Implementing cognitive screenings for outpatients with bipolar disorder following optimised treatment in a specialised mood disorder clinic

Bipolar disorder (BD) is often accompanied by persistent cognitive impairment. However, screening for cognitive impairment in the clinic is challenged by a lack of consensus on screening procedures. This study assesses cognitive impairment prevalence and screening feasibility in alignment with the International Society for Bipolar Disorder Targeting Cognition Task Force recommendations. Between January 2022 and May 2023, 136 newly diagnosed BD outpatients were assessed with the Screen for Cognitive Impairment in Psychiatry after 15 – 20 months of specialised care at the Copenhagen Affective Disorder Clinic. Cognitive impairment patterns and associations with cognitive complaints, perceived stress, and functioning were examined. Most screened patients (73 %) achieved full or partial remission, with 51 % being cognitively normal, 38 % showing global impairments, and 11 % displaying selective impairments. Among remitted patients, 56 % were cognitively normal, while 31 % and 13 % exhibited global or selective impairments, respectively. Both objectively impaired patient groups re-ported more subjective cognitive difficulties than those who were cognitively normal. The globally impaired group also demonstrated poorer functioning, more depressive symptoms and lower quality of life than cognitively normal patients. Across all patients, lower cognitive performance correlated with more cognitive complaints, lower functioning, lower quality of life, and more depressive symptoms. Cognitive screenings were relatively easily implementable, involving only a 1.5 h session including mood ratings, feedback and cognitive strategy discussion. The study highlights the clinical relevance and feasibility of cognitive screenings in BD patients, emphasizing the need for tailored interventions given frequent cognitive impairment in clinically stable individuals.


Introduction
Cognitive impairment, a core symptom in various psychiatric illnesses, has emerged as a key treatment target for enhancing functional recovery.In individuals with bipolar disorder (BD), moderate cognitive impairments have been observed across multiple domains, including attention, verbal memory, and executive functions, when compared to norms, both during and between mood episodes (Bora et al., 2013;Bourne et al., 2013).However, the impairments occur to varying degrees and with substantial inter-individual heterogeneity (Bourne et al., 2013), which can be explained the existence of underlying subgroups.Indeed, several studies of remitted BD patients indicate the existence of three distinct neurocognitive subgroups: one with global impairments (12-37 %), another with selective moderate impairments-typically in verbal memory and/or attention-(30-40 %), and a third with relatively normal cognitive functioning (33-48 %) (Burdick et al., 2014;Jensen et al., 2016;Kjaerstad et al., 2021;Solé et al., 2016).Importantly, patients who exhibited either global or selective cognitive impairments reported lower quality of life, increased perceived stress, and diminished vocational function compared to those who were cognitively spared, even when exhibiting similar levels of residual mood symptoms (Jensen et al., 2016;Solé et al., 2016).These findings concord with identified links between cognitive impairment and functional outcomes, quality of life, persistent affective symptoms, and implications for future prognosis in patients with mood disorders (Bonnín et al., 2012;Miskowiak et al., 2018;Rosa et al., 2009;Sankar et al., 2023;Tse et al., 2014).
Suffering from undetected cognitive impairment during remission following initial treatment is stressful for the individual patient leaving them unprepared and puzzled when challenged by daily life, education, and work.Thus, the inter-individual cognitive variability and the associations between cognitive impairments and poor psychosocial outcomes and prognosis warrant broad implementation of cognitive screening in the clinical management of BD.However, to achieve this goal there is a need for consensus on the screening and management of cognitive impairment in clinical practice.The International Society for Bipolar Disorders (ISBD) Targeting Cognition Task Force therefore developed a set of consensus-based clinical recommendations for clinicians to guide their choices on whether, when and how to assess and address cognition in their patients (Miskowiak et al., 2018).Key recommendations were to: (i) conduct formal cognition assessments for all patients in remission, (ii) use brief tools like Screen for Cognitive Impairment in Psychiatry (SCIP) (Purdon, 2005) and Cognitive Complaints in Bipolar Disorder Rating Assessment (COBRA) (Rosa et al., 2013), freely available in multiple languages through the ISBD website, (iii) consider medication, comorbidity, and symptoms when detecting impairment, refer for comprehensive neuropsychological evaluation when necessary, and use assessments for encouraging compensatory strategies and promoting cognitive reserve.
Implementing ISBD Targeting Cognition Task Force recommendations, we conducted cognitive screenings with SCIP and COBRA for newly diagnosed BD outpatients at the Copenhagen Affective Disorder Clinic (Kessing et al., 2013) in 2022-2023.
The aims of the current study were to: (I) examine the prevalence, severity and pattern of residual cognitive impairments among recently diagnosed BD patients following 15-20 months of optimised pharmacological treatment and group-based psychoeducation in a specialised, multidisciplinary hospital outpatient clinic, and (II) investigate the implications of residual cognitive impairments by comparing clinical characteristics of patients with and without impairments and by exploring associations between patients' cognitive status and functioning, subjective cognitive complaints, quality of life, and clinical variables.

The copenhagen affective disorder clinic, participants, and recruitment
The Copenhagen Affective Disorder Clinic, Mental Health Services, Capital region of Denmark, was established in 2004 based on positive findings from a pragmatic randomised controlled trial (Kessing et al., 2013) and covers the entire population of 1.8 million people in the Capital Region of Denmark.A total of 18 fulltime clinicians are employed in the Copenhagen Affective Disorder Clinic (7 specialists in psychiatry, nurses, psychologists, a halftime social adviser and physiotherapist) treating approximately 300 newly diagnosed patients with bipolar disorder yearly in addition to other groups of patients with BD.Newly diagnosed BD is defined as less than two years since diagnosis or age below 30 years and patients are referred to the Clinic by GPs, MDs, and specialists in psychiatry.The diagnosis is confirmed/disconfirmed by specialists in psychiatry in the Clinic following weekly conferences with regular attendance by senior specialists in psychiatry (including LH and LVK).
Outpatients with newly diagnosed BD were offered a cognitive screening with verbal and written feedback as part of their two-year treatment program in the Copenhagen Affective Disorder Clinic between January 2022 and May 2023.Patients were offered a cognitive screening between 15 and 20 months after their enrollment in the clinic, towards the end of their specialised care which involved optimization of pharmacological treatment, group-based psychoeducation over 16 weeks and, when needed, additional individual care from nurses, psychologists, a physiotherapist, and a social worker.For the recruitment strategy, patients of all ages, including patients with psychiatric or somatic comorbidities, who had been receiving treatment at the clinic for 15-20 months were contacted, assuming that the majority were in partial or full remission given their optimised treatment duration.However, it was acknowledged that even patients not in remission could benefit from cognitive screening at this point in their specialised treatment, as cognitive impairments could contribute to mood symptom instability and hinder treatment efficacy.
The study protocol included contacting the local ethics committee in the Mental Health services, Capital Region of Denmark, via email to inquire about the necessity of ethics approval.In response, the committee provided a written statement stating that formal ethics approval was not required as the study involved no biomedical devices or invasive procedures and thus posed minimal risks for participants.Nonetheless, in accordance with the protocol, all participants provided written informed consent, demonstrating their voluntary agreement to participate after receiving information about the study's purpose, procedures, and potential risks.The study received approval in compliance with current GDPR regulations (P-2022-356), and data storage adhered to secure web database regulations in REDCap.All patients provided written informed consent.

Procedures
Master's level psychology students, supervised by KWM, administered cognitive screenings, including the SCIP, clinical mood symptom and functioning ratings, and questionnaires on subjective cognition and quality of life.During screening, patients promptly received cognitive performance feedback and psychoeducation about BD-related cognitive impairments, linking them to symptoms, illness, medication, and lifestyle factors.Assessors then connected screening results to real-life challenges, discussed strategies for coping with cognitive difficulties, and gave them a copy of the ISBD Cognition Booklet that was made freely available in multiple languages on the ISBD homepage (https:// www.isbd.org/TaskForcePatientResources).On average, the screening and feedback sessions lasted 1.5 h.Following screening sessions, assessors, supervised by KWM, composed written feedback reports that were shared with patients and clinicians via electronic case files.A feedback session case is included in the supplementary materials.

Materials
The objective cognitive measure SCIP consists of five subtests measuring verbal learning (VLT-L), working memory (WMT), verbal fluency (VFT), verbal memory (VLT-D) and psychomotor speed (PMT) (Ott et al., 2021).Scores on the SCIP were analysed in relation to Danish norms (Ott et al., 2021).Regression-based formulas were employed to predict each patient's expected SCIP scores based on their age, sex, and years of education (Ott et al., 2021).The regression-based formula enables precise estimation for each patient of whether their cognition scores deviate from or align with the expected scores of a person with matched demographic characteristics.Subjective cognitive complaints were measured with the Cognitive Complaints in Bipolar Disorder Rating Assessment (COBRA) with scores ≥14 indicating substantial subjective cognitive impairment (Jensen et al., 2015).Quality of life was measured with the World Health Quality of Life questionnaire (WHO--QOL) (Nørholm and Bech, 2001).Hamilton Depression Rating Scale (HDRS) (Hamilton, 1960) and Young Mania Rating Scale (YMRS) (Young et al., 1978) were used to assess mood symptoms.Finally, Functional Assessment Short Test (FAST), with scores ≥12 indicating substantial subjective functional impairment, was used to assess daily functioning during the last 14 days (Rosa et al., 2007).

Data analysis
Aim (I) -the pattern, severity and prevalence of cognitive impairments in BD patients in optimal specialised treatment -was investigated with paired-samples t-test comparing the patients' actual SCIP scores with their individual expected scores based on age, gender and years of education (Ott et al., 2021).These expected scores were calculated using regression-based formulas for demographically corrected normative scores that are applicable at an individual level as described extensively in the Danish validation study (Ott et al., 2021).We used the difference between expected SCIP scores and actual SCIP scores to assess patients' cognitive status as either (1) cognitively normal (SCIP total score within 1 SD from expected total score), (2) selective impairment (performance ≥1 SD below expected score on ≥2 SCIP subtests), or (3) global impairment (SCIP total score ≥1 SD below expected total score).The proportion of patients with significant selective or global cognitive impairments despite optimal pharmacological and psychological treatment in a specialised outpatient clinic was established.We repeated all analyses in a subsample of patients in full or partial remission to examine the representativeness of the findings in this population.Aim (II) -the implications of cognitive impairments -was investigated in two ways: (a) we examined differences between globally impaired, selectively impaired, and cognitively normal patients in functioning, subjective cognitive complaints, quality of life, and clinical variables with one-way ANOVAs or Kruskal-Wallis tests on continuous variables (for normally-and non-normally distributed data, respectively) and Chi-Square analyses on categorical variables; (b) we explored associations between SCIP scores and functioning, subjective cognitive complaints, quality of life, and clinical variables with Pearson (correlation coefficients noted as r p ) and Spearman correlations (correlation coefficients noted as r s ) for normally and non-normally distributed data, respectively, across the entire sample.The association analyses were repeated in the subgroup of patients in full or partial remission.
Statistical analyses were completed using IBM SPSS statistics v28 for windows (IBM Corporation, Armonk, New York).Prior to analysis, statistical significance for all analyses was set to an alpha-level of p < .05(two-tailed).Effect sizes for significant differences were estimated with Cohen's d.We decided a priori to not adjust for multiple comparisons in the primary analyses due to the exploratory nature of the study.However, for significant findings, we applied posthoc Bonferroni adjustments to examine their robustness.

Participant characteristics
Fig. 1 depicts the flow-chart for the inclusion process.A total of 308 patients were identified from patient lists at the clinic and contacted by phone with the offer to get a cognitive screening, which was followed up by a written letter sent to their secure electronic mailbox (E-boks).A total of 157 (51 %) were enrolled in cognitive screening assessments.Of these, 21 were excluded from the present report for the following reasons: n = 4 completed the screening in English language; n = 15 had previously been screened with SCIP (to rule out the influence of repeated testing, despite the availability of alternate forms of SCIP); n = 2 patients had the diagnosis reevaluated to be unipolar disorder.Thus, 136 patients, aged from 20 to 68 years (median=29, interquartile range (IQR)= 24, 36), screened on a median of 19 months (IQR=17, 21) after initiation of specialised care, were included in the final analyses (see flowchart below for process details including reasons for non-attendance).Table 1 displays demographic and clinical characteristics of the patient group.Most patients (N = 99, 73 %) were in full or partial remission (HDRS and YMRS scores ≤14) after 15-20 months of specialised care in the clinic.Notably, 55 % presented with self-reported somatic comorbidities.Patients who completed a screening were slightly younger (median age=29, IQR=24, 36) than patients who did not complete a screening (median age=32, IQR=27, 38, p=.044) but did not differ in sex (chi-square=0.016,p=.905)

What is the prevalence, severity and pattern of cognitive impairments among bd patients in optimal treatment?
Fig. 2 illustrates the proportion of the 136 included patients who were cognitively normal or presented with clinically relevant global and selective cognitive impairment, respectively.A total of 70 (51 %) patients were cognitively normal, 51 (38 %) patients presented with globally impaired cognition, and 15 (11 %) patients had selective cognitive impairments.This distribution was highly similar for patients in full or partial remission (Figure S1).Specifically, 55 (56 %) patients were cognitively normal, 31 (31 %) presented with globally impaired cognition, and 13 (13 %) patients had selective cognitive impairments.
Fig. 3 illustrates differences between the patients' obtained SCIP scores and their demographically adjusted expected normative scores on the SCIP total and five subtests based on the regression-based formula (Ott et al., 2021).On average, the patients scored lower than expected according to normative data on the SCIP total, with a moderate effect

What are the associations between patients' cognitive status and subjective cognitive complaints, daily functioning, quality of life and clinical variables?
Table 2 presents comparisons between cognitive groups on subjective cognition, daily functioning, quality of life, and mood symptoms.Regardless of cognitive status, 87 % patients reported experiencing substantial subjective cognitive difficulties in daily life (COBRA score ≥14).Patients in the globally and selectively impaired groups presented with greater subjective cognitive difficulties compared to the cognitively normal group (Global: p < .001,d = 0.67; selective: p = .024,d = 0.67).The globally impaired group presented with poorer psychosocial functioning compared to the cognitively normal group (p=.002, d = 0.58), while the selectively impaired group showed intermediate functioning that did not differ from either of the other groups (p > .31).Finally, the globally impaired group reported lower quality of life compared to the cognitively normal group (p < .001,d = 0.69) and selectively impaired group (p=.012, d = 0.78), respectively.These differences between the globally impaired and cognitively normal group in subjective cognitive difficulties, psychosocial function, quality of life would have survived Bonferroni correction for multiple comparisons.These findings were largely replicated in fully or partially remitted patients (supplementary materials).
Regarding demographic and clinical variables, the globally impaired group presented with more depressive symptoms than the cognitively normal group (p = .001,d = 0.57), while there were no group differences in (hypo)manic symptoms (p ≥ .156).Similarly, we found no differences between the cognitive subgroups in gender, employment status, bipolar type, drug and alcohol use, types of medication, psychiatric or somatic comorbidities.In patients in full or partial remission, there were no group differences in mood symptoms or any demographic or clinical variables (supplementary materials).

Discussion
This study investigated cognitive functions in newly diagnosed bipolar disorder (BD) patients having undergone treatment in a specialised outpatint clinic over 15-20 months (N = 136).Most patients (73 %) achieved full or partial remission, with 51 % being cognitively normal, 38 % showing global impairments, and 11 % displaying selective impairments.In the remitted group, 56 % were cognitively normal, while 31 % and 13 % exhibited global or selective impairments, respectively.Globally and selectively impaired patients reported more daily cognitive difficulties.The globally impaired group also showed poorer psychosocial function, more subsyndromal depressive symptoms, and lower quality of life than the cognitively normal group.Across all patients, lower cognitive performance correlated with more subjective cognitive complaints, lower psychosocial functioning, lower quality of life, and more depressive symptoms.The study supports the clinical relevance of cognitive screenings in BD patients.
The proportions of BD patients that experience substantial cognitive impairment during full or partial remission have been reported in a range of studies to be about 50-70 % (Burdick et al., 2014;Jensen et al., 2016;Kjaerstad et al., 2021).We show here that these proportions are consistent with the prevalence of cognitive in newly diagnosed BD outpatients having undergone 15-20 months of specialised care in a multidisciplinary specialised outpatient clinic.The observed pattern of cognitive impairments in our cohortincluding in those in partial or full remission -is consistent with prior studies that demonstrated generally moderate deficits in attention, verbal memory, and working memory (Bourne et al., 2013).Although executive dysfunction is frequently cited as a prevalent cognitive impairment in BD patients, we observed no statistically significant impairment in verbal fluency, the sole executive function measure in the SCIP.This discrepancy might be attributed to the constrained scope of SCIP for screening purposes only, excluding supplementary executive function tests for the sake of brevity (Miskowiak et al., 2017).
While many studies found a discrepancy between subjectively and objectively measured cognitive difficulties in BD (Ott et al., 2021), this study stands in contrast by revealing significant albeit weak correlations between these two types of cognition measures.A reason for this unexpected concordance may be that our recently diagnosed patients had received specialised care including group-based psychoeducation, which may have contributed to heightened awareness of their symptoms, possibly indicating improved illness insight.Indeed, all patients enrolled in the Copenhagen Affective Disorders Clinic receive group-based psychoeducation on various aspects of BD, including cognitive impairment and its relation to the disorder and daily life function.
The present findings align with these observations, showing that greater deficits in global cognition, verbal learning, working memory, and verbal fluency are associated with increased functional impairment in daily life.This study further establishes a link between cognitive impairment and diminished quality of life, an association that was observed in the full sample as well as in the subsample of fully or partially remitted patients.In concert, these results substantiate the presumed associations between cognitive impairment in BD patients and adverse functional outcomes, reduced quality of life, and hurdles in achieving and sustaining remission for those affected (Burdick et al., 2014;Tse et al., 2014).This highlights the practical value of cognitive screenings for identifying patients in need of additional intervention, such as medication adjustments, strategies to aid treatment adherence, or lifestyle modifications, to optimize their cognition, psychosocial function, and quality of life.A clinical benefit of SCIP lies in its psychoeducational nature.A patient's SCIP results can serve to initiate constructive conversations about cognition in BD, emphasizing understanding, realistic expectations, and strategies to compensate for cognitive challenges.Future qualitative studies including patient outcome interviews post-screening are warranted to further elucidate the clinical impact of counseling on cognitive difficulties.Given the prevalence of cognitive impairments, such investigations can inform clinical and therapeutic approaches, examining the effectiveness of tailored interventions across diagnoses.
The determination of clinically relevant cognitive impairments is contentious, with various proposed cut-offs in the literature (Miskowiak et al., 2018;Ott et al., 2021), contingent on their context of application.In research settings, a cut-off of ≥0.5 standard deviations (SD) below the expected SCIP total score, adjusted for age, sex, and education, has been suggested to enhance sensitivity in identifying patients eligible for pro-cognitive interventions (Ott et al., 2021).Conversely, for clinical purposes, a more conservative cut-off of ≥1.0 SD has been recommended to minimize false positives.Following ISBD Task Force recommendations (Miskowiak et al., 2018), selective cognitive impairments were identified by performance ≥1 SD below expected scores on ≥2 individual SCIP tasks.It is noteworthy that the selective impairment The feasibility of cognitive screenings in outpatient care for individuals with BD underscores their clinical importance.The SCIP screening, which takes around 15 min to administer, was completed for approximately 50 % of patients in the clinic, indicating broad acceptability among this population.The feedback included in the session facilitated productive discussions, validating experiences for those with objective cognitive impairment and alleviating concerns for those with normal cognition (See case example in the supplementary materials).Patients reported high satisfaction, particularly appreciating the psychoeducation and cognitive strategy discussions, which empowered them by fostering a sense of agency.Clinicians also found the screenings valuable, enriching their understanding of patients' cognitive status and contributing lifestyle factors, while enhancing the overall treatment approach.The decision to conduct cognitive screening between 15 and 20 months after enrollment in the clinic allowed for patients to first receive optimal pharmacological and psychological treatment, resulting in relative symptom stability.Consequently, the cognitive assessments provided insights into patients' cognitive functioning under optimised treatment conditions.While early screening may offer benefits, our later approach prioritised a cost effectiveness perspective (to not screen patients unnecessarily), also capturing a more accurate representation of cognitive status following sustained treatment efforts.
A notable limitation was the 50 % acceptance rate among invited patients for cognitive screening, reflecting challenges in engaging patients with BD in research activities.Most of the unscreened patients were not well enough to participate due to mood instability and impaired functioning, suggesting a potential bias toward recruiting more engaged and motivated individuals actively involved in their treatment.Consequently, those who declined or were unreachable were likely to be older and more unwell.On the other hand, it cannot be excluded that some non-participants did not want a cognitive assessment as they may have felt it was not of relevance to them -e.g., they felt they were functioning cognitively well.These uncertainties may limit the generalisability of our findings to the broader spectrum of BD outpatients.Despite these limitations, it is noteworthy that approximately 50 % of those screened displayed clinically relevant cognitive impairments, a figure that is consistent with previous findings regarding the prevalence of cognitive impairment among partially or fully remitted patients with BD.Efforts to improve participation rates and address potential biases are warranted to enhance the representativeness of future studies in this population.Additionally, the brevity of the SCIP and lack of more robust executive function measures is a limitation, emphasizing its role as a screening tool for clinic-based monitoring rather than a substitute for comprehensive neuropsychological assessments.If suspicions of neurological or organic causes for cognitive impairment arise, referral for a thorough neuropsychological assessment, rather than the SCIP, would thus be recommended.
In conclusion, adhering to the ISBD Task Force on Cognition recommendations (Miskowiak et al., 2018), for systematic cognitive screenings in BD patients, this study examined impairments, feasibility, and clinical relevance at the Copenhagen Affective Disorder Clinic.Patients exhibited lower-than-expected scores, with impaired individuals showing more depressive symptoms, reduced psychosocial function, and lower quality of life.The study affirms the practicality and importance of incorporating cognitive screenings into clinical settings, particularly in specialised outpatient clinics.However, the relatively low acceptance rate among invited patients, particularly those with more severe conditions, poses challenges.Future efforts should focus on integrating cognitive screenings as a standard part of clinical practice to ensure comprehensive patient care and improve participation rates.This approach will be valuable in identifying individuals requiring additional care for optimal treatment and facilitating functional recovery.

Role of funding source
No funding was provided for this study.

Contributors
KWM and LVK designed the study and wrote the protocol with input from LH and BV.JLB, JM, TR, ZKO conducted the cognitive screenings and feedback with supervision from KWM. JM and JLB conducted the statistical analyses supervised by KWM.KWM wrote the first draft of the manuscript with input from TR, JM and JLB.All authors contributed to and have approved the final manuscript.

Declaration of competing interest
KWM reports having received consultancy fees from Lundbeck, Angelini, Gideon Richter and Janssen-Cilag in the past three years.LVK reports having received consultancy fees from Lundbeck and Teva in the past three years.JM, TR, JLB, ZKO, LH and report no conflicts of interest.

Fig. 2 .
Fig. 2. Proportion of patients with clinically relevant global or selective cognitive impairments following ISBD guidelines, using a conservative cut-off for determining the clinical relevance of impairments (1≥SD under the expected score of the SCIP total for global impairment and 2≥Subtests under 1≥SD of the expected score for selective impairment).

Fig. 3 .
Fig. 3. Pattern of cognitive impairments in patients (n = 136) at ≥15 months of treatment at AAL with normative scores adjusted for age, sex and education estimated with regression models.Most pronounced impairments were seen in working memory (WMT), verbal memory (VLT-D) and SCIP total scores.Graphs represent the mean and error bars the standard error of the mean.* p < 0-05; ** p < .01;*** p < .001.

Table 1
Demographic and clinical characteristics in included patients.

Table 2
Associations between cognitive status, subjective cognition, daily functioning, quality of life, and mood symptoms (N = 136).Standard deviation.IQR = Interquartile range.COBRA = Cognitive Complaints in Bipolar Disorder Rating Assessment.FAST = Functional Assessment Short Test.WHO-QOL = World Health Quality of Life questionnaire.HDRS-17: Hamilton Depression Rating Scale 17-items.YMRS = Young Mania Rating Scale.Test value (F/K) = F or K value for normally and non-normally distributed data, respectively.Test p-value = P-value for one-way ANOVA or Kruskal Wallis test for normally and non-normally distributed data, respectively.groupexhibited comparable subjective cognitive challenges in everyday functioning to the global impairment group.However, distinct differences emerged regarding psychosocial disability and quality of life, with only the globally impaired group demonstrating significant impairment.The selective impairment group showed intermediate scores, that differed from neither the more impaired global group nor the healthy controls, suggesting that these less impaired patients are able maintain daily function and quality of life.