0
Kelly quit one of her jobs due to being uncomfortable with sexual innuendos made by her co‐workers (comorbid)	13.03b (5.41)	16.34a (5.07)	15.85ab (5.04)	16.10ab (5.35)	14.71 (5.42)	0
Kelly used to enjoy practicing piano and playing softball, but it has been several years since she has got any enjoyment from them (BPD)	16.17a (5.08)	19.56b (4.56)	17.32a (4.62)	13.93a (6.18)	16.76 (5.17)	.25
Kelly worries that she may never feel sufficiently motivated to determine a career path for herself (comorbid)	18.68a (4.18)	17.18ab (5.50)	19.32ac (3.87)	17.63abc (5.03)	18.61 (4.40)	0
Kelly stated that she thinks she will “probably die of AIDS by 40” so figuring out a career path is irrelevant (comorbid)	18.33a (5.23)	16.56ab (4.98)	17.87abc (5.15)	17.00abc (4.80)	17.82 (5.15)	0
John told Kelly that he could no longer take her jealous tantrums (BPD)	22.09a (3.43)	13.92b (7.38)	18.92c (5.77)	21.40a (3.93)	19.92 (5.63)	0
CPTSD combined symptoms	.83 (.44)	1.50 (1.79)	.39 (.76)		.87 (.60)
BPD combined symptoms	.81 (.80)	6.90 (1.75)	3.07 (1.55)		3.51 (2.84)
Comorbid combined symptoms	.96 (.72)	5.15 (2.11)	1.78 (1.46)		1.23 (1.50)

Note: Matching superscript letters within measure rows denote means that are equal.

We aimed to understand how participants ranked symptoms in relation to the disorders they are typically assigned. Symptoms were categorized as belonging to BPD, cPTSD or a combination of both based on our initial assessment of the symptoms within the vignette. Table 4 presents the regression results that examined whether participants' symptom rankings predicted the diagnosis they assigned. The higher participants ranked BPD symptoms, the more likely they were to give a BPD or comorbid diagnosis, and the less likely they were to give a cPTSD diagnosis (χ ²(3) = 171.64, p < .001, Nagelkerke R ² = .382). Next, the higher participants ranked cPTSD symptoms, the more likely they were to diagnose cPTSD and the less likely they were to diagnose BPD and comorbid (χ ²(3) = 176.83, p < .001, Nagelkerke R ² = .391). Finally, the higher comorbid symptoms were ranked, the more likely participants were to diagnose BPD (χ ²(3) = 12.26, p = .007, Nagelkerke R ² = .033).

TABLE 4.

Symptom rank order and diagnostic decision.

Disorder ranking	Diagnosis given	β	SE	95% CI		β	Sig.
				Lower bound	Upper bound
BPD	CPTSD	.192	.095	1.212	1.460	1.212	.044
	BPD	−.842	.125	.431	.551	.431	.001
	comorbid	−.276	.093	.759	.910	.759	.003
CPTSD	CPTSD	−.165	.069	.740	.972	.848	.018
	BPD	.599	.090	1.526	2.172	1.821	.001
	Comorbid	.171	.066	1.186	1.350	1.186	.010
Comorbid	CPTSD	.043	.084	.886	1.229	1.044	.610
	BPD	−.223	.102	.655	.977	.800	.029
	Comorbid	.007	.086	.852	1.191	1.007	.933

Stigma towards BPD did not significantly predict cPTSD or comorbid symptom rankings. However, stigma towards BPD did significantly predict BPD symptom rankings (χ ²(138) = 189.47, p = .002, Nagelkerke R ² = .386). cPTSD stigma did significantly predict cPTSD symptom rankings (χ ²(100) = 142.38, p = .003, Nagelkerke R ² = .308). cPTSD stigma also predicted BPD symptom rankings (χ ²(138) = 200.03, p < .001, Nagelkerke R ² = .404). However, cPTSD stigma did not predict comorbid symptom rankings.

A forward stepwise regression identified symptoms that participants ranked as most important when providing their diagnostic decision. The initial model, which includes only the intercept, demonstrated poor fit (−2 Log Likelihood = 464.66). The symptom “Kelly has always felt wary of her neighbours and whenever she moves, she immediately has additional locks installed on the doors” (a cPTSD symptom) was the first to be added, significantly improving model fit (Δ −2 Log Likelihood = 122.86, df = 69, p < .001). Adding a second symptom, “Kelly reverted to her old pattern of drinking heavily and indiscriminately picking up men in bars” (a BPD symptom), further enhanced the model (Δ −2 Log Likelihood = 119.636, df = 63, p < .001). No additional symptoms were added. The final model explained a substantial proportion of the variance in participants' rankings (Nagelkerke R ² = .793). However, the presence of singularities in the Hessian matrix suggests that some predictors may overlap in meaning, indicating that symptoms belonging to each diagnosis were largely redundant with each other in determining clinicians' diagnostic decisions.

Table 5 presents the results of symptom belongingness based on diagnoses. Table 6 presents the multinomial logistic regression results of whether participants' identified symptom belongingness (i.e., to which diagnosis they attributed the various symptoms) predicted the diagnosis they assigned. When participants attributed symptoms to BPD, they were more likely to diagnose BPD and comorbid disorders (χ ²(3) = 278.86, p < .001, Nagelkerke R ² = .503). However, when symptoms were grouped according to cPTSD, the model did not significantly align with the diagnosis given. Lastly, when clinicians attributed symptoms as comorbid, they were still more likely to diagnose BPD over cPTSD or comorbid (χ ²(3) = 187.581, p < .001, Nagelkerke R ² = .392). The most frequent selection by participants matched 7/8 cPTSD symptoms, 5/9 BPD symptoms, 5/7 comorbid symptoms and 1/1 symptom that fell into neither category.

TABLE 5.

Symptom belonging.

Symptom	CPTSD f (%)	BPD f (%)	Comorbid f (%)	Neither f (%)	Most commonly chosen
Kelly consistently avoids sexual activity with John as it makes her feel “icky” (CPTSD)	349 (82.70%)	3 (.71%)	58 (13.74%)	12 (2.84%)	Complex PTSD
John failed to call, and when he did the next day, Kelly met him with accusations of betrayal and screamed and swore at him (BPD)	40 (9.48%)	224 (53.08%)	146 (34.60%)	12 (2.84%)	BPD
The arguments between Kelly and John increased in frequency and intensified (BPD)	39 (9.24%)	182 (42.13%)	153 (35.26%)	48 (11.37%)	BPD
After the relationship ended, Kelly reverted to her old pattern of drinking heavily and indiscriminately picking up men in bars. (BPD)	69 (16.35%)	120 (28.44%)	223 (52.84%)	10 (2.37%)	Comorbid
Kelly called John several times—once to see if a woman answered, and occasionally when she was unable to sleep. (BPD)	35 (8.31%)	231 (54.87%)	111 (26.37%)	44 (10.45%)	BPD
Kelly has woken up from a nightmare in which she was brutally attacked (CPTSD)	380 (90.05%)	1 (.24%)	30 (7.11%)	11 (2.61%)	Complex PTSD
Kelly reports experiencing frequent insomnia (CPTSD)	315 (75.00%)	0 (0%)	80 (19.05%)	25 (5.95%)	Complex PTSD
Kelly experienced feelings of worthlessness, shame and disgust (CPTSD)	155 (36.73%)	17 (4.03%)	243 (57.58%)	7 (1.66%)	Comorbid
Kelly became frightened of what might happen to her if she continued to have insomnia (comorbid)	216 (51.43%)	12 (2.86%)	75 (17.86%)	117 (27.86%)	Complex PTSD
Kelly, an intelligent and articulate woman who was an outstanding student and athlete in high school, says she has felt adrift for years (comorbid)	124 (29.38%)	50 (11.85%)	165 (39.10%)	83 (19.67%)	Comorbid
Although she had been accepted for admission by several prestigious colleges, Kelly chose instead to work as a data entry operator and to pick up classes at a local community college (Neither)	119 (28.20%)	58 (13.74%)	107 (25.36%)	138 (32.70%)	Neither
Kelly has quit three jobs in the past year (comorbid)	40 (9.50%)	137 (32.54%)	189 (44.89%)	55 (13.06%)	Comorbid
Two of the jobs Kelly quit were because embarrassment over angry outbursts directed at her supervisors (comorbid)	58 (13.78%)	156 (37.05%)	192 (45.61%)	15 (3.56%)	Comorbid
Kelly quit one of her jobs due to being uncomfortable with sexual innuendos made by her co‐workers (comorbid)	270 (63.98%)	14 (3.32%)	84 (19.91%)	54 (12.80%)	Complex PTSD
Kelly used to enjoy practicing piano and playing softball, but it has been several years since she has got any enjoyment from them (BPD)	219 (51.90%)	15 (3.55%)	94 (22.27%)	94 (22.27%)	Complex PTSD
Kelly worries that she may never feel sufficiently motivated to determine a career path for herself (comorbid)	128 (30.40%)	44 (10.45%)	139 (33.02%)	110 (26.13%)	Comorbid
Kelly stated that she thinks she will “probably die of AIDS by 40” so figuring out a career path is irrelevant (comorbid)	117 (27.90%)	107 (25.42%)	132 (31.35%)	65 (15.44%)	Comorbid
Kelly says she doesn't have much hope that she will stick with therapy and wonders if a person with problems such as hers is even capable of being helped (BPD)	82 (19.48%)	91 (21.62%)	204 (48.46%)	44 (10.45%)	Comorbid
John told Kelly that he could no longer take her jealous tantrums (BPD)	27 (6.41%)	233 (55.34%)	99 (23.52%)	62 (14.73%)	BPD
John said that the only times Kelly looked animated had been when she was fighting with him (BPD)	86 (20.53%)	154 (36.75%)	119 (28.40%)	60 (14.32%)	BPD
Kelly said that her relationship with John was not the first time a relationship had ended “this way” (BPD)	41 (9.76%)	174 (41.43%)	192 (45.10%)	13 (3.10%)	Comorbid
When Kelly was 10, she was sexually abused by a next‐door neighbour (CPTSD)	285 (67.54%)	1 (.24%)	125 (29.62%)	11 (2.61%)	Complex PTSD
Kelly has always felt wary of her neighbours, and whenever she moves, she immediately has additional locks installed on the doors. (CPTSD)	363 (86.02%)	3 (.71%)	48 (11.37%)	8 (1.90%)	Complex PTSD
Kelly states that she has never told anyone in the family about the abuse because she feels “too ashamed” (CPTSD)	310 (73.46%)	9 (2.13%)	82 (19.43%)	21 (4.98%)	Complex PTSD
Currently, Kelly lives with her parents but avoids speaking with them most days of the week. (CPTSD)	141 (33.56%)	57 (13.57%)	120 (28.67%)	102 (24.90%)	Complex PTSD

TABLE 6.

Symptom belongingness and diagnostic decision.

Belonging disorder	Diagnosis given	β	SE	95% CI		β	Sig.
				Lower bound	Upper bound
BPD	CPTSD	−.190	.087	.697	.981	.827	.029
	BPD	1.079	.143	2.222	3.898	2.94	.001
	Comorbid	.577	.094	1.482	2.139	1.78	.001
CPTSD	CPTSD	.221	.297	.696	2.233	1.247	.458
	BPD	.172	.347	.592	2.384	1.188	.628
	Comorbid	.573	.305	.976	3.224	1.771	.060
Comorbid	CPTSD	−.312	.215	.480	1.116	.732	.147
	BPD	1.499	.233	2.836	7.071	4.478	.001
	Comorbid	.891	.202	1.639	3.625	2.438	.001

Stigma towards BPD significantly predicted BPD symptom belonging (χ ²(9) = 24.358, p = .004, Nagelkerke R ² = .057). However, BPD stigma did not predict cPTSD or comorbid symptom belonging. When looking at cPTSD stigma, it did significantly predict BPD symptom belonging (χ ²(4) = 28.744, p < .001, Nagelkerke R ² = .80), but not comorbid or cPTSD symptom belonging.

DISCUSSION

Clinicians most frequently diagnosed cPTSD, either as a standalone diagnosis or comorbid with BPD. Interestingly, the clinical vignette contained symptoms of both BPD and cPTSD, making a comorbid diagnosis the most accurate. However, the majority of clinicians underdiagnosed BPD by selecting cPTSD alone, overlooking BPD symptoms. One possible explanation for this pattern is that some clinicians may conceptualize cPTSD as a more appropriate and accurate diagnosis for BPD‐like symptoms, believing its symptoms were more reflective of trauma rather than personality pathology (Miller, 2014). This could suggest that clinicians in the current study intentionally or subconsciously avoided diagnosing BPD. Another explanation for the underdiagnosis of comorbid BPD and cPTSD is that some clinicians may not believe in the validity of comorbid mental health diagnoses (Maj, 2005), or may prefer single diagnostic models over comorbid frameworks (Nordgaard et al., 2023).

Additionally, there was no significant difference in clinicians' confidence levels across diagnostic decisions. Overall, clinicians reported high confidence, suggesting they strongly believed in their chosen diagnoses. However, if diagnostic decisions were incorrect but confidence remained high, this could indicate a lack of awareness of potential biases in decision‐making (Miller et al., 2015; Spengler et al., 2009).

Familiarity

Clinician familiarity with BPD and cPTSD was shown to have many impacts on stigma and diagnostic outcomes. Surprisingly, clinicians were significantly less familiar with BPD than with cPTSD. Clinicians' lack of familiarity with the BPD diagnosis is consistent with previous literature (Sisti et al., 2016). However, clinicians being more familiar with cPTSD than BPD was unexpected due to cPTSD's recent addition to the ICD‐11, and lack of representation in the DSM‐5. One explanation is that American mental health professionals may be more familiar with cPTSD than originally thought. Given that cPTSD was the most suggested disorder to be added to the ICD‐11 across the globe, it is possible that many U.S. clinicians are familiar with this disorder (Robles et al., 2014).

Familiarity with BPD did not predict diagnostic decisions, whereas familiarity with cPTSD did. Greater familiarity with BPD demonstrated lower stigma, consistent with prior research which suggests that increased familiarity with mental disorders is generally linked to reduced stigma (Corrigan & Nieweglowski, 2019; Potts & Henderson, 2020). Similarly, Miller (2014) found that clinicians more familiar with cPTSD viewed clients with both BPD and cPTSD more favourably. Familiarity with mental disorders has also been linked to lower perceived dangerousness (Link & Cullen, 1986), greater comfort in interactions (Corrigan et al., 2001) and increased empathy and prosocial attitudes (Angermeyer & Matschinger, 1996).

Stigma and diagnosis

The study also identified that clinicians reported higher stigma for BPD than for cPTSD, which is consistent with previous literature that BPD is associated with more negativity (Miller, 2014; Sisti et al., 2016). However, in terms of predicting diagnosis, stigma towards cPTSD predicted diagnosis, whereas stigma towards BPD did not. This may suggest that clinicians diagnosing BPD may have higher levels of stigma towards cPTSD, which could reflect a general bias or scepticism about the validity of cPTSD as a diagnosis (Resick et al., 2012). Another possibility is that participants in the current study may have indicated less variability in stigma towards BPD, which would limit the statistical ability to predict a significant outcome. Thus, we may not have seen the significance of stigma towards BPD due to the restriction of range. Another explanation may be that clinicians may be more open to diagnosing cPTSD than BPD due to the preconceived stigma associated with BPD (Miller, 2014).

Ranked symptoms and belongingness

Lastly, participants ranked symptoms from most helpful to least and categorized symptoms based on the disorder to which they felt it belonged. Across all diagnoses, the symptom ranked most helpful was related to a history of childhood sexual abuse, with a majority of those diagnosing cPTSD selecting this as the most important symptom. However, a notable proportion of participants who diagnosed BPD or comorbid conditions also prioritized this symptom, suggesting that trauma exposure is broadly considered relevant across multiple diagnoses. This finding highlights the potential difficulty in distinguishing BPD and cPTSD based on trauma history alone, furthering concerns about diagnostic overlap and potential misdiagnosis (Ford & Courtois, 2021). Symptom ranking also strongly influenced their respective diagnoses, and symptom rankings were largely redundant in this task (i.e., if they ranked one BPD symptom highly, they also ranked others highly). This has implications for clinical training as clinicians may be relying on heuristics, such as the representative heuristic, to make their diagnosis (Garb, 1998).

Additionally, we found that stigma influenced symptom ranking and symptom belonging. BPD stigma predicted the ranking of BPD symptoms, while cPTSD stigma predicted both BPD and cPTSD symptom rankings. This might suggest that previous attitudes towards BPD might reduce clinicians' ability to assess symptoms objectively (Aviram et al., 2006). Symptom belonging significantly predicted diagnostic decisions, specifically for BPD. Clinicians may be relying more heavily on symptom patterns associated with BPD when making differential diagnoses. We also found that clinicians correctly identified symptoms as belonging to cPTSD or both disorders but only identified about half of BPD symptoms. The symptoms clinicians most often missed were assigned as comorbid, as they related to impulsivity and relationship difficulties—features seen in both BPD and cPTSD. This suggests that clinicians may be unable to correctly identify symptoms of BPD, which might explain why cPTSD was diagnosed more frequently and BPD was underdiagnosed.

Limitations

A large limitation of this study was that it is difficult to get honest answers about stigma related to mental health disorders. While the survey was anonymous, clinicians may have been hesitant to disclose their true biases related to BPD and cPTSD. The Hawthorne effect is when individuals change their response due to being observed (Adair, 1984) and may have been present in this study. There are also limitations with using case vignettes to measure diagnostic decisions due to limited information and the inability to ask clarifying questions that may aid in diagnostic decisions.

Another limitation was that the study sample was volunteers; thus, those who chose to take the study may be fundamentally different than clinicians who chose not to take the study. Examples may include clinicians caring more about diagnosis, or the diagnosis of these specific disorders, being more familiar or comfortable with a variety of psychopathology, or being more conscientious. These differences could lead to a different pattern of results and could explain some of the patterns in the findings. Also, the demographics of the sample were primarily White cisgender females. Becker and Lamb (1994) found that female clinicians diagnosed PTSD over BPD more often than male clinicians. Thus, it is possible that the sample is not representative and may have bias towards the proposed psychopathology. Lastly, a limitation is that psychiatrists were not targeted in the recruitment strategy from this study. While this sample of mental health professionals is permitted to provide mental health diagnoses in Florida and Rhode Island, diagnostic‐decision patterns may differ among psychiatrists who may be the profession responsible for diagnosis in other world regions.

Future studies

Understanding clinicians' beliefs related to diagnosing BPD and cPTSD should continue to be studied. As mentioned earlier, a limitation of the present study is that the Hawthorne effect may have been present when clinicians were answering questions related to stigma. Finding another way to study stigma, like investigating reaction time to measure implicit bias (Karpinski & Hilton, 2001), may result in more accurate measurement. Future studies may also consider that the ICD‐11 has moved towards a dimensional model for personality disorders, which may reduce stigma towards BPD relative to the DSM‐5 (Monaghan & Bizumic, 2023). Future studies may use archival data or real‐life interviews to measure diagnostic decision‐making rather than a clinical vignette, which prohibits clinicians from gathering data that may help in their decision.

Implications

Understanding clinicians' beliefs related to diagnosing BPD and cPTSD is important in understanding how clinicians differentiate the two, recognize symptoms, and the role of stigma. These findings highlight the critical role of familiarity in shaping both diagnostic decisions and stigma, with familiarity with cPTSD having a stronger impact than familiarity with BPD. Further, training and education may play a role in how clinicians diagnose and their stigma. Research suggests that brief clinician training focused on knowledge and skills for working with individuals with BPD is associated with more positive attitudes towards this population (Krawitz, 2004). These shifts in attitudes may also have implications for clinical conceptualization and treatment planning, as trauma‐based frameworks such as cPTSD emphasize staged, trauma‐focused interventions that prioritize stabilization and emotion regulation, which may differ from approaches more commonly associated with BPD (Cloitre et al., 2014). Addressing this gap in training through the graduate or post‐graduate level may reduce diagnostic uncertainty and foster more accurate, unbiased clinical assessments. The potential for cognitive biases, such as the representativeness heuristic, to influence diagnostic decisions is an important consideration. Clinicians may unconsciously rely on the representativeness of symptoms to determine a diagnosis, particularly when symptoms overlap across disorders. Training in unconscious bias, which focuses on recognizing and mitigating cognitive biases, has been beneficial in changing clinician behavior (Ogunyemi, 2021) and could improve diagnostic accuracy. Further, anti‐stigma interventions have decreased negative attitudes in clinicians towards BPD and may be beneficial in diagnostic decision‐making (Knaak et al., 2015). Finally, the strong influence of symptom ranking on diagnosis indicates that clinicians may be overly influenced by salient symptoms rather than by an objective assessment of the entire clinical picture. This has significant implications for clinical training. Developing training programs that emphasize comprehensive assessments or reliance on symptom checklists could foster more accurate and holistic diagnoses.

CONCLUSION

The present study investigated beliefs related to diagnosing BPD and cPTSD, familiarity with each disorder, symptom ranking and symptom belongingness. Results identified the importance of understanding clinician bias in diagnosing BPD and cPTSD, with many related factors such as familiarity with each disorder. The findings underscore the importance of familiarity, training and awareness in shaping clinical decision‐making. As the understanding of both BPD and cPTSD continues to evolve, addressing the challenges posed by stigma, cognitive biases and diagnostic overlap will be crucial for improving diagnostic accuracy and treatment outcomes in clinical settings. Future studies should further explore training programs at the graduate and postgraduate level and other sources of bias, like implicit bias.

AUTHOR CONTRIBUTIONS

Bailey Pascuzzi: Conceptualization; investigation; writing – original draft; methodology; writing – review and editing; formal analysis; project administration; data curation. Jared Keeley: Conceptualization; investigation; methodology; writing – review and editing; formal analysis; project administration; data curation; supervision.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

ETHICS STATEMENT

Approved by Virginia Commonwealth University's institutional review board (HM20031318).

Supporting information

Data S1.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.