10. Clinical Vignettes

The following composite cases are drawn from published literature and clinical patterns described in the research reviewed for this module. No case represents a single patient. Each illustrates decision points where the evidence covered in this module directly affects clinical outcomes. Vignette 1: The Diagnosis That Depended on Who Made It A 28-year-old Black man is brought to the emergency department by his family after three days of decreased sleep, pressured speech, grandiose statements about a business venture, and disorganized behavior including spending $14,000 on equipment for the venture. He is intermittently irritable and has made several vague references to "people working against me." His family reports that he was treated for a major depressive episode two years ago and that his maternal aunt has bipolar I disorder. The overnight psychiatry resident diagnoses schizoaffective disorder, citing the paranoid ideation and disorganized behavior. She starts risperidone 2 mg and plans discharge to the community mental health center with an antipsychotic prescription. No mood stabilizer is discussed. The clinical decision points this module addresses: The diagnostic bias literature (Section 1) documents that Black patients presenting with affective psychoses are 2.4 to 4 times more likely to receive a schizophrenia-spectrum diagnosis than White patients with identical symptom profiles (Olbert et al., 2018). The diagnostic bias research (Section 1.1) documents that implicit cognitive shortcuts lead clinicians to equate psychosis in Black patients with schizophrenia while interpreting identical psychotic features in White patients as manifestations of a mood disorder (Gara et al., 2019). This patient's presentation, including grandiosity, decreased sleep, pressured speech, excessive spending, a clear mood episode trajectory, and a first-degree relative with bipolar I, is more consistent with a manic episode with mood-congruent psychotic features than with a primary psychotic disorder. The vague paranoid ideation does not resolve this differential. Psychotic features occur in approximately 50-70% of manic episodes (Module 1, Section 1). The presence of paranoia during a manic episode does not indicate schizophrenia or schizoaffective disorder. The critical clinical question is whether the psychosis occurs exclusively within the context of mood episodes (bipolar I with psychotic features) or independently of mood episodes (schizoaffective disorder). This distinction requires longitudinal observation that cannot be made in a single emergency department visit. Teaching points: The prescribing disparity data (Section 1.3) shows that this patient is less likely to receive a mood stabilizer and more likely to receive a high-dose antipsychotic without lithium or valproate, perpetuating a treatment trajectory that may worsen outcomes. Structured diagnostic instruments and systematic collateral history (asking about prior hypomanic or manic episodes, not just the current presentation) reduce the influence of implicit bias on diagnostic decision-making. A clinician trained in the content of Sections 1 and 3 would recognize this as a case requiring careful differential diagnosis rather than rapid diagnostic closure. For non-psychiatrist providers encountering this presentation: this patient warrants urgent psychiatric consultation for diagnostic clarification and mood stabilizer consideration alongside the antipsychotic, rather than discharge on antipsychotic monotherapy with a psychotic disorder diagnosis. Vignette 2: The Patient Who Wants to Stop Lithium After Seven Good Years Dr. Patel has been treating Sandra, a 52-year-old teacher with bipolar I disorder, for nine years. Sandra has been on lithium 900 mg daily (serum level 0.7 mEq/L) and has been euthymic for seven consecutive years. She is physically active, has a supportive marriage, and has not been hospitalized since her last manic episode at age 45. At today's appointment, Sandra says she has been reading about long-term lithium effects on kidney function and wants to discuss tapering off. She also mentions that a friend with bipolar disorder successfully discontinued medication using "a genetic test that showed she didn't need it anymore." The clinical decision points this module addresses: The maintenance discontinuation data (Section 9.3) provides the evidence Sandra and Dr. Patel need: in Sharma et al.'s (2014) study, 87% of patients who had been stable for five or more years relapsed within two years of discontinuation, with a median time to recurrence of 10 months. All relapses were manic. Goodwin's meta-analysis found 63% relapse within one year after lithium discontinuation versus 10% among those who continued, and a 20-fold increase in suicidal behavior in the first year post-discontinuation. Sandra's concern about renal effects is clinically legitimate, and renal monitoring should be ongoing (Module 5 covers lithium nephrotoxicity in detail). But the pharmacogenomic testing her friend mentioned (Section 4) does not currently offer guidance on whether a given patient can safely discontinue a mood stabilizer. Commercial PGx panels can provide information about drug metabolism (CYP enzymes) and hypersensitivity risk (HLA-B1502 for carbamazepine), but they do not predict bipolar illness course, episode recurrence risk, or lithium necessity. The APA's critical review found that only 27% of PGx clinical trials (3 of 11) demonstrated statistically significant efficacy on their primary outcome measures (APA 2023 critical review; Bousman et al., 2023, Molecular Psychiatry). A patient who discontinues lithium based on a PGx result showing "normal metabolism" has misunderstood what the test measures. Teaching points: If Sandra and Dr. Patel proceed with a taper despite the evidence, Section 9.4 outlines the approach: hyperbolic tapering over months (not weeks), monitoring for manic relapse (which typically precedes depressive relapse after lithium discontinuation), and clear criteria for re-initiation. The 2025 Joint Clinical Practice Guideline on psychotropic tapering recommends dose reductions that become progressively smaller as the serum level approaches zero. Dr. Patel should also address Sandra's renal concerns directly with current data rather than allowing them to drive a discontinuation decision by default. Vignette 3: The Bipolar II Patient Asking About Psilocybin and a Mood-Tracking App James, a 34-year-old software engineer with bipolar II disorder, comes to a follow-up appointment on lamotrigine 200 mg and bupropion 300 mg. His hypomania has been well-controlled, but he continues to experience low-grade depressive symptoms despite adequate trials of lamotrigine, quetiapine (discontinued for weight gain), and the current bupropion augmentation. He brings two requests: first, he has been reading about psilocybin therapy for depression and wants to know whether it could help his bipolar depression. Second, he has been using a mood-tracking app that claims to predict depressive episodes 7-10 days in advance using phone sensor data and wants his clinician to incorporate the app's predictions into treatment decisions. The clinical decision points this module addresses: For the psilocybin question, Section 5 provides the current evidence. Bipolar patients have been systematically excluded from the major psilocybin clinical trials (COMP360, Johns Hopkins, NYU). The only published data specific to bipolar psilocybin use comes from a single open-label study of 15 bipolar II patients (Aaronson et al., 2024). Antidepressant effects were rapid, and 9 of 15 participants (60%) maintained remission off pharmacotherapy through 12 weeks of follow-up, but several participants experienced treatment-emergent hypomania (Aaronson et al., 2024). The critical safety concern is the lithium-psychedelic contraindication: a 47% seizure rate has been documented in naturalistic reports of patients combining lithium with psychedelics (Nayak et al., 2021, Pharmacopsychiatry). James is not on lithium, but if psilocybin were ever considered, his lamotrigine and bupropion would require evaluation. At present, the evidence does not support psilocybin use for bipolar depression outside of specialized research settings with strict protocols, and James should be counseled accordingly. For the mood-tracking app, Section 7 provides the context he needs. Machine learning models applied to smartphone sensor data (voice patterns, sleep timing, activity levels) have shown impressive accuracy metrics in research settings, with some models reaching 85-92% sensitivity for episode prediction. However, the OXTEXT trial (Section 7.4) provides a critical real-world counterpoint: when smartphone-based monitoring was deployed to actual patients, the false positive rate exceeded 90%. In clinical terms, more than 9 out of 10 alerts were false alarms. For a patient like James who experiences residual depressive symptoms, an app generating frequent false positive depression warnings could increase anxiety, prompt unnecessary medication changes, or erode trust in the monitoring system entirely. The appropriate clinical response is not to dismiss the app but to contextualize its predictions: app data can supplement clinical judgment by identifying trends (sleep shortening over several weeks, for instance) but should not drive treatment decisions in isolation. Teaching points: James represents a common clinical encounter: a well-informed patient whose treatment has been partially effective and who is seeking emerging options. The clinician's role is not to dismiss his interest in psilocybin or digital tools but to provide an accurate, evidence-based assessment that distinguishes genuine clinical promise from media hype and commercial marketing. For James's refractory bipolar depression specifically, the evidence better supports considering lurasidone (Section 9.5: NNT of 3 for mixed features, though James presents with pure depression) or rTMS (Section 6: NNT of 8, with manic switch rates not significantly elevated above sham in pooled trials) as next steps within the established evidence base. Key Takeaways for Clinical Practice 1. Diagnostic bias in bipolar disorder is quantitative, persistent, and directly affects your patients. Black patients presenting with affective psychoses remain 2.4 to 4 times more likely to receive a schizophrenia-spectrum diagnosis than White patients with identical symptoms, a disparity confirmed by 2024-2025 electronic health record data. Structured assessment instruments and systematic collateral history reduce the influence of implicit bias on diagnostic decision-making. Every practice should audit its own diagnostic and prescribing patterns for racial and ethnic disparities. School psychologists conducting evaluations should be aware that these biases affect the referral pipeline: a student referred with a schizophrenia-spectrum label may warrant reconsideration for a mood disorder, particularly if there is episodic course, family history of bipolar disorder, or prominent mood symptoms alongside psychotic features. (Section 1) 2. The bipolar overdiagnosis and underdiagnosis problems coexist; they do not cancel each other out. Community studies suggest 40% of bipolar diagnoses may not meet strict DSM criteria, while the average diagnostic delay from first mood episode to correct bipolar diagnosis remains 5-10 years. These patterns reflect different clinical settings (overdiagnosis in outpatient practice, underdiagnosis in emergency and primary care settings) and different patient populations. Addressing one without the other provides an incomplete picture. (Section 3) 3. Gender and reproductive factors create specific pharmacotherapy challenges that require proactive management. The lamotrigine-oral contraceptive interaction (estrogen induces UGT enzymes, potentially halving lamotrigine levels) and valproate's reproductive toxicity (absolute contraindication in women of reproductive potential per multiple international guidelines) are not optional knowledge for any prescriber treating bipolar disorder in women. Menstrual cycle mood destabilization affects approximately 45-68% of women with bipolar disorder, and testosterone therapy in transgender men can precipitate manic episodes, particularly with supraphysiological peak-and-trough dosing from intramuscular depot injections. Exogenous estrogen in transfeminine patients induces UGT enzymes that accelerate lamotrigine clearance, making non-oral estrogen routes strongly preferred. (Section 2) 4. Precision psychiatry tools for bipolar disorder are mostly investigational; temper your patients' expectations. No single biomarker (neuroimaging, blood-based, electrophysiological, or genetic) approaches clinical diagnostic utility for bipolar disorder. HLA-B1502 screening before carbamazepine in Southeast Asian-ancestry patients is the one pharmacogenomic application with clear clinical utility. Commercial PGx panels may provide some guidance on drug metabolism, but only 27% of PGx clinical trials (3 of 11 reviewed by the APA) demonstrated statistically significant results on their primary outcome measures. Patients requesting "a genetic test to guide my medication" deserve an accurate explanation of what these tests can and cannot do. (Section 4) 5. Counsel patients accurately about psychedelic-assisted therapy: the evidence for bipolar is almost nonexistent, and one combination is actively dangerous. Bipolar patients have been excluded from major psilocybin trials. The only published bipolar-specific data involves 15 patients. Lithium combined with psychedelics carries a 47% seizure rate in naturalistic reports. If a patient discloses intent to use psychedelics, the lithium contraindication is the single most important clinical message. (Section 5) 6. Among neuromodulation options, rTMS should be considered before ECT for most non-emergent bipolar depression. In pooled trial data, the odds of manic switch with rTMS did not reach statistical significance compared to sham (OR 1.3, 95% CI: 0.7-2.4), rTMS has an NNT of 8 for treatment response, and it avoids the cognitive effects of ECT. ECT remains the gold standard for life-threatening presentations (catatonia, severe suicidality, psychotic mania), with response achievable in a mean of 7 sessions for bilateral pulse and exceptional efficacy in mixed episodes (greater than 75% improvement in 63% of patients). Magnetic seizure therapy (MST) may eventually replace ECT, with comparable efficacy (86.4% vs. 95.0%) and superior cognitive preservation, but awaits FDA approval. (Section 6) 7. Digital mood monitoring shows promise but is not ready for clinical decision-making. The OXTEXT trial demonstrated false positive rates exceeding 90% in real-world deployment of smartphone-based episode prediction. App-generated data can supplement clinical judgment by identifying longitudinal trends (particularly in sleep patterns) but should not trigger treatment changes in isolation. Algorithmic bias, with training datasets that underrepresent minority populations and older adults, is an additional concern. (Section 7) 8. The pharmaceutical pipeline for bipolar disorder has mostly disappointed, but a few repurposed medications show enough promise in treatment-resistant cases to warrant familiarity. Most novel approaches (anti-inflammatory agents, TAAR1 agonists) have failed or stalled in rigorous trials. A small number of medications developed for other conditions (pramipexole, modafinil, supraphysiologic levothyroxine, NAC) have generated preliminary evidence as augmentation strategies for specific treatment-resistant phenotypes, but none have large-scale definitive bipolar trials, and all require careful patient selection and monitoring. The detailed evidence, including its limitations, is in Section 8 for clinicians considering these options. (Section 8) 9. The antidepressant question in bipolar depression remains genuinely unresolved, and transparency about this uncertainty serves patients better than false confidence. Major international guidelines (BAP, NICE, CANMAT, APA) diverge substantially on antidepressant recommendations. The 2023 Lancet Psychiatry network meta-analysis (Yildiz et al.) identified traditional antidepressant monotherapy as among the least efficacious approaches, while the Danish target trial emulation (Rohde et al., 2024) found no significant association with manic switching but also no therapeutic benefit. What is settled: antidepressant monotherapy without a mood stabilizer remains contraindicated in bipolar I. What is not settled: whether adjunctive antidepressants offer meaningful benefit beyond first-line options, and for whom. (Section 9) 10. If maintenance medication discontinuation is attempted, the method matters as much as the decision. In small studies, rapid discontinuation produced relapse rates as high as 94% versus approximately 53% with gradual tapering. The 2025 Joint Clinical Practice Guideline on psychotropic tapering (originally developed for benzodiazepines, with principles applicable to mood stabilizer discontinuation per the Horowitz & Taylor framework) recommends hyperbolic dose reductions (progressively smaller and slower decrements as serum levels approach zero) spanning months, not weeks. The 20-fold increase in suicidal behavior during the first year post-lithium-discontinuation is data that patients deserve to know before making this decision. (Section 9)

PsychHQ

Clinical Vignettes