Acta Psychiatr Scand 2016: 1–7 © 2016 John Wiley & Sons A/S.

© 2016 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
All rights reserved ACTA PSYCHIATRICA SCANDINAVICA
DOI: 10.1111/acps.12678

Clinical vs. DSM diagnosis of bipolar

disorder, borderline personality disorder
and their co-occurrence
Bayes AJ, Parker GB. Clinical vs. DSM diagnosis of bipolar disorder, A. J. Bayes1, G. B. Parker1,2
borderline personality disorder and their co-occurrence. 1
School of Psychiatry, University of New South Wales,
Sydney, NSW, and 2Black Dog Institute, Randwick,
Objective: To investigate the extent and reasons contributing to NSW, Australia
discrepancies between those receiving a DSM as against a clinical
diagnosis of a bipolar disorder (BP) and/or a borderline personality
disorder (BPD).
Method: We interviewed participants previously receiving a BP or BPD
diagnosis, studying those who met DSM or clinical criteria for one or
both conditions. We compared the numbers of participants allocated to
the three diagnostic categories according to rater strategy to calculate
concordance rates and determine reasons for discordance.
Results: Rates of assignment to BP, BPD and comorbid BP/BPD varied
according to the diagnostic strategy. Concordance rates were reduced as
BP disorder duration criteria were relaxed, with discordance mainly
arising from clinical allocation of a BP disorder for those DSM assigned
as unipolar depression. Rates of BPD allocation varied marginally, with
discordance mostly arising from so clinically diagnosed receiving a Key words: bipolar disorder; borderline personality
disorder; comorbidity; diagnosis
comorbid BP/BPD DSM diagnosis. Finally, DSM overestimated
comorbidity compared with clinician diagnoses. Of central importance, Adam John Bayes, Hospital Rd, Randwick, NSW, 2031,
not imposing the DSM duration criteria for BP did not increase the Australia.
E-mail: a.bayes@unsw.edu.au
prevalence of misdiagnosing BPD, a finding at variance with the
literature.
Conclusion: Rates and reasons for discordance between clinical and
DSM diagnosis are detailed, which should assist clinical decision-making. Accepted for publication November 14, 2016

Significant outcomes
• DSM compared with clinical diagnosis overestimates comorbidity of BP/BPD, with discordant par-
ticipants instead clinically assigned as having BP, BPD or comorbid BP disorder with borderline per-
sonality traits.
• Rates of concordance reduced as BP disorder duration criteria were relaxed, with discordance pre-
dominantly from clinical allocation to BP disorder for those assigned a unipolar depressive DSM dis-
order. Not imposing the DSM duration criteria for BP did not increase the prevalence of
misdiagnosing BPD.
• Rates of BPD allocation varied marginally between groups, with discordance mostly arising from
clinical BPD being diagnosed by DSM criteria as having comorbid BP/BPD status.

Limitations
• There was only one rater for the structured diagnostic measures, with an absence of an independent
diagnostic rater.
• Clinical diagnoses were generated by one rater, and thus, the extent of potential clinician bias or diag-
nostic inaccuracy cannot be determined.
• The discordant diagnoses were determined with reference to clinical diagnoses, with the assumption
that the latter diagnoses were valid.

1
Bayes and Parker
Introduction
Differentiation of the bipolar disorders (BPs)
from borderline personality disorder (BPD) can
be a diagnostic challenge, particularly owing to
phenomenological features such as emotional
dysregulation (ED) and impulsivity being com-
mon to each condition (1) and leading to bipo-
lar disorder being misdiagnosed as BPD or the
converse. Inter-episode residual symptoms in
those with a BP condition, including chronic
dysphoria (2, 3), can resemble the chronic empti-
ness of BPD while periods of ED in individuals
with BPD can resemble hypomania (4), also
leading to misdiagnosis (5, 6). Perugi et al.’s (7)
finding of BP and BPD sharing a common
cyclothymic temperament might also obscure
diagnostic clarity. When both conditions are pre-
sent, diagnostic parsimony can operate with the
clinician seeking to identify the condition rather
than consider that both may be present. Con-
versely, pan-diagnostic features present in each
disorder may lead to an invalid diagnosis of
comorbidity.
Differing diagnostic criteria for bipolar disor-
ders may also have an impact on diagnosis. For
example, DSM-5 criteria mandate a 4-day mini-
mum duration to assign a diagnosis of hypomania
and 7 days for mania. By contrast, ‘bipolar spec-
trum’ models allow bipolar disorder to be diag-
nosed when briefer hypo/manic states are
experienced. The DSM-imposed duration criteria
for hypo/mania have been challenged in several
empirical studies [see (8)], and those not meeting
such criteria risk being invalidly assigned a BPD
diagnosis. While DSM-5 and ICD-10 provide for-
malized diagnostic criteria, their dominant
cross-sectional symptom set criteria (and with
each criterion viewed polythetically) offer a more
formulaic view to diagnosis than might be gener-
ated by an assessing clinician. Clinician-based
diagnoses have the advantage of ranging over a
wider set of candidate variables, including the
patient’s lifetime history and narrative. In addi-
tion, a patient’s behaviour during interview and
the resultant transference might assist with diag-
nosis, for example, a rapidly overvaluing or
devaluing interaction might indicate attachment
disturbance and a borderline rather than a bipolar
disorder (1). Disadvantages of clinician-based
diagnoses include their subjective risk, including
the possibility of ‘diagnostic expectation’ influ-
enced by the specific clinical setting, and thus
being open to potential biases and idiosyncratic
decisions made by the diagnostician. Indeed,
Saunders et al. (9) found clinician practices in
2
diagnosing BP and BPD were inconsistent and
often failed to address key symptom domains. In
clarifying the diagnostic dilemma of BP vs. BPD,
there is an advantage in moving beyond symptom
sets alone. For example, Parker (10) describes a
number of clinical features which weight a bipolar
disorder as against BPD and other states of emo-
tional dysregulation, including the presence of a
family history of bipolar disorder, a distinct clini-
cal onset or ‘trend break’ vs. pervasive emotional
problems from childhood, absence of distinct dys-
regulated personality style, more autonomous vs.
reactive mood episodes, melancholic vs. non-mel-
ancholic depressive episodes, ‘self’ vs. ‘other’
agency and response to a mood stabilizer vs. non-
specific medication effects. A set of symptom and
non-symptom clinical features strongly differenti-
ating of BP vs. BPD were delineated in our previ-
ous report (11), while we also demonstrated their
utility in differentiating those with one condition
from those who had comorbid BP/BPD condi-
tions (12, 13). These studies showed that the
choice of the diagnostic strategy (i.e. DSM vs.
clinical judgment) had a distinct effect on estimat-
ing the prevalence of the conditions, in differenti-
ating one from the other and in quantifying the
percentage of the sample who had comorbid diag-
noses. ‘Comorbidity’ is generally defined as the
coterminous presence of two independent condi-
tions. While DSM uses the term ‘comorbidity’, it
remains unclear whether this reflects the presence
of multiple distinct clinical entities or refers to
multiple manifestations of a single clinical entity
(14). Thus, ‘comorbidity’ might be artefactual due
to shared pan-diagnostic symptoms across psychi-
atric disorders, in this instance BP and BPD. Con-
versely, and as noted earlier, clinicians often
operate instead to a rule of parsimony, seeking to
derive one diagnosis for management, and are less
likely to make diagnoses of comorbid states. Both
factors risk compromising true estimates of the
prevalence of the individual and coterminous con-
ditions. A recent meta-analysis (15) quantified a
wide range in rates of comorbidity (i.e. 0–50% of
BPD in BP and 0–62.5% of BP in BPD) in 42
studies while also quantifying the prevalence of
BPD in those with bipolar disorder at 21.6% and
the converse at 18.5%.
Aims of the study
In this study, we sought to quantify rates of
discordance between clinician-based and DSM-
assigned BP, BPD and comorbid states and,
more importantly, to identify reasons for such
non-concordance.

Material and methods
As the study design has been reported in detail
(11), we only summarize key aspects. We sought
participants aged 18 years and older who had
received a prior diagnosis of BP (I or II) disorder
or BPD, recruiting participants from a number of
clinical services, including a tertiary referral
depression clinic, two public psychiatric hospital
in-patient services, three dialectical behaviour
therapy out-patient clinics, a public hospital out-
patient psychotherapy clinic and three private psy-
chiatric clinics. In addition, an invitation placed on
the institute’s Volunteer Research Register, as well
as advertisements placed on our clinical institute’s
website and in newspapers. Site ethical approval
was ratified by the University of New South Wales
Human Research Ethics Committee, and written
consent was obtained from all participants. Exclu-
sion criteria included English language limitations,
psychotic or significant organic features and cur-
rent substance dependence. Data were collected
over the period April 2012 to December 2014.
A detailed semistructured interview was under-
taken by a psychiatrist (A.B.) experienced in assess-
ing patients with mood and borderline personality
disorders. The diagnostic aim of the interview was
to establish the presence or absence of a BP I, BP II
and/or BPD diagnosis, while reasons for assigning
each clinical diagnosis were documented. Study
‘DSM’ diagnoses respected DSM-IV criteria for
both conditions. In practice, a DSM-IV BP I or BP
II disorder diagnosis was determined by administer-
ing the Mini International Neuropsychiatric Index
(MINI) (16) to all participants. A DSM-IV diagno-
sis of BPD was generated after administering the
Diagnostic Interview for Personality Disorders IV
(DIPD-IV) – Borderline Personality Disorder sec-
tion (17) to all interviewees, and those scoring two
on five or more of the nine criteria received a DSM-
IV BPD diagnosis.
Our ‘clinical diagnoses’ were generated from a
comprehensive interview which, in addition to
respecting DSM-IV symptom criteria, also consid-
ered general history data. For clinical judgment of
a bipolar disorder, consideration was given to (i)
the illness course, including whether there was a
distinct onset or ‘trend break,’ (ii) cyclical mood
changes with inter-episode remission, (iii) the phe-
nomenology of ‘highs’ and (iv) the absence of a
history of pervasive instability in relationships
(and its absence in the interviewee’s interpersonal
behaviour during the clinical assessment).
If the psychiatrist judged a bipolar disorder as
present, diagnostic subtyping and allocation
occurred – a clinical diagnosis of a BP I disorder
Clinical vs DSM diagnosis of BP, BPD or both
was assigned on the basis of a diagnosed BP partic-
ipant describing psychotic features during previous
manic states, while a clinical BP II diagnosis
required the absence of psychotic features during a
high. We generated two clinical BP subgroups (i) a
‘BP CLIN STRICT’ group comprised those partic-
ipants receiving a clinical bipolar (BP I or BP II)
diagnosis and with DSM duration criteria (i.e.
hypomania lasting 4 or more days; mania lasting 7
or more days) respected, and (ii) a ‘BP CLIN
EXTEND’ group that also included (along with
those in the ‘STRICT’ group) those reporting
briefer hypo/manic episodes than stipulated by
DSM.
A BPD ‘clinical’ diagnosis was determined by
the clinician weighting DSM-IV criteria but also
considering corroborative referrer information if
available. For a clinical diagnosis of BPD, addi-
tional consideration was given to (i) the illness
course – with the presence of pervasive mood diffi-
culties present since a young age and thus without
a clear onset or ‘trend break’, (ii) the pervasive
presence of typical borderline personality features
that were independent of mood episodes, and
included impulsivity, relationship difficulties or
anger (not occurring during a hypo/manic epi-
sode), or emptiness, identity disturbance and delib-
erate self-harm (not necessarily occurring during a
depressive episode), (iii) pervasive instability in
relationships and which might also be evidenced in
the interpersonal behaviour of the participant dur-
ing the clinical interview. Clinician diagnosis also
allocated ‘disorder’ status or ‘trait’ status to those
judged as having a borderline personality, reflect-
ing the judged presence or absence of functional
impairment.
DSM-defined diagnoses were compared with
both CLIN STRICT and CLIN EXTEND diag-
noses. Concordance was defined as the DSM diag-
nosis agreeing with clinical diagnosis (e.g. DSM
BP = CLIN STRICT BP). Discordance was
defined as a DSM diagnosis disagreeing with a
clinical diagnosis (used as the standard).
Results
DSM allocated 82 participants (44%) to a BP diag-
nosis, increasing to 98 (55%) using CLIN STRICT
criteria and 125 (61%), using broader CLIN
EXTEND criteria (see Table 1). Concordance
rates are detailed in Table 1. For bipolar disorder,
using CLIN STRICT defined criteria, there was a
76% concordance rate with DSM, reducing to
62% when CLIN EXTEND criteria were applied.
Assuming our reference clinical diagnoses are
valid, we now explore discordance with DSM
3
Bayes and Parker

Table 1. Allocation of BP, BPD and comorbid BP/BPD according to diagnostic strategy and associated concordance rates

CLIN STRICT concordance CLIN EXTEND concordance

DSM CLIN STRICT CLIN EXTEND with DSMa with DSMb
N (%) N (%) N (%) N (%) N (%)

BP 82 (44%) 98 (55%) 125 (61%) 74 (76%) 78 (62%)

BPD 53 (28%) 58 (33%) 52 (25%) 46 (79%) 42 (81%)
Comorbid 52 (28%) 23 (13%) 29 (14%) 20 (87%) 22 (76%)
Total 187 (100%) 179 (100%) 206 (100%) 140 (79%) 142 (69%)

BP, bipolar disorder; BPD, borderline personality disorder; comorbid, bipolar disorder and borderline personality disorder; CLIN STRICT, those receiving a clinical diagnosis of a
bipolar disorder and meeting DSM duration criterion; CLIN EXTEND, those with a clinical diagnosis of a bipolar disorder but not necessarily meeting the DSM minimum duration
criterion.
a
Percentage of clinically diagnosed participants in the CLIN STRICT sample with the same DSM and clinical diagnosis.
b
Percentage of clinically diagnosed participants in the CLIN EXTEND sample with the same DSM and clinical diagnosis.

decisions. For the CLIN STRICT bipolar group, groups. In the CLIN STRICT sample, 12 partici-
two were assigned a BPD diagnosis, two were diag- pants who were clinically diagnosed with a BPD
nosed as having unipolar depression, two failed to were discordant with DSM-defined BPD – 11
receive a DSM diagnosis, and 18 were assigned a participants assigned to comorbid BP/BPD and
comorbid BP/BPD diagnosis. Of the latter group, one assigned to unipolar depression. In the CLIN
nine were clinically diagnosed as having borderline EXTEND sample, 10 participants clinically diag-
personality traits (but not the disorder) in addition nosed with a BPD were discordant with DSM-
to bipolar disorder. Applying CLIN EXTEND cri- defined BPD – nine assigned to comorbid BP/BPD
teria for BP, the proportion discordant with DSM and one assigned to unipolar depression. The sub-
rose to 38%, with DSM allocating five a BPD set of those clinically diagnosed with a sole BPD,
diagnosis, 20 were assigned as having unipolar but receiving a DSM-defined diagnosis of comor-
depression, three did not receive a DSM diagnosis, bid BP/BPD, most commonly was due to the
and 19 were assigned a comorbid BP/BPD diagno- suggested hypo/manic symptoms more reflecting
sis. Of the latter group, nine were clinically diag- mood reactivity secondary to the external environ-
nosed as having borderline personality traits in ment, immature or regressed behaviours, disinhibi-
addition to having a BP disorder. tion, irritability or anger related to alcohol misuse
The subsets clinically diagnosed with a bipolar or being part of a ‘manic defence’ with one partici-
disorder only (CLIN STRICT or CLIN pant describing the latter state as ‘feeling fake’.
EXTEND), but receiving a DSM diagnosis of For comorbid BP/BPD, DSM allocated 52 par-
comorbidity, were generally a consequence of ticipants (28%), decreasing to 23 (13%) using
reported ‘borderline’ personality features being CLIN STRICT and 29 (14%) using CLIN
only mood state related and not enduring (e.g. in EXTEND criteria. Utilizing CLIN STRICT crite-
such clinically diagnosed BP participants, ‘empti- ria, there was 87% concordance with DSM, reduc-
ness’ and transient paranoid ideation or dissocia- ing to 76% when CLIN EXTEND criteria were
tion was only reported during depressive episodes, applied. The discordant individuals received the
impulsivity occurred only during highs, anger was following DSM diagnoses: two were assigned to a
mood state dependent – when depressed or high – sole BP disorder (in both the CLIN STRICT and
or symptoms such as emptiness disappeared when CLIN EXTEND groups), with a sole BPD
the participant had been treated with a mood sta- assigned to one participant in the CLIN STRICT
bilizer). The subsets clinically diagnosed as having group and five participants in the CLIN EXTEND
BP disorder with comorbid borderline personality group. In the CLIN STRICT group, the increased
traits, but receiving a DSM diagnosis of comorbid- DSM BP/BPD allocations are accounted for by
ity, reflected the clinician judging insufficient func- DSM reclassification of nine BP, 13 BPD and nine
tional impairment to warrant ‘disorder’ status. BP/borderline personality trait participants. In the
For borderline personality disorder, 53 (28%) CLIN EXTEND group, DSM reclassified as
participants were allocated by DSM, 58 (33%) comorbid BP/BPD, 10 BP, nine BPD and nine BP/
were clinically allocated in the CLIN STRICT borderline personality trait participants.
sample and 52 (25%) were clinically allocated in
the CLIN EXTEND group. Using clinical diagno-
Discussion
sis compared with DSM, there were very similar
concordance rates of 79% and 81%, respectively, Study limitations and strengths are first noted.
for the CLIN STRICT and CLIN EXTEND Participants were volunteers and, thus, not

4

necessarily representative, and we might assume
weighted towards those with less severe disorders.
For the majority of participants, there was only
one rater for the structured DSM diagnoses –
thus, in the absence of an independent diagnostic
rater – there is the potential of measurement
error. Similarly, clinical diagnoses, including the
CLIN STRICT and CLIN EXTEND categories,
were generated by one rater and thus risk clini-
cian bias or diagnostic inaccuracy. In addition,
the discordant diagnoses were predicated on the
assumption that the clinician diagnosis was valid
and thus may limit generalizability of the find-
ings. Potential clinician rater biases include diag-
nostic expectation related to the differing
recruitment portals, while participants may have
biased their reporting of symptoms in line with
the prior diagnosis they had received. The cross-
sectional nature of the study, meant that the
diagnostic stability of the disorders over time
could not be established. Future studies would
ideally utilize more than one rater to quantify the
extent of interrater variability on the structured
measure and more than one clinician to generate
clinician diagnoses and quantify respective levels
of interrater agreement. One of the study’s
strengths was that it controlled for interrater
variability with respect to comparison of struc-
tured and clinical diagnoses by the use of only
one rater. A further strength is the use of three
differing diagnostic strategies which allowed eval-
uation of how varying the decision rules might
impact on diagnostic allocation.
Turning to study findings, we first examined the
influence of utilizing DSM-defined BP disorder vs.
two clinical diagnoses (CLIN STRICT and CLIN
EXTEND), with an increased rate of discordance
as the BP definition broadened. A key finding, and
at variance with a common view [e.g. Zimmerman
and Morgan (18)], was that by defining bipolar dis-
order broadly – in not of necessity imposing DSM
minimum hypo/mania duration criteria – there
was no increase in the prevalence of misdiagnosing
borderline personality disorder. The most distinc-
tive increase of 20 (24%) additional diagnoses of
bipolar disorder arose from where DSM instead
diagnosed major depressive disorder. Relaxing the
minimum duration criteria for hypo/mania can
inflate the number of bipolar disorder diagnoses by
reclassifying DSM-defined major depressive disor-
der as bipolar disorder. As has been reported pre-
viously, duration of hypo/manic states as a means
of ruling out BP disorder is unlikely to be useful as
the minimum DSM duration criteria appear inva-
lid (8). Thus, rather than duration, consideration
of phenomenology of the hypo/manic state
Clinical vs DSM diagnosis of BP, BPD or both
including episodes being more likely to be sponta-
neous (vs. reactive), associated with reduction
rather than an increase in anxiety, and additional
corroborative information might enable greater
confidence in detecting true hypo/manic episodes
and affirming or rejecting the presence of a bipolar
disorder. Our previous study (12) found BP, BPD
and comorbid groups did not differ on a range of
symptoms related to depression history (e.g. family
history of depression, age of first depressive epi-
sode, number of deliberate self-harm/suicide
attempts, antidepressant effectiveness), so that dif-
ferentiation based on depression history is unlikely
to be helpful.
Similar numbers of BP participants in both the
CLIN STRICT and CLIN EXTEND groups were
allocated by DSM to comorbid BP/BPD. Where
DSM diagnosed comorbidity, the clinician judged
that the apparent ‘borderline’ symptoms could be
explained solely by the presence of a bipolar disor-
der. Thus using a hierarchical approach, ‘state’
borderline phenomena described in the history
were considered as having occurred secondary to
bipolar mood episodes. Therefore, correlation of
‘personality’ symptoms with either depressive or
hypo/manic episodes or improvement with mood
stabilizer treatment was able to clarify a bipolar
disorder diagnosis.
In both CLIN STRICT and CLIN EXTEND
groups, approximately one-half of BP participants
allocated by DSM as comorbid were considered
clinically to have comorbid borderline personality
traits (i.e. were falsely assigned personality ‘disor-
der’ status). These participants described comorbid
personality dysfunction outside of discrete mood
episodes, but without sufficient severity to impact
on their functioning. By delineating personality
traits vs. disorder, this can assist in sequencing of
treatment, whereby priority would be given to
treating the bipolar illness with a mood stabilizer,
and management of the personality traits, if
needed, considered secondary.
For borderline personality disorder, rates of
assignment varied marginally (25–33%) across the
three groups. Assessment of discordant diagnoses
reveals the majority of clinically defined BPD in
both the CLIN STRICT and CLIN EXTEND
samples were allocated to comorbid BP/BPD by
DSM criteria. The additional ‘false-positive’ bipo-
lar disorder and the apparent ‘bipolar’ symptoms
were determined by the clinician to be explained
solely by the presence of a borderline personality
disorder. Clinically, this scenario arose when the
affirmed ‘high’ was not judged to be truly hypo-
manic but considered to be a situationally driven
affective change reactive to the external
5
Bayes and Parker
environment, ‘immature’ or regressed behaviour,
conceptualized as a ‘manic defence’ to ward off
uncomfortable affects or related to emotional dys-
regulation, anger or giddiness. We next investi-
gated classification of DSM-defined and clinically
defined comorbid BP and BPD. The increased pro-
portion of DSM allocation to BP/BPD compared
with CLIN STRICT and CLIN EXTEND is made
up of DSM assigning comorbidity to approxi-
mately one-third with a bipolar disorder only, one-
third having a borderline personality disorder only
and one-third with bipolar disorder and borderline
personality traits. It is likely DSM overestimates
comorbidity because it uses a symptom-based
approach whereas a clinician may overcome any
such artefact created by pan-diagnostic features
using a hierarchical diagnostic approach. Con-
versely, when a clinician might diagnose a ‘pure’
BP or BPD and DSM criteria indicated their
comorbidity, such findings might reflect either the
clinician missing a truly comorbid presentation or
diagnostic overshadowing obscuring identification
of the coterminous conditions.
Our sample members were generally not facing
any acute mood-perturbing disturbance when
interviewed, but a clinical dilemma may arise if
assessment is made when the patient is experienc-
ing a crisis or at episode nadir, and risking invalid
assessment of personality if weighted on cross-
sectional reporting. During a hypo/manic episode,
especially an ‘irritable’ high or a mixed affective
state, a bipolar patient might display classic bor-
derline features (e.g. challenging interpersonal
interactions, anger and prominent emotional dys-
regulation) and risk a false-positive BPD diagno-
sis. Conversely, a patient experiencing bipolar
depression might present having self-harmed,
report emptiness, identity disturbance and suicidal-
ity and also risk false-positive assignment of a
BPD diagnosis. However, with longitudinal obser-
vation and resolution of the mood episode, the
apparent personality disruption may dissipate and
be recast as a secondary phenomenon. Addition-
ally, seeking collateral history from a carer or rela-
tive to determine whether such behaviours are ‘out
of character’, and occur only episodically, will
sharpen diagnostic accuracy.
Clinical delineation of comorbid BP and BPD
is important, as the two conditions require differ-
ing treatment modalities – respectively, a mood
stabilizer and a psychotherapy being prioritized.
False assignment can lead to the unnecessary pre-
scribing of a mood stabilizer for a bipolar condi-
tion when a borderline condition is present, while
invalidly diagnosing an individual with a bipolar
condition as having borderline personality
6
disorder risks a pejorative diagnosis as well as
inappropriate management. Clinical acumen
remains an important tool of psychiatric diagno-
sis as it utilizes both a broader set of clinical fea-
tures and longitudinal history data, whereas
DSM diagnoses focus on polythetic symptom cri-
teria. This study – which not only identifies rates
of discordance in assigning BP, BPD and comor-
bid states – identifies factors that can contribute
to diagnostic error and markers that can advance
diagnostic precision in the assessment of these
two psychiatric conditions and their comorbid
expression.
Acknowledgements
The study was supported by NHMRC Program Grant
(1037196) funding. We thank those who assisted with study
recruitment: Amelia Paterson, Dr Rebecca Graham, Stacey
McGraw, Rosemary Clancy, Jonathan Levy, Dr Rene de
Monchy, Dr Michael Hong, Dr Caryl Barnes, Wayne
Borg, Dr Farzaan Mehta, Robert Glassick, and Dr Laurie
Power.
Declaration of interest
There are no conflict of interests to be disclosed.
References
1. Phelps J. A Spectrum Approach to Mood Disorders. New
York, NY: W. W. Norton & Company, 2016.
2. Paykel ES, Abbott R, Morriss R, Hayhurst H, Scott J.
Sub-syndromal and syndromal symptoms in the longitudi-
nal course of bipolar disorder. Brit J Psychiatry
2006;189:118–123.
3. Judd LL, Akiskal HS, Schettler PJ et al. A prospective
investigation of the natural history of the long-term weekly
symptomatic status of bipolar II disorder. Arch Gen Psy-
chiatry 2003;60:261–269.
4. Kernberg OF, Yeomans FE. Borderline personality disor-
der, bipolar disorder, depression, attention deficit/hyper-
activity disorder, and narcissistic personality disorder:
practical differential diagnosis. Bull Menninger Clin
2013;77:1–22.
5. Ruggero CJ, Zimmerman M, Chelminski I, Young D. Border-
line personality disorder and the misdiagnosis of bipolar
disorder. J Psychiatr Res 2010;44:405–408.
6. Zimmerman M, Ruggero CJ, Chelminski I, Young D. Psychi-
atric diagnoses in patients previously overdiagnosed with
bipolar disorder. J Clin Psychiatry 2010;71:
26–31.
7. Perugi G. Are atypical depression, borderline personality
disorder and bipolar II disorder overlapping manifesta-
tions of a common cyclothymic diathesis? World Psychia-
try 2011;10:45–51.
8. Parker G, Graham R, Synnott H, Anderson J. Is the DSM-
5 duration criterion valid for the definition of hypomania?
J Affect Disord 2014;156:87–91.
9. Saunders KE, Bilderbeck AC, Price J, Goodwin GM.
Distinguishing bipolar disorder from borderline personal-
ity disorder: a study of current clinical practice. Eur Psy-
chiatry 2015;30:965–974.

10. Parker G. Clinical differentiation of bipolar II disorder
from personality-based “emotional dysregulation” condi-
tions. J Affect Disord 2011;133:16–21.
11. Bayes AJ, McClure G, Fletcher K et al. Differentiating
the bipolar disorders from borderline personality disorder.
Acta Psychiatr Scand 2016;133:187–195.
12. Parker G, Bayes A, McClure G, del Moral YR, Stevenson
J. Clinical status of comorbid bipolar disorder and border-
line personality disorder. Brit J Psychiatry 2016;209:209–
215.
13. Bayes A, Parker G, McClure G. Emotional dysregulation
in those with bipolar disorder, borderline personality dis-
order and their comorbid expression. J Affect Disord
2016;204:103–111.
14. Maj M. “Psychiatric comorbidity”: an artefact of current
diagnostic systems? Brit J Psychiatry 2005;186:182–184.
Clinical vs DSM diagnosis of BP, BPD or both
15. Fornaro M, Orsolini L, Marini S et al. The prevalence and
predictors of bipolar and borderline personality disorders
comorbidity: systematic review and meta-analysis. J Affect
Disord 2016;195:105–118.
16. Sheehan DV, Lecrubier Y, Sheehan KH et al. The Mini-
International Neuropsychiatric Interview (M.I.N.I.): the
development and validation of a structured diagnostic psy-
chiatric interview for DSM-IV and ICD-10. J Clin Psychi-
atry 1998;59(Suppl 20):22–33; quiz 4-57.
17. Zanarini M. The Diagnostic Interview for DSM-IV Per-
sonality Disorders. Belmont, MA: McLean Hospital and
Harvard Medical School, 1996.
18. Zimmerman M, Morgan TA. Problematic boundaries in the
diagnosis of bipolar disorder: the interface with borderline
personality disorder. Curr Psychiatry Rep 2013;15:
422.