A beloved, familiar self

Refusal of treatment is not simple ‘non-compliance’. Jamison suggests that this familiar clinical frustration is a reaction to preserve a beloved self. She articulates – with a precision that patients often cannot reach in consultation – that prescription can provide a stability that feels like diminishment.

Jamison’s refusal of lithium, sustained over years and at significant personal cost, is evidence of this point. She, like many of our patients, understood the diagnosis and accepted the evidence, yet still refused the prescription. In a tone of bereavement, she writes that to accept lithium was to relinquish a familiar self and its manic capacities, where ideas arrived with felt certainty and consequences seemed negotiable. She resisted treatment with the particular anger of someone required to mourn a self they had not chosen to lose.

A more accurate description than ‘non-compliance’ (or even ‘non-adherence’) is therefore ambivalence: the patient may want relief from symptoms but they also want to keep something that the illness seems to confer. Recovery-oriented writing has made a similar point: terms such as ‘non-compliance’ and ‘drop-out’ can obscure the meanings patients attach to discontinuing treatment (Roe Reference Roe and Davidson2017). Confrontation fails, because it speaks to the clinician’s model of reality. From the patient’s phenomenology, a clinician’s reality-testing does not register as reason, but dispossession.

Treatment non-adherence remains a significant unresolved problem, with a central estimate for bipolar disorder sitting around 40% (Jawad Reference Jawad, Watson and Haddad2018).

Yet the relapse-prevention conversation continues in the language of risk and evidence, as though non-adherence is a deficit of insight. Jamison argues that it must be conducted in the language of loss. The more accurate clinical work is motivational and existential: name what the patient believes medication will steal and acknowledge what the illness has stolen. The adherence conversation reframed in these terms becomes not persuasion, but a relationship built on shared decision-making. Motivational interviewing is useful here because it works with ambivalence rather than against it and resists the ‘righting reflex’ that Jamison’s account shows so clearly is ineffective (Miller Reference Miller and Rollnick2013).

This proposition has only become more pertinent. Through social media algorithms and, increasingly, artificial intelligence (AI)-generated health information, patients are exposed to medication discourse that can alter expectations before a prescription is written. Analyses of social media discussion of lithium, quetiapine and valproic acid on X (formerly Twitter) illustrate how narratives of medication harm and discontinuation circulate rapidly, potentially shaping adherence (Chart-Pascual Reference Chart-Pascual, Alvarez-Mon and Montero-Torres2025). We must recognise that the patient’s private argument against medication can be socially reinforced.

The ill clinician

The second discomfort in An Unquiet Mind is that it forces consideration of the clinician’s stigma. Writing the book – a public identification with mania, suicidality and psychosis – as a clinician and clinical professor is not merely ‘brave’; in the professional culture of the mid-1990s, this was career-jeopardising, with no precedent or safety net. The disapproval of a physician publicly admitting their illness, viewed as an affront to the required invulnerability of a physician, is not only a matter of history. That discomfort persists.

The clinician with lived experience of mental illness may bring something valuable to practice. But the diagnosed clinician is not, by virtue of illness, an impaired clinician. The first is diagnostic; the second is functional. To collapse them into one another is itself a form of stigma. Poor clinician well-being is associated with worse patient safety outcomes, and systems that rely on individual resilience rather than structural support are inadequate responses to that risk (Hall Reference Hall, Johnson and Watt2016). For psychiatrists and other mental health clinicians, the difficulty may be sharper still. Their work is conducted in the same conceptual territory as the illness they may disclose, so the fear is not only of being unwell, but of being judged as less credible or as less safe because of it.

Medical boards have historically posed punitive, overly diagnostic questions about psychiatric history – a barrier to physician help-seeking (Dyrbye Reference Dyrbye, West and Sinsky2017). Recently, there have been attempts to reduce the sense that disclosure automatically triggers punishment. In the UK, the General Medical Council (GMC) now requires disclosure of functional impact only (General Medical Council 2026). That is an improvement. But Jamison’s memoir – and the data – suggest that formal reassurance alone does not abolish informal peril. The legacy of those questions has shaped professional culture long after they were removed. Culture lags behind policy.

Today, Jamison would also face what she did not: permanent digital traces that are searchable, decontextualisable and subject to recirculation at a speed no author controls. What is said cannot be unsaid, nor contained. The trainee with lived experience of mental illness faces a version of Jamison’s dilemma at greater velocity and considerably higher exposure. A clinician’s history of mental illness may be irrelevant to competence if there is no active impairment. However, untreated illness, burnout or functional impairment may affect judgement and patient safety, particularly when clinicians work with highly vulnerable patients (Hall Reference Hall, Johnson and Watt2016). The correct response is proportionate, with early support and confidentiality where possible.

A caution against overgeneralisation

Memoirs carry a peculiar authority. They are persuasive precisely because they are not clinical texts: the texture of first-person experience, the specificity of detail and the literary exceptionalism of An Unquiet Mind in particular render it more compelling than any symptom checklist. DSM and ICD symptom criteria are, by design, syndromic and aggregative; they communicate what is necessary for reliable classification. Dimensional and transdiagnostic models may correct some of the blunter categorical boundaries, but overall can remain insufficient for a holistic clinical understanding of the patient. The phenomenological richness of Jamison’s first-person narrative does the opposite – the luminosity of mania turning dysphoric, the suicidal ideation, the specific texture of lithium refusal – conveys what syndromic lists cannot offer. This is its quite extraordinary value.

Yes, but. An Unquiet Mind renders a single phenomenology with uncommon vividness, and it perhaps risks being absorbed in the trainee’s working model as a representative account. Jamison’s mania is, in her own account, luminous, productive and suffused with creative and relational intensity. It is not the only mania. To absorb her account as the typical presentation is to misread the patient who does not resemble her.

The solution is to use it for what it is: one fully inhabited account of one illness, which sharpens the trainee’s capacity to resist syndromic flatness, rather than to pattern-match. The memoir is not a template; it is a corrective.

To conclude

An Unquiet Mind is something considerably hard to replace. It describes medication non-adherence, ambivalence within a mind struggling with mental illness, and clinician stigma from within the career it threatened. For trainees, it clarifies why confrontation fails and why concealment persists. Read carefully, and critically, three lessons, then, from a book that has been in print for 30 years without being properly absorbed into psychiatric training are: first, so-called non-adherence is often ambivalence, and requires a relational response that can tolerate loss, conflict and meaning. Second, institutional stigma is reproduced through training cultures, informal hierarchy and silence; it shapes how clinicians respond to illness in themselves and in colleagues, and will persist unless it is named and interrupted. Third, no diagnostic system yet devised closes the gap between classification and experience. Jamison illuminates the affective dimensions of the psychiatric profession. That is a highly thought-provoking series of messages to receive from one book.