Language and implicit assumptions

Words do not merely describe reality; they organise it. Insisting on linguistic seriousness is not semantic pedantry; it is an argument about respect, clarity and epistemic power.

We argue that in contemporary clinical and scientific psychiatry talk, the phrase self-medication often functions less as a testable mechanism and more as a metaphor, with ethical and epistemic implications.

The use of prescribed medications and non-prescribed psychoactive substances is interpreted differently across communities, substances and contexts. The same or similar psychoactive effect may therefore acquire different meanings depending on whether it occurs within a prescription, self-directed use of medication, recreational use, ritual practice or an illicit market. Within healthcare systems, prescribed medication is ideally embedded in an institutionally authorised clinical practice involving assessment, diagnosis, prescription, professional accountability, regulatory safeguards and therapeutic expectations. ^{Reference Cruess and Cruess1} By contrast, substance use is often interpreted through less stable and more contested frameworks, including individual choice, dependence, risk, illegality and negative prognostic attributions, such as loss of control, relapse, chronicity, reduced agency or difficult treatability. ^{Reference Merrick, Louie, Cleary, Molloy, Baillie and Haber2,Reference van Boekel, Brouwers, van Weeghel and Garretsen3}

This distinction is not absolute. Prescribed medication can be ineffective, harmful, misused or shaped by commercial and institutional interests; conversely, non-prescribed psychoactive substance use may be socially accepted, culturally meaningful, subjectively helpful or embedded in collective practices. Nevertheless, the word medication generally carries a therapeutic and institutional register. When attached to substance use through the term self-medication it may import that register into behaviours whose meanings are multiple and not necessarily medical. The resulting narrative can sound empathic – they are coping – and mechanistic – they are correcting symptoms – while quietly narrowing psychological formulations about the patients’ motives, contexts and constraints. This narrative also appears intuitively coherent, yet its coherence is largely rhetorical.

Our argument is not that self-medication never occurs, nor that clinicians routinely misuse the term, but that the term should not be asked to carry more explanatory and clinical meaning than the evidence can support.

What the self-medication paradigm claims and what it leaves vague

The self-medication hypothesis was influentially formalised by Khantzian (1997) as a psychodynamic hypothesis, suggesting that individuals use specific substances to relieve painful and intolerable affective states. ^{Reference Khantzian4} Indeed, high comorbidity between mental disorders and substance use disorders has been consistently documented and linked with worse outcomes than either alone. ^{Reference Burdzovic Andreas, Lauritzen and Nordfjaern5} Yet high co-occurrence does not specify mechanism. Self-medication can be a patient-valid description and sometimes a risk marker, but it is a poor default mechanism.

If self-medication is treated as more than a metaphor, it implies at least some empirical assumptions:

1. (a) Intentionality: the person primarily uses a substance to manage an initial problem, such as distress, insomnia, anxiety or inattention. This may include continuing medication beyond medical advice or obtaining non-prescribed substances when access to prescribed treatment is limited. However, although some substance use may be motivated by attempts to manage negative affect or other perceived problems, this cannot be assumed for all people, all substances or all episodes of use. ^{Reference Sjödin, Molander, Ingesson-Hammarberg and Hammarberg6}

2. (b) Specificity: the substance is preferentially selected because of a perceived fit between its pharmacology and the symptom profile (e.g. stimulants for inattention). This is one of the most problematic assumptions, since drug choice is influenced by many factors, including availability, cost, social context, habit, expected pleasure, peer networks, previous exposure and polysubstance use. ^{Reference Han, Volkow, Jones, Dowell, Baldwin and Einstein7}

3. (c) Perceived benefit: the substance is experienced as helping to manage the perceived problem. Even prescribed medication does not invariably improve symptoms or functioning; however, perceived short-term benefit does not, by itself, establish sustained symptom improvement, functional benefit or therapeutic specificity. ^{Reference Faraone, Rostain, Montano, Mason, Antshel and Newcorn8}

4. (d) Temporal direction: symptoms precede use, and use persists because it is experienced as symptom-relieving. This temporal sequence may occur in some cases. However, it should not be generalised to all patients or all forms of substance use, especially when mental health symptoms, intoxication, withdrawal and social adversity may influence one another bidirectionally over time. ^{Reference Robinson, Sareen, Cox and Bolton9}

The core scientific problem is that the paradigm often mixes motive, mechanism and clinical interpretation without keeping these levels distinct. A patient may report using a substance to cope; this does not by itself establish why the disorder and substance use co-occur, nor whether the substance has therapeutic meaning.

Self-medication is commonly treated as established despite limited direct, prospective, mechanism-specific evidence. Some longitudinal epidemiological studies have supported self-medication as a predictor of later substance use disorder. ^{Reference Robinson, Sareen, Cox and Bolton9,Reference Broman, Wright, Broman and Bista10} Conversely, ecological momentary assessments (EMAs) of motives of substance use research have cautioned that trait-like motive questionnaires and momentary motives may not measure the same construct. ^{Reference Votaw and Witkiewitz11} Taken together, these findings suggest different risk pathways, of which self-medication is only one of the possible subjective accounts.

Epistemic inflation

Earlier reports have raised concerns about the limited empirical support for broad applications of self-medication to explain psychoactive substance use among individuals with mental disorders. ^{Reference Lembke12,Reference Levy13} Still, the narrative remains a recognisable explanatory frame in patient accounts, clinical discourse and research literature. ^{Reference Merrick, Louie, Cleary, Molloy, Baillie and Haber2,Reference Turner, Mota, Bolton and Sareen14} We argue that the narrative carries an inferential risk, as it can make one possible interpretation of substance use appear more coherent than the available evidence warrants.

The self-medication story tends to overgeneralise from three kinds of findings: ^{Reference Robinson, Sareen, Cox and Bolton9,Reference Carrigan and Randall15,Reference Hien, Jiang, Campbell, Hu, Miele and Cohen16} (a) people endorse coping motives, (b) coping is associated with worse outcomes and (c) longitudinally, self-identified self-medication predicts later disorder.

Acute psychoactive relief may reflect withdrawal relief, sedation, stimulation, cognitive narrowing or short-lived affect suppression with long-term rebound. Similar acute relief may also occur with prescribed psychoactive medication, for example during a short course of benzodiazepines. However, in clinical prescribing, expected benefits are ideally situated within the therapeutic relationship, in which indication, formulation, dose, duration, monitoring, adverse effects and functional outcomes, can be made explicit and reviewed. ^{Reference Marchi, Venturi, Visentini, Pinelli, Priebe and Galeazzi17} By contrast, self-medication reasoning can retrospectively convert subjective relief into evidence that the substance was treating an underlying disorder. ^{Reference Khantzian18}

The smoking literature illustrates this inferential risk, particularly when the self-medication hypothesis is applied loosely in clinical discourse. Smokers often experience cigarettes as psychologically beneficial, yet a large systematic review and meta-analysis found cessation is associated with improvements in depression, anxiety, stress and quality of life, suggesting that perceived self-treatment can sometimes be misattributed relief from withdrawal cycles rather than symptom remediation. ^{Reference Taylor, McNeill, Girling, Farley, Lindson-Hawley and Aveyard19} Therefore, acute relief can be clinically plausible and pharmacologically real, but it does not by itself establish therapeutic action, disease specificity or diagnostic meaning.

Agency and paternalism: ‘fun’ is not an afterthought

The phrase self-medication sounds patient-centred, but it can be patronising in at least two ways.

First, it attributes a medical logic to behaviour that may be driven by many logics: pleasure, belonging, curiosity, performance pressure, habit, despair or a varying mixture of these. Empirical work on motives for substance use repeatedly found that coping is only one motive among several, and not always the dominant one. For alcohol, a large meta-analysis (229 studies; >130 000 participants) found that enhancement motives – drinking to increase positive affect – and coping motives – drinking to reduce negative affect – are among the strongest predictors of drinking problems, with enhancement motives especially strongly associated with alcohol use itself. ^{Reference Bresin and Mekawi20} Therefore, pleasure seeking and positive reinforcement are central, often as central as coping.

Second, self-medication frames the patient as a quasi-clinician who attempts treatment – but also as someone who does so incompetently. The patient is simultaneously credited with insight – they are treating symptoms – and discredited as unsafe – they choose the wrong drug. This ambivalence is not neutral; it structures the clinical encounter and flattens clinicians’ curiosity about patient experiences. Clinical knowledge may make the term self-medication descriptively more appropriate in some cases, for example when a clinician uses a psychoactive medication for themselves. However, knowledge alone does not eliminate concerns about objectivity, nor does it replace clinical formulation. Literature on psychiatrists’ self-treatment suggested that, despite professional knowledge, self-medication may be shaped by concerns about confidentiality, social and professional stigma and help-seeking attitudes. ^{Reference Balon21} Thus, even in informed users, self-medication can remain a clinical interpretation that risks overlooking complex motives.

Stigma

Self-medication can be invoked to reduce blaming accounts of substance use. In the addiction field, the brain disease model has similarly been promoted partly as a destigmatising framework. ^{Reference Buchman and Reiner22} However, medicalised causal stories do not necessarily reduce stigma in a straightforward way. Previous work suggested that neurobiological narratives may reshape, rather than remove, stigmatising identities, whereby reduced blame may coexist with greater prognostic pessimism, perceived dangerousness or assumptions of reduced agency. ^{Reference Hall, Carter and Forlini23}

If self-medication is used as a quasi-biological story – e.g. dopamine deficit to cocaine as correction – it may inadvertently increase perceptions of chronic defectiveness, while obscuring social determinants and personal agency.

Explanatory distraction and shadowing of real problems

Overpromoting self-medication risks funnelling research into a single motivational lane (i.e. negative affect relief), while underweighting positive reinforcement, social learning, cue reactivity, habit formation, trauma-related processes, socioeconomic and network determinants and the pharmacology of withdrawal and rebound. It distracts from competing mechanisms and from multi-level causation. This narrowing is difficult to justify given evidence that substance-use motives are heterogeneous, ^{Reference Sjödin, Molander, Ingesson-Hammarberg and Hammarberg6} momentary motives may differ from trait-like reports ^{Reference Votaw and Witkiewitz11} and patterns of use challenge substance-specific explanations. ^{Reference Han, Volkow, Jones, Dowell, Baldwin and Einstein7}

Shadowing may also occur at the level of suffering. If drugs are treated as medication, clinicians may under-attend to what else is being managed: trauma symptoms, shame, loneliness, precarious housing, workplace demands or other personal problems.

Inadvertent legitimation of substances as coping tools

From an ethical perspective, we should be aware of the risk of iatrogenic communication. Overreliance on labelling and suggestive language in psychiatry and psychotherapy may lead patients to literalise the metaphor such that it permanently transforms their self and world perceptions into fixed views. ^{Reference Boisvert and Faust27} Additionally, the descriptor may promote excuse making, in that individuals may avoid responsibilities and justify their behaviour through evoking their psychiatric condition. ^{Reference Levy13}

Diagnostic drift

The self-medication narrative can promote diagnostic shortcuts: stimulant use to ADHD, cannabis use to psychosis vulnerability and alcohol use to social anxiety. This can produce overdiagnosis – by diagnosing a comorbid disorder based on the substance used – and missed diagnosis – by attributing problems to self-medication rather than assessing them on their own terms.

Additionally, interpreting substance use as a misguided attempt to treat an underlying illness may create a discursive environment where pharmacological substitution feels logically mandatory and overprescription becomes plausible, with consequent ethical and social implications. ^{Reference Faraone, Rostain, Montano, Mason, Antshel and Newcorn8}

This can happen as a result of a fundamental confusion between two conceptual models of psychopharmacology: the drug-centred model and the disease-centred model. ^{Reference Moncrieff and Cohen28} When a patient experiences temporary relief from distress or inattention after using a stimulant (e.g. cocaine), this is a predictable, non-specific psychoactive effect of the drug on mental states (the drug-centred perspective). However, self-medication reasoning often retrospectively filters this altered state through a disease-centred lens, erroneously assuming the substance is bridging a pathological deficit or correcting an underlying chemical imbalance. ^{Reference Moncrieff, Cooper, Stockmann, Amendola, Hengartner and Horowitz29} This rationale is then used to justify the prescription of medically substitute medications. Ultimately, reducing the complex cycle of addiction to a missing neurochemical deficit often leads to suboptimal, if not unsafe, pharmacological outcomes.