Misdiagnosing bipolar disorder--do clinicians show heuristic biases?

A diagnostic model based on bioinformatics and machine learning to differentiate bipolar disorder from schizophrenia and major depressive disorder

Bipolar disorder (BD) showed the highest suicide rate of all psychiatric disorders, and its underlying causative genes and effective treatments remain unclear. During diagnosis, BD is often confused with schizophrenia (SC) and major depressive disorder (MDD), due to which patients may receive inadequate or inappropriate treatment, which is detrimental to their prognosis. This study aims to establish a diagnostic model to distinguish BD from SC and MDD in multiple public datasets through bioinformatics and machine learning and to provide new ideas for diagnosing BD in the future. Three brain tissue datasets containing BD, SC, and MDD were chosen from the Gene Expression Omnibus database (GEO), and two peripheral blood datasets were selected for validation. Linear Models for Microarray Data (Limma) analysis was carried out to identify differentially expressed genes (DEGs). Functional enrichment analysis and machine learning were utilized to identify. Least absolute shrinkage and selection operator (LASSO) regression was employed for identifying candidate immune-associated central genes, constructing protein-protein interaction networks (PPI), building artificial neural networks (ANN) for validation, and plotting receiver operating characteristic curve (ROC curve) for differentiating BD from SC and MDD and creating immune cell infiltration to study immune cell dysregulation in the three diseases. RBM10 was obtained as a candidate gene to distinguish BD from SC. Five candidate genes (LYPD1, HMBS, HEBP2, SETD3, and ECM2) were obtained to distinguish BD from MDD. The validation was performed by ANN, and ROC curves were plotted for diagnostic value assessment. The outcomes exhibited the prediction model to have a promising diagnostic value. In the immune infiltration analysis, Naive B, Resting NK, and Activated Mast Cells were found to be substantially different between BD and SC. Naive B and Memory B cells were prominently variant between BD and MDD. In this study, RBM10 was found as a candidate gene to distinguish BD from SC; LYPD1, HMBS, HEBP2, SETD3, and ECM2 serve as five candidate genes to distinguish BD from MDD. The results obtained from the ANN network showed that these candidate genes could perfectly distinguish BD from SC and MDD (76.923% and 81.538%, respectively).

Generalizing the Prediction of Bipolar Disorder Onset Across High-Risk Populations

OBJECTIVE

Risk calculators (RC) to predict clinical outcomes are gaining interest. An RC to estimate risk of bipolar spectrum disorders (BPSD) could help reduce the duration of undiagnosed BPSD and improve outcomes. Our objective was to adapt an RC previously validated in the Pittsburgh Bipolar Offspring Study (BIOS) sample to achieve adequate predictive ability in both familial high-risk and clinical high-risk youths.

METHOD

Participants (aged 6-12 years at baseline) from the Longitudinal Assessment of Manic Symptoms (LAMS) study (N = 473) were evaluated semi-annually. Evaluations included a Kiddie Schedule for Affective Disorders (K-SADS) interview. After testing an RC that closely approximated the original, we made modifications to improve model prediction. Models were trained in the BIOS data, which included biennial K-SADS assessments, and tested in LAMS. The final model was then trained in LAMS participants, including family history of BPSD as a predictor, and tested in the familial high-risk sample.

RESULTS

Over follow-up, 65 youths newly met criteria for BPSD. The original RC identified youths who developed BPSD only moderately well (area under the curve [AUC] = 0.67). Eliminating predictors other than the K-SADS screening items for mania and depression improved accuracy (AUC = 0.73) and generalizability. The model trained in LAMS, including family history as a predictor, performed well in the BIOS sample (AUC = 0.74).

CONCLUSION

The clinical circumstances under which the assessment of symptoms occurs affects RC accuracy; focusing on symptoms related to the onset of BPSD improved generalizability. Validation of the RC under clinically realistic circumstances will be an important next step.

Brain grey-matter volume alteration in adult patients with bipolar disorder under different conditions: a voxel-based meta-analysis

BACKGROUND

The literature on grey-matter volume alterations in bipolar disorder is heterogeneous in its findings.

METHODS

Using effect-size differential mapping, we conducted a meta-analysis of grey-matter volume alterations in patients with bipolar disorder compared with healthy controls.

RESULTS

We analyzed data from 50 studies that included 1843 patients with bipolar disorder and 2289 controls. Findings revealed lower grey-matter volumes in the bilateral superior frontal gyri, left anterior cingulate cortex and right insula in patients with bipolar disorder and in patients with bipolar disorder type I. Patients with bipolar disorder in the euthymic and depressive phases had spatially distinct regions of altered grey-matter volume. Meta-regression revealed that the proportion of female patients with bipolar disorder or bipolar disorder type I was negatively correlated with regional grey-matter alteration in the right insula; the proportion of patients with bipolar disorder or bipolar disorder type I taking lithium was positively correlated with regional grey-matter alterations in the left anterior cingulate/paracingulate gyri; and the proportion of patients taking antipsychotic medications was negatively correlated with alterations in the anterior cingulate/paracingulate gyri.

LIMITATIONS

This study was cross-sectional; analysis techniques, patient characteristics and clinical variables in the included studies were heterogeneous.

CONCLUSION

Structural grey-matter abnormalities in patients with bipolar disorder and bipolar disorder type I were mainly in the prefrontal cortex and insula. Patients' mood state might affect grey-matter alterations. Abnormalities in regional grey-matter volume could be correlated with patients' specific demographic and clinical features.

Factors associated with false positives in MDQ screening for bipolar disorder: Insight into the construct validity of the scale

BACKGROUND

Identifying bipolar patients in the first phases of the illness is essential to establish adequate treatment. The goal of this study was to examine the discriminant ability of the Mood Disorders Questionnaire (MDQ) in recognizing bipolar patients referred to a tertiary care structure.

METHODS

Between 2006 and 2012, we assessed 843 individuals referred to the Mood Disorders Program by family physicians in the community. The Structured Clinical Interview for DSM-IV-TR (SCID) was used to assess diagnoses. A nurse collected the information about lifetime symptoms of (hypo)mania in 759 individuals using the MDQ. Univariate chi-square test and logistic regression were used for the statistical analysis.

RESULTS

Overall, 86% of the sample had a current anxiety or depressive disorder. When compared to the diagnoses formulated through the SCID, the sensitivity of the MDQ was 75.0%, the specificity was 74%, the positive predictive value was 55%, and the negative predictive value was 88%. Among non-bipolar patients, current post-traumatic stress disorder, borderline personality disorder, current or early remission substance use disorder, and the history of childhood abuse were independently associated with false positive screening using the MDQ.

LIMITATIONS

Individuals with current substance use disorders were under-represented, whether or not the patients were aware of their diagnosis of bipolar disorder was not recorded, and the history of childhood abuse was collected based on an open interview.

CONCLUSIONS

The self-rated measure of the symptoms listed by the MDQ seems to measure a dimension shared by both bipolar disorder and other conditions characterized by affective instability and impulsivity.

Strukturierte klinische Interviews zur Erfassung psychischer Störungen über die Lebensspanne

Zusammenfassung. Psychische Störungen können über die ganze Lebensspanne auftreten. Strukturierte klinische Interviews sind zentrale Hilfsmittel für ihre rasche, zuverlässige und umfassende Diagnostik. Im deutschsprachigen Raum stehen mit den Verfahren der DIPS-Familie Interviews zur Diagnostik psychischer Störungen über die gesamte Lebensspanne zur Verfügung, die seit den 90er Jahren regelmäßig aktualisiert wurden. Ihre Reliabilität, Validität und Akzeptanz wurde wiederholt in großen Stichproben aus ambulanten, stationären und Forschungssettings überprüft. Die Einführung des DSM-5 erforderte eine umfassende Überarbeitung der DIPS-Interviews, deren wesentliche Merkmale dargestellt werden. Um die breitere Verwendung von strukturierten klinischen Interviews zu fördern, werden die Verfahren der DIPS-Familie neu als „Open Access-Dokumente“ zur Verfügung gestellt. Abschließend werden weitere Entwicklungen zu Training, Dissemination und Computerisierung im Ausblick angesprochen.

Hypomania Checklist-32 - cross-validation of shorter versions screening for bipolar disorders in an epidemiological study

OBJECTIVE

Self-reports such as Hypomania Checklist (HCL-32) can be used to enhance recognition of bipolar disorders, but they are often too long and only validated in clinical samples. The objectives of this study are therefore to test whether (i) the HCL-32 can be used for screening in the community and (ii) whether two previously suggested shorter versions would do as well.

METHOD

Data stemmed from the CoLaus|PsyColaus, a prospective cohort study which included randomly selected residents aged 35-66 years from an urban area. Participants underwent semistructured interviews to assess DSM-IV disorders and 1712 of them completed the HCL-32.

RESULTS

Forty individuals (2.3%) were diagnosed as having BD. Compared to others, participants with BD scored significantly higher on the HCL-32. The HCL-32 had a sensitivity of 0.78 and specificity of 0.68. Very similar figures were found for two previously proposed shorter versions with 16 and 20 items. The results of confirmatory factor analysis and item response theory (IRT) models supported the postulated two-factor structure for the three HCL versions.

CONCLUSION

Despite the low base rate of BD in this sample, the screening properties of the HCL-32 remained almost as good. Importantly, two previously proposed shorter versions performed as well, suggesting that those could be used without losing essential information.

Information order effects in clinical psychological diagnoses

Despite the wide application and long history of diagnostic systems, several sources of diagnostic errors remain in the criterion-based diagnosing of mental disorders. The aim of this study was to investigate whether the presentational order of diagnosis-relevant information and pretreatment reports predict diagnostic errors. One hundred twenty psychotherapists participated in the present online study. The study employed a 2 (symptom presentation: core symptoms at vignette's beginning vs. core symptoms at the end of the case vignette) × 2 (pretreatment report: receiving a pretreatment report with an incongruent diagnosis to the case vignette vs. receiving no pretreatment report) between-subjects experimental design, with random assignment. Participants were asked to make diagnoses after reading three case vignettes describing patients with different disorder constellations. Additionally, participants rated their confidence in the diagnoses and their estimation of the severity of each diagnosed condition. Results indicated that order of symptom descriptions predicted the correctness of diagnostic decisions, with a recency effect causing more fully correct diagnostic decisions in cases where diagnostic information was presented last. Receiving incongruent pretreatment reports was predictive for diagnostic errors. In conclusion, the results of this study indicate that diagnoses of mental disorders can depend on the way symptoms are presented or reported.

KEY PRACTITIONER MESSAGE

Therapists' diagnostic decisions are not influenced by pretreatment reports. Diagnostic decisions are affected by information order effects. Diagnostic accuracy of psychotherapists is debatable. High rate of misdiagnoses in case vignette with comorbid disorders.

An Investigation of Diagnostic Accuracy and Confidence Associated with Diagnostic Checklists as Well as Gender Biases in Relation to Mental Disorders

This study examines the utility of checklists in attaining more accurate diagnoses in the context of diagnostic decision-making for mental disorders. The study also aimed to replicate results from a meta-analysis indicating that there is no association between patients’ gender and misdiagnoses. To this end, 475 psychotherapists were asked to judge three case vignettes describing patients with Major Depressive Disorder (MDD), Generalized Anxiety Disorder, and Borderline Personality Disorder. Therapists were randomly assigned to experimental conditions in a 2 (diagnostic method: with using diagnostic checklists vs. without using diagnostic checklists) × 2 (gender: male vs. female case vignettes) between-subjects design. Multinomial logistic and linear regression analyses were used to examine the association between the usage of diagnostic checklists as well as patients’ gender and diagnostic decisions. The results showed that when checklists were used, fewer incorrect co-morbid diagnoses were made, but clinicians were less likely to diagnose MDD even when the criteria were met. Additionally, checklists improved therapists’ confidence with diagnostic decisions, but were not associated with estimations of patients’ characteristics. As expected, there were no significant associations between gender and diagnostic decisions.

Teaching Cognitive Bias in a Hurry: Single-Session Workshop Approach for Psychiatry Residents and Students

INTRODUCTION

Diagnostic ambiguity is commonplace in psychiatric practice. With limited etiological understanding behind the illnesses that we treat, it can be a daunting task to diagnostically approach a complex patient. To avoid diagnostic pitfalls, it is paramount to employ carefully constructed diagnostic cognitive processes. The aim of this resource is to decrease diagnostic thought errors through education via acquired knowledge and skills.

METHODS

This workshop is intended as a primer for psychiatric medical students, interns, residents, and practicing psychiatrists who have not yet been exposed to this topic. The resource consists of a brief didactic session (knowledge portion of the topic) followed by small-group sessions (skills portion of the topic), led by discussion of clinical vignettes incorporating complex medical-psychiatric patients. The cases highlight the various cognitive biases and anchor points that can contribute to diagnostic error. In order to make it easier for educators to implement, the workshop can be delivered in a single 1-hour session. With the provided material, this educational session can be easily delivered by a single facilitator.

RESULTS

Formal and informal feedback from learners demonstrates that this resource is effective at teaching cognitive bias to learners.

DISCUSSION

While psychiatric learners are used to metacognitive approaches in clinical settings, they often lag behind other medical specialties in discussing these topics related to their own diagnostic reasoning and acumen. This is especially important in psychiatry given the evolving nature of our specialty.

Misclassification of Self-Directed Violence

BACKGROUND

Inconsistent nomenclature and classification of suicidal behaviour have plagued the field of suicidology for a long time. Recently, the United States Centers for Disease Control (CDC) advocated for the usage of a specific classification system. Aim of the current study was to determine the extent of misdiagnosed acts of self-directed violence-controlling for the level of expertise in psychology/psychotherapy. Additionally, the effect of gender and diagnosis on misclassifications was assessed.

METHOD

A total of 426 participants (laypersons, psychology students, psychotherapists-in-training, licensed psychotherapists) were presented with an array of case vignettes describing different acts of self-directed violence (e.g., non-suicidal self-directed violence, suicide attempt, suicide ideation) and were asked to make a classification. Gender and given diagnosis were varied systematically in two vignettes.

RESULTS

Overall 51.6% of the cases were misclassified (according to the Self-Directed Violence Classification System). The level of expertise was almost unrelated to classification correctness. Yet, psychotherapists were more confident about their judgments. Female gender of the character described in the vignette and an ascribed diagnosis of Borderline Personality Disorder were associated with higher misclassification rates.

LIMITATIONS

The validity of case vignettes is discussible.

CONCLUSIONS

The results highlight the importance of more methodological and diagnostic training of psychologists regarding suicidal issues. Copyright © 2016 John Wiley & Sons, Ltd.

KEY PRACTITIONER MESSAGE

Misclassification of non-suicidal and suicidal events is common. Expertise is only weakly associated with classification correctness. Misclassification of suicide attempts occurs more often in women. Misclassification of suicide attempts occurs more often in Borderline Personality Disorder. The use of standardized diagnostic tools in relation to self-directed violence is highly recommended.

Grey matter volume abnormalities in patients with bipolar I depressive disorder and unipolar depressive disorder: a voxel-based morphometry study

Bipolar disorder and unipolar depressive disorder (UD) may be different in brain structure. In the present study, we performed voxel-based morphometry (VBM) to quantify the grey matter volumes in 23 patients with bipolar I depressive disorder (BP1) and 23 patients with UD, and 23 age-, gender-, and education-matched healthy controls (HCs) using magnetic resonance imaging. We found that compared with the HC and UD groups, the BP1 group showed reduced grey matter volumes in the right inferior frontal gyrus and middle cingulate gyrus, while the UD group showed reduced volume in the right inferior frontal gyrus compared to HCs. In addition, correlation analyses revealed that the grey matter volumes of these regions were negatively correlated with the Hamilton depression rating scores. Taken together, the results of our study suggest that decreased grey matter volume of the right inferior frontal gyrus is a common abnormality in BP1 and UD, and decreased grey matter volume in the right middle cingulate gyrus may be specific to BP1.

The Hypomania Checklist (HCL) - systematic review of its properties to screen for bipolar disorders

OBJECTIVE

Bipolar disorders (BDs) are often not recognised with potentially drastic consequences for the individuals and their families. In clinical practice self-reports can be used to screen to enhance recognition. We therefore present a systematic review of the screening properties for the Hypomania Checklist (HCL-32).

METHODS

A systematic literature search was conducted to identify all relevant studies looking at the screening properties of the HCL-32 in adults.

RESULTS

Out of 196 papers 21 papers reported data on 22 independent samples. We narratively reviewed these studies. Weighted estimated Sensitivity was 80% regardless of whether a BD diagnosis was compared to unipolar depression or any other non-bipolar diagnosis. Specificity indicated that the HCL-32 was better when comparing BD to unipolar depression (65.3%) than to any other diagnostic category (57.3%). Fewer studies provided estimates for predictive powers, leading to less reliable overall estimates for these indicators.

CONCLUSIONS

Despite some limitations, using the HCL-32 as a first screening in patients seeking help for depression can be recommended, but should never be used on its own for diagnosing. Future research should examine whether screening properties can be improved by developing an algorithm incorporating the negative consequences reported for different areas in the HCL-32.

Predictors of primary care physicians' self-reported intention to conduct suicide risk assessments

Primary care physicians play a significant role in depression care, suicide assessment, and suicide prevention. However, little is known about what factors relate to and predict quality of depression care (assessment, diagnosis, and treatment), including suicide assessment. The authors explored the extent to which select patient and physician factors increase the probability of one element of quality of care: namely, intention to conduct suicide assessment. Data were collected from 404 randomly selected primary care physicians after their interaction with CD-ROM vignettes of actors portraying major depression with moderate levels of severity. The authors examined which patient factors and physician factors increase the likelihood of physicians' intention to conduct a suicide assessment. Data from the study revealed that physician-participants inquired about suicide 36% of the time. A random effects logistic model indicated that several factors were predictive of physicians' intention to conduct a suicide assessment: patient's comorbidity status (odds ratio (OR) = 0.61; 95% confidence interval (CI) = 0.37-1.00), physicians' age (OR = 0.67; 95% CI = 0.49-0.92), physicians' race (OR = 1.84; 95% CI = 1.08-3.13), and how depressed the physician perceived the virtual patient to be (OR = 0.58; 95% CI = 0.39-0.87). A substantial number of primary care physicians in this study indicated they would not assess for suicide, even though most physicians perceived the virtual patient to be depressed or very depressed. Further study is needed to establish factors that may be modified and targeted to increase the likelihood of physicians' providing one element of quality of care--suicide assessment--for depressed patients.

Diagnostic disagreements in bipolar disorder: the role of substance abuse comorbidities

Substance abuse can produce symptoms similar to other psychiatric disorders, thus confusing the diagnostic picture. This paper attempts to elucidate how misdiagnosis in bipolar disorder might be explained by the presence of substance abuse comorbidities. The overlap of symptoms, limited information about symptom onset, and inexperienced clinicians can result in the misinterpretation of symptoms of substance abuse disorders for bipolar disorder. The present study found that the presence of a substance abuse comorbidity, the polarity of last episode (depressed, manic, mixed, not otherwise specified), and the total number of comorbidities affected the reliability of a bipolar disorder diagnosis.