What do Patients Wish Their Psychiatrist Had Asked—But Never Asked? A Narrative Review of the Unspoken Agenda in Psychiatric Consultations

1. Introduction

Every psychiatric consultation contains two agendas. One belongs to the clinician: clarify the diagnosis, quantify risk, titrate the medication, document the review. The other belongs to the person in the chair: Is this drug going to change who I am? Can I still work? Will this make me gain weight? Does what happened to me as a child matter? Am I allowed to say that I sometimes think about dying?

Three decades of communication research converge on a single, uncomfortable finding: the second agenda is routinely present but rarely voiced, and the silence is produced largely by the structure of the encounter rather than by patient reticence. Patients and psychiatrists disagree systematically about what matters most in an outpatient consultation, and it is the patient’s priorities that decide perceived service quality [1]. Poor communication and discontinuity remain among the most enduring complaints of mental health service users [2], and inadequate information, with consequent exclusion from decisions, is a persistent theme in their accounts [3].

The clinical stakes are established. Most studies relating physician–patient communication to health outcomes demonstrate a positive association [4]; quality communication strengthens trust, emotional regulation and treatment adherence [5], and adherence improves when physicians receive communication training [6]. Yet performance is uneven even in well-resourced settings: in one hospital study only 33% of physicians were rated excellent on all four assessed communication behaviours [7].

This review does not add another lament about poor communication. It answers four specific questions: (i) what do patients wish their psychiatrist had asked but never asked; (ii) why do these questions go unasked; (iii) what follows from the unasked question; and (iv) what works to close the gap. Fig. 1 summarises the conceptual model that organises the synthesis.

Figure 1. Conceptual model of the unspoken agenda in psychiatric consultations. The patient’s and the clinician’s agendas pass through a consultation filter shaped by five suppressing mechanisms; what is filtered out becomes the unasked question, with five categories of consequence. Evidence-based remedies act by re-opening the filter.

Figure 1. Conceptual model of the unspoken agenda in psychiatric consultations. The patient’s and the clinician’s agendas pass through a consultation filter shaped by five suppressing mechanisms; what is filtered out becomes the unasked question, with five categories of consequence. Evidence-based remedies act by re-opening the filter.

2. Methods

2.1. Design

We conducted a narrative, evidence-based review, reported in accordance with the Scale for the Assessment of Narrative Review Articles (SANRA) [8]; the completed checklist is provided as Supplementary File 1. A narrative rather than systematic design was chosen deliberately, because the question spans conversation analysis, service-user experience, epidemiology and intervention research, whose designs and outcomes cannot be meaningfully pooled.

2.2. Sources and Selection

PubMed/MEDLINE, Scopus, the Cochrane Library and Google Scholar were searched for English-language literature to November 2025, combining terms for psychiatrist–patient communication, shared decision making (SDM), information needs, unmet needs, disclosure and concealment, sexual side effects, physical health, suicidal ideation, trauma enquiry, stigma, coercion, question prompt lists, decision aids and routine outcome monitoring. Reference lists of key reviews were hand-searched (backward and forward citation chaining).

Four complementary evidence streams were integrated: (i) conversation-analytic and observational studies of naturally occurring consultations, which analyse what participants do rather than what they later report [9,10]; (ii) quantitative surveys and qualitative studies of service-user experience, satisfaction and unmet need [11,12]; (iii) systematic reviews and meta-analyses of communication, alliance and adherence [13,14,15]; and (iv) randomised trials and Cochrane reviews of communication-focused and decision-support interventions [16,17,18]. Systematic reviews, trials and widely cited primary studies were prioritised for every substantive claim; smaller studies were retained only where they supplied the sole direct evidence on a specific question. Findings were synthesised thematically into mechanisms, domains, consequences and interventions.

A caution frames all that follows. Patients are not a homogeneous block: preferences for involvement vary with personality, background, experience, problem, context and even the mood of the day [19]. Nothing here argues that every patient wants the same conversation; the argument is that the offer of the conversation should be routine and explicit.

3. Results

3.1. Why the Question Goes Unasked: Five Mechanisms

Question design. The most actionable finding in this literature is that clinicians’ question design constrains what patients can say. “Any questions?” is built not to elicit further talk, whereas “Do you have some questions?” invites it; “No problems with the medication?” invites confirmation that there are none, making concerns hard to raise [10]. In consultations with patients with psychosis, patients repeatedly attempted to talk about the content of their symptoms and were deflected, generating interactional tension [20]. Many “unasked” questions were therefore asked, and then closed down. The collaborative clarification of meaning (“repair”) is associated with adherence [21], and it is trainable [18].

Divergent agendas. In 200 consecutive psychiatric consultations, between 21% and 57% of patients, depending on the clinic, asked no questions at all [22]. In a large Japanese survey, what patients most wanted their psychiatrists to change was behavioural rather than therapeutic: to stop “only watching the computer” and to build trust [23]. The same two-sided “don’t ask, don’t tell” dynamic, clinician non-enquiry plus patient non-disclosure, has been described for insomnia [24] and recurs across the domains below.

Power and prior coercion. Thirty-four per cent of voluntary patients felt coerced into admission, half still did so a month later, and no patient characteristic predicted it; the drivers were not participating in decisions and not feeling respected [25]. Coercion research has long operationalised the patient’s experience as a lack of control, choice, influence and freedom [26], and informal coercion, through persuasion, leverage, threat and repeated requests, is pervasive and often invisible to clinicians, irrespective of legal status [27]. Its communicative legacy is durable: negative experiences of psychiatric power, such as involuntary admission, prevented patients from expressing preferences years later [28]. The question a patient most needs to ask may be precisely the one that a previous episode of coercion taught them not to ask.

Stigma and self-silencing. Public stigma operates through stereotypes, prejudice and discrimination [29], and when internalised produces the “why try” effect of diminished self-esteem and self-efficacy [30,31]. People with mental illness thereby face reduced life opportunities over and above the impairments of the disorder itself [32]. Self-stigma reduces help-seeking, is highest among people with schizophrenia, and is on average higher in South-East Asian and Middle Eastern than in European and North American countries [33]. A patient with high self-stigma will not spontaneously raise sexual dysfunction, suicidal thoughts or childhood abuse; the invitation must come from the clinician.

Time, structure and misperception. Time pressure is the barrier to SDM that clinicians cite most often [34], and psychiatric prescribers perceive SDM visits as longer than standard appointments [35]. The evidence does not support the perception: a Cochrane review of SDM interventions in mental health found no effect on consultation length [34], and clinicians who embed SDM in routine practice report that it takes no extra time [36]. The attitudinal barriers of “my patients don’t want it” and “I already do it” [28] are similarly contestable, because patients and providers both find SDM acceptable [37].

3.2. What Patients Wish Had Been Asked: Eight Domains

Table 1 maps each domain to a representative question, the documented gap and the key evidence.

Medication: purpose, duration and exit. The question patients most often manage to ask is about stopping: in 200 consultations, most wanted to know when medication could be reduced or stopped [22]. The deeper question, what the drug is doing and what the plan is, is seldom addressed. Over 60% of patients with bipolar disorder were dissatisfied with information about side-effect risks, including effects on their sex lives, and lower satisfaction with information was associated with lower adherence [11]. Clinicians overestimate their own performance: 57% said they always or nearly always offered a choice of medication, but only 34% of service users reported being offered one, and only 24% received enough information [12]. Patients who understand the purpose of a prescription are twice as likely to collect it, and patients recall about half of what they are told [54]. The stakes are high. Non-adherence affects more than one-third of people with schizophrenia each year, is often covert and is underestimated by clinicians [55]; discontinuation is followed by relapse in 77% of patients within one year and 90% within two, against 18% among adherent patients, with risk rising after as few as 10 days of missed medication [56]. Discontinuation rates for oral antipsychotics range from 26% to 44%, and up to two-thirds of patients are at least partially non-adherent [57]. Tellingly, the recommended remedies are conversational: shared decision making, regular assessment of adherence, management of side effects and a positive therapeutic alliance [55].

Sexual function and intimacy. This is the most clearly documented unasked question. Among severely mentally ill outpatients, 62.5% of men and 38.5% of women believed their medication caused sexual side effects, yet 50% never or infrequently discussed sexual functioning with their provider, 80% of affected women had never raised it, and 41.7% of men and 15.4% of women had stopped medication because of it [38]. Detection depends on asking: in 344 patients taking selective serotonin reuptake inhibitors (SSRIs), sexual dysfunction was reported by 58% when clinicians asked directly but by only 14% spontaneously [39]. Meta-analytic evidence confirms that treatment-emergent sexual dysfunction ranges from 15% to 80% across trials and is under-reported when no structured enquiry is used [58], with rates of up to 80% for SSRIs in some series [59,60], and systematic enquiry has been recommended because dysfunction is an unrecognised cause of non-compliance [40]. Patients wait for the clinician to raise the subject; when clinicians regard it as low priority or fear that honesty will encourage non-adherence, the burden of disclosure returns to an embarrassed patient [61].

Physical health, weight and metabolic risk. Mortality in serious mental illness is two to three times that of the general population, largely from cardiovascular disease driven by inactivity, diet, smoking and antipsychotic-related weight gain [42]; over 60% of people with schizophrenia smoke [43]. Service users want this addressed: they are concerned about weight and fitness, uncertain why assessments are done, report refused requests for help with weight gain, and expect clinicians to act proactively when illness compromises self-care [41]. Every patient should have a professional responsible for their physical health [62], and screening triggered by diagnosis or antipsychotic use is recommended [42,63].

Suicidal thoughts and self-harm. Approximately 45% of people who die by suicide consult a primary care physician in the preceding month, rarely with documented enquiry or disclosure [45]. Concealment is calculated rather than merely shameful: 70% of psychotherapy clients who concealed suicidal ideation feared practical consequences, chiefly involuntary hospitalisation, and nearly half would disclose only if that threat were reduced [44]. Across samples, people withholding information from clinicians cite embarrassment, fear of judgement, reluctance to accept follow-up treatment and not wanting it in their record [64]. The patient’s unasked question is therefore twofold: what happens if I tell you? and will you still treat me as a person afterwards?

Trauma and adversity. Among mental health service users, 47% have experienced physical and 37% sexual abuse [46]. Patients are comfortable being asked: most primary care patients accepted direct or questionnaire-based trauma enquiry and its inclusion in the record [47]. Yet 35 of 96 service users with childhood adversity were never asked [48], and unrecognised trauma exposes patients to re-traumatisation through removal of choice, judgemental responses, seclusion and restraint [46]. Facilitators of disclosure lie largely with the professional and the setting rather than the patient [48]. When disclosure occurs, recommended responses are specific: affirm that disclosing was right, do not probe for details, ask whether anyone has been told before and how that went, check current safety and emotional state, and offer follow-up and referral [65].

Work, money, housing and company. Unmet needs cluster outside symptoms: daytime activity, company, intimate relationships, employment, housing and money [49,66]. Patient–provider agreement on individual needs is only slight to fair [50], and only meeting client-stated needs has been shown to improve quality of life [67]. The needs perspective matters clinically, not only ethically: assessing needs requires the patient’s own view of what is unmet, turns assessment into negotiation, and better staff–patient agreement on needs independently predicts treatment outcome [66]; comprehensive needs assessments in severe mental illness likewise foreground daily functioning, work, education and social inclusion [68]. Although as many as 90% of people with mental health problems would like to work [51], only 15% of people with serious mental illness in the UK are employed [69]; the relationship is bidirectional, since poor mental health predicts later job loss and job loss predicts poorer mental health [70], and supported employment is an evidence-based intervention in its own right [71,72].

Identity and stigma. Person-centred care rests on the principle that people with severe mental illness are people, with histories, strengths and values [73]. When Dutch outpatients defined good care, they added elements absent from academic frameworks, including “professionals listen without judgment” and “patients are seen as persons with positive sides and strengths” [52]; some argue that help can only ever be defined by the person receiving it [74].

The encounter itself. In 80 routine psychiatric consultations rated with the OPTION (“observing patient involvement”) scale, the item assessing whether the clinician explored the patient’s preferred approach to receiving information was scored at the lowest level in every one [53]. Patients prioritise being listened to, explanation and reassurance over technical intervention: in Karachi, being listened to ranked first among the issues most important to patients [75], and in Polish ambulatory care, expectations of explanation and reassurance (mean 9.42) and emotional support (9.39) exceeded those of investigation and treatment (7.64) [76]. Expectations also vary by sex, socio-economic status and education [77]. Patients describe adequate information as “basic courtesy”: “I wasn’t told what the medication did positively or negatively – you were just told to take it” [78]. On acute wards, patients and relatives ranked information provision as a higher priority than most professionals did [3].

3.3. What Patients Want When They Are Asked

Patients seek information sufficient to feel respected and to take part, not exhaustive technical detail [78,79]. SDM, endorsed as the standard of care by the National Academy of Medicine, the National Institute for Health and Care Excellence, the American Psychiatric Association and SAMHSA [80], yields benefits in mental health comparable to those in general medicine [37]. For patients who take medication inconsistently, SDM is best understood as a long-term endeavour embedded in the therapeutic partnership rather than a single transaction [81].

A counterweight is essential. Some patients want the psychiatrist to decide [75], and only 7% of patients with severe mental illness answered “no” to a National Health Service community survey question on whether they had been involved as much as they wanted [19]. The resolution is procedural rather than substantive: whether a patient wishes to lead, share or delegate is itself a question that must be asked, and it is the question the OPTION data show is almost never asked [53].

Relational qualities complete the picture. The alliance, comprising agreement on goals and tasks and an emotional bond, robustly predicts outcome, largely through the clinician’s facilitative interpersonal skills [14]; in psychosis, task-oriented elements (collaboration, agreement on tasks, regular contact) appear most active for adherence [15], although the outcome evidence is less consistent [13]. The therapeutic relationship may act both through adherence and directly as a therapeutic ingredient [82], and clinical empathy is a core professional competence [83]. Trials in which practitioners enhanced how they expressed empathy have reported reduced pain and anxiety and improved satisfaction and adherence [84], and a theoretical model now specifies how empathic communication produces these effects [85]. Yet psychiatrists receive little specific instruction or supervision in communication [82].

3.4. Consequences of the Unasked Question

The consequences are measurable across five domains (Fig. 2). Concealed risk: suicidal ideation and trauma histories remain hidden when enquiry is absent or its consequences are opaque [44,45,48]. Non-adherence and relapse: 41.7% of men with perceived sexual side effects stopped medication [38]; in the CATIE study, 74% of patients discontinued medication prematurely, and 38% of surveyed patients stopped antipsychotics without telling their psychiatrist [21]; non-adherence in first-episode schizophrenia carries a roughly fivefold increase in relapse risk [86]. Unmet need and reduced quality of life: needs identified by patients but not clinicians remain unmet [50,66,67]. Premature mortality: physical health questions not asked become cardiovascular deaths not prevented [42,62]. Disengagement and perceived coercion: not feeling heard is itself a driver of perceived coercion [25], and its effects persist for years [28].

Figure 2. The measured gap. (A) Paired comparisons showing that direct questioning multiplies detection of sexual dysfunction [39] and that clinicians overestimate the choice and involvement they offer [12]. (B) Indicators of the unasked question [38,44,48,53] and of information sufficiency [12]. Values are drawn from different populations and are not directly comparable.

Figure 2. The measured gap. (A) Paired comparisons showing that direct questioning multiplies detection of sexual dysfunction [39] and that clinicians overestimate the choice and involvement they offer [12]. (B) Indicators of the unasked question [38,44,48,53] and of information sufficiency [12]. Values are drawn from different populations and are not directly comparable.

3.5. Interventions That Change The Encounter

Six intervention families have supporting evidence (Table 2). Structuring the consultation gives the patient’s agenda a guaranteed slot: DIALOG+, in which patients rate satisfaction with eight life domains and three treatment aspects and then choose what to discuss, improved quality of life in a cluster randomised trial [87] and has been associated with reduced treatment costs [88]. Preparing patients with question prompt lists or coaching produces small but consistent increases in question-asking and satisfaction [16,89], particularly when the clinician endorses the list [90], and does not lengthen consultations [91]. Decision aids improve knowledge, risk perception and values-consistent choice while reducing decisional conflict [17]. Training clinicians works when it is specific: in the TEMPO cluster randomised trial, psychiatrists trained to explore patients’ psychotic experiences used 44% more self-repair and achieved better therapeutic relationships [18], and simulation-based education improved trainees’ agenda setting and empathic skills [92]; generic communication training, however, has shown only modest effects on satisfaction with the relationship and none on mental state [93], and it has been called staggering that, a century after talking therapies emerged, so few trials examine how the conduct of consultations affects outcomes in severe mental illness [94]. Trauma-informed enquiry is explicit and manualisable: ask everyone, at initial assessment rather than in crisis, with a normalising preface, specific examples, the person’s own pace and an explicit right to decline [65]. Routine outcome monitoring improves outcomes and reduces dropout, with the greatest benefit for patients who are not improving [95,96,97], precisely the group least likely to volunteer their concerns.

4. Discussion and Conclusions

4.1. Discussion

This review shows that “what patients wish their psychiatrist had asked” has an evidence-based answer that is narrower, and more actionable, than the familiar complaint of poor communication. Eight domains recur across designs, countries and diagnoses, and the gap in each is quantified rather than anecdotal. Three insights reshape the problem. First, many unasked questions were in fact attempted by patients and closed down by question design [10,20]; the target is therefore clinician behaviour, not patient courage. Second, clinicians consistently overestimate their own performance [12], so self-report cannot serve as the measure of improvement. Third, the principal barrier, time, is largely perceived rather than real [34,91].

The findings also reframe what “asking” means. The question is bidirectional: what patients wish had been asked is inseparable from whether the clinician demonstrably works to understand what the patient does say, which is why self-repair, a measurable index of communicative engagement, rose with specific training [18]. It is equally procedural: because preferences for involvement differ [19,75], the first question is not “what would you like to decide?” but “how would you like us to decide?”, the behaviour rated at the lowest possible level in all 80 recorded consultations [53]. These principles are low-cost and portable, which makes them especially relevant to high-volume, time-constrained services in low- and middle-income countries, where self-stigma is on average higher [33].

Five research priorities follow. (i) Quantify the counterfactual: trials should test whether routinely asking a fixed set of questions changes adherence, disclosure and outcome, since current evidence is largely correlational [93]. (ii) Specify mechanisms: conversation analysis already provides sensitive, systematic measures of shared understanding [9,21], and specific behaviours should replace umbrella concepts [19]. (iii) Measure what clinicians do, not what they say: individual clinician style explains more variance in involvement than patient or structural factors [10]. (iv) Test the time objection empirically with objective consultation-length feedback [34,35]. (v) Build equity into enquiry: because self-stigma, unmet need and dissatisfaction are concentrated among the most disadvantaged [33,50,77], interventions that shift the burden of asking onto patients risk widening inequality; clinician-side change deserves equal weight [16,89].

The review has limitations. It is narrative rather than systematic, so selection bias cannot be excluded, although triangulation across four evidence streams mitigates it. Prioritising highly cited work favours older studies. Heterogeneity is substantial, for example estimates of antidepressant-associated sexual dysfunction range from 15% to over 80%, largely due to measurement [58,59]. Much of the evidence derives from high-income Western settings; the available Asian data suggest that both the content of unspoken questions and the acceptability of asking them are culturally patterned [23,33,50], which matters for low- and middle-income countries, including India, where consultation volumes are high and time is scarce. Finally, preference heterogeneity is real [19,75]: the defensible conclusion is not that all patients want SDM, but that all patients should be asked how they want information and decisions to be handled.

4.2. Conclusions

The unasked question is not an incidental lapse of courtesy but a systematic, measurable and clinically consequential feature of psychiatric practice, linked to concealed risk, non-adherence, relapse, unmet need, preventable physical morbidity and disengagement. Remedies with trial evidence already exist. The deficit is one of implementation, not knowledge, and it lies in the design of the routine consultation, which is the most tractable object of reform in psychiatric practice.

4.3. Practice Implications

Clinicians should (1) replace closed, disconfirming questions (“No problems with the medication?”) with open invitations (“What concerns do you have about the medication?”); (2) ask every patient how they wish information and decisions to be handled; (3) enquire directly and routinely about sexual function, weight and smoking, work and money, and trauma, using the questions in Table 1; (4) explain transparently what happens after disclosure of suicidal thoughts before asking about them; and (5) look up from the screen. Services should embed a structured patient-agenda tool (such as DIALOG+ or a question prompt list) and routine outcome monitoring, and should train clinicians in specific, observable communication behaviours rather than generic skills.