The UK’s National Institute for Health and Care Excellence (NICE) currently advises that antidepressants should not be routinely offered as a first-line treatment for less severe depression ‘unless that is the person’s preference’.1,2 In this article we examine this 'patient preference clause’, arguing that while it may appear to affirm patient autonomy, aligning with the NHS’s broader commitment to shared decision making (SDM), it is not the same thing and, in practice, such guidance risks inflating unnecessary antidepressant prescribing, escalating healthcare costs, undermining evidence-based medicine, and exposing patients to avoidable and potentially long-term harms.
Below, we set out five arguments as to why, on ethical and clinical grounds, it is advisable for the ‘patient preference clause’ to be removed.
“... in practice, such guidance risks inflating unnecessary antidepressant prescribing, escalating healthcare costs, undermining evidence-based medicine, and exposing patients to avoidable and potentially long-term harms.”
Misinterpretation of shared decision making
The idea of SDM denotes a structured, evidence-informed dialogue in which a clinician’s expertise and the research evidence is integrated with a patient’s values and circumstances to arrive at a mutual decision.3 The idea of SDM is not the same as ‘patient preference’ (PP), which simply denotes an individual’s stated desire for a given treatment, and implies that ‘preference’, by itself, is sufficient to render a prescription appropriate. While there are occasions when a PP may be well-informed (being rooted in full awareness of evidence-based risks, limitations, and alternatives), this cannot be guaranteed, as patient preferences may also be shaped by prior experience or misleading cultural messaging that contravenes the evidence.
While PP may both inform and be a component of SDM, it is insufficient, by itself, to provide a robust justification for any clinical intervention. We therefore contend that NICE’s guidance, by allowing antidepressants to be prescribed on the basis of patient preference, cuts across the best intentions of SDM and risks allowing evidence-based practice to be superseded by a preference-led model of care; one that can undermine clinical judgement and best practice by endorsing a patient’s request even if that rests on limited understanding of treatment risks, benefits, and alternatives.
Reinforcement of medicalisation
Since the biomedical model of emotional/mental distress is the dominant model in mental health care, enabling PP may further reinforce the unnecessary medicalisation of ‘situational distress’, such as that arising from personal loss, unemployment, financial duress, social isolation, violence, abuse, relational breakdown, and so on. Such medicalisation of what anthropologists have called ‘social suffering’ aligns with a neoliberal vision of health care, which frames emotional suffering as rooted in individual failings or personal dysfunction (rather than in, say, adverse socio-cultural factors), thereby threatening to deflect attention from the social determinants of mental distress. This may reinforce the legitimacy of pharmaceutical interventions even when non-pharmacological options may be equally or more effective, or when the evidence base does not support pharmacological treatment, as is the case with antidepressants for ‘less severe depression’ (a phrase synonymous with ‘mild to moderate’ depression, which was NICE’s previous descriptor).
Furthermore, the patient preference clause also underestimates how PPs are not formed in a vacuum but are shaped by a multiplicity of factors, such as broader sociocultural and institutional contexts, time-constrained consultations, systemic underfunding of alternatives such as psychological therapy, pharmaceutical marketing, and pervasive bio-narratives (for example, the discredited ‘chemical imbalance’ theory) that portray antidepressants as necessary or curative. In privileging these preferences without sufficient interrogation of the cultural myths and practices propagating them, the guidance risks endorsing choices that are less based on informed consent than on questionable ideas, messaging, and expectations.
Erosion of alternatives
A large body of research demonstrates that patients’ preferences significantly drive physician prescribing across a broad range of medications, including antidepressants.4 For instance, in one study more than half of patients received antidepressants when requested, regardless of whether their symptoms warranted them;5 while in another, antidepressant prescribing for major depression was 76% when a specific brand was requested; 53% when a general request was made; and only 31% when not requested.6
When medication becomes the de facto first-line response to distress (simply because it is requested), other therapeutic modalities may be deprioritised. Such de-prioritisation, when scaled up, may intensify a self-perpetuating system where investment in therapy, community-based services, peer support, and primary prevention programmes further diminishes, making medication seem the only viable or accessible option. As a result, what appears to be an outcome of free choice in fact results from being subject to a system that has quietly eroded meaningful alternatives. At the same time, prescribers under increasing pressure to deliver quick solutions in time-limited consultations often default to medications as the path of least resistance. In this context, what is described as ‘preference’ may, for the clinician, reflect the internalisation of dominant medical and cultural norms; and for the patient, it may represent lowered expectations or resignation in the face of limited access to meaningful alternatives.
“While honouring a PP [patient preference] may enhance engagement and modestly improve satisfaction or adherence, there is no strong evidence that it improves clinical outcomes.”
No evidence of improved clinical outcomes
While honouring a PP may enhance engagement and modestly improve satisfaction or adherence, there is no strong evidence that it improves clinical outcomes. A 2024 meta-analysis of 26 randomised trials involving approximately 3670 patients with depression or anxiety found that treatment preference had no significant impact on symptom recovery (effect size d = 0.06, P = 0.16), despite small gains in satisfaction and adherence.7 In populations with multimorbidity (particularly older adults), patients often base choices on perceived risk rather than on realistic trade-offs, and may opt for treatments that increase the likelihood of harm, believing they offer greater benefit.8 As already mentioned, these risks are exacerbated when preferences are shaped by misleading marketing, cultural narratives, or outdated clinical models, influences shown to prompt inappropriate or even dangerous treatment decisions.9 Together, these findings indicate that while respecting PP is important, it cannot substitute for sound clinical judgement, especially when the evidence base is weak, where risks are significant, or where patient understanding is limited.
Conflict with professional standards and ethics
There is a regulatory and ethical contradiction at the heart of current NICE guidance. According to the British Medical Association (BMA), ‘A patient is entitled to drugs that the GP believes are necessary, not those which the patient feels should be prescribed.’
10 This principle places prescribing authority firmly with clinical judgement, not PP. The General Medical Council (GMC) also reinforces this principle in its prescribing guidance: ‘Sometimes, patients will ask for treatment or care ... If, after discussion, you still think the treatment or care would not serve the patient’s needs, you should not provide or propose it. You should explain your reasons to the patient and explore other options that might be available, including their right to seek a second opinion.’ Also, ‘You must only propose, prescribe or provide medicines, treatments or devices if it is safe to do so.’
11
Both the BMA and GMC make clear that clinicians are not obligated to prescribe medication in the absence of sound clinical justification, even if the patient requests it. This position is underpinned by several considerations: patient preferences may conflict with evidence-based guidelines, may be based on misinformation or may lead to harm. These concerns are especially relevant in light of mounting research on adverse antidepressant effects (including protracted withdrawal effects and long-term sexual dysfunction), not all widely known, leading many patients to request prescriptions without being fully informed about all potential risks.12 Since preferences can diverge from what is ethically or clinically appropriate, they must not override the medical duty to avoid doing harm (primum non nocere). Allowing PP to supersede clinical evidence risks eroding the role of medical expertise and the principle of non-maleficence, particularly if prescribers feel compelled to meet expectations that could result in long-term harm.
From preference to partnership
There are, of course, circumstances where patient preference may ethically support prescribing: when the patient is fully informed of evidence-based risks, limitations, and alternatives; when non-pharmacological options have been offered, tried, or declined; when symptoms are clearly distressing and impairing; and when adequate follow-up and monitoring are in place. However, such decisions must arise from SDM; not from deference to the principle of PP. In this sense, PP, while influencing ethical care, cannot stand alone as a reason for prescribing antidepressants, especially where evidence of efficacy is limited and the risk of harm is significant.
In this light, the NICE guidance risks conflating responsiveness and deference with genuine care. By allowing antidepressants to be prescribed solely based on PP, it undermines clinical judgement, medical ethics, and the principles of SDM.
We strongly recommend NICE revises its current guideline by removing the ‘patient preference’ clause. Its inclusion is not only inconsistent with evidence-based practice, but may actively contribute to overprescribing, rising healthcare costs, and avoidable harm. Treating mild and moderate depression and anxiety must involve resisting the appeal of quick pharmaceutical fixes and investing instead in long-term, non-medicalised, evidenced-based approaches to healing.
Notes
Provenance
Comissioned; externally peer reviewed.
Acknowledgements
AI support (ChatGPT) for literature search and minor editing assistance.
Competing interests
James Davies is Lead Researcher at the Beyond Pills All-Party Parliamentary Group, is a practicing psychotherapist, and receives royalties on book publications. John Read declares no competing interests.
- Received September 10, 2025.
- Revision received October 6, 2025.
- Accepted October 7, 2025.
- © British Journal of General Practice 2025
