Introduction
Schizophrenia is considered a severe and chronic mental illness with complex symptomatology and a prevalence of ~1% in the general population and is associated with a significant societal burden [Reference Solmi, Seitidis, Mavridis, Correll, Dragioti and Guimond1]. The aetiology of schizophrenia is complex [Reference van Os, Rutten and Poulton2] and, in general, the clinical presentation of the illness is very heterogeneous. While some of the symptoms may be visible during the premorbid phase [Reference Hemager, Plessen, Thorup, Christiani, Ellersgaard and Spang3], in most cases, symptoms evolve in adolescence with the prodromal stage of the first acute psychotic episode. The description of prodromal states has been extensively detailed since the inception of the proto-concept of schizophrenia, that is, dementia praecox, by Kraepelin, while operationalizing these symptoms into reliable diagnostic criteria has remained a challenge. The International Classification of Disorders, 11th revision, describes a prodromal phase that often precedes the onset of psychotic symptoms by weeks or months under schizophrenia, 6A20.Z Schizophrenia, episode unspecified. The characteristic features of this phase often include “loss of interest in work or social activities, neglect of personal appearance or hygiene, inversion of the sleep cycle, and attenuated psychotic symptoms, accompanied by negative symptoms, anxiety/agitation, or varying degrees of depressive symptoms” [4]. As for the Diagnostic and Statistical Manual, 5th revision, despite three decades of active clinical research in the field of prodromal/at-risk mental states, it was only in 2013 that the notion of attenuated psychosis syndrome was included [5] in Section III under “conditions for further study” with the aim of recognizing psychosis risk syndromes for treatment to prevent transition to psychosis in ultra-high-risk groups [Reference Tsuang, Van Os, Tandon, Barch, Bustillo and Gaebel6, Reference Carpenter and van Os7]. Besides the original ultra-high-risk state concept that encompasses genetic risk, brief limited intermittent psychotic symptoms, and attenuated psychotic syndrome [Reference Yung, Phillips, Yuen and McGorry8, Reference Yung, Yung, Pan Yuen, PD, Phillips and Kelly9], the broader, evidence-based notion of clinical risk for psychosis incorporates two additional criteria derived from “the basic symptom approach,” that is, cognitive-perceptive basic symptoms and cognitive disturbances [Reference Schultze-Lutter10, Reference Schultze-Lutter, Ruhrmann, Fusar-Poli, Bechdolf, Schimmelmann and Klosterkotter11].
The first psychotic episode is usually characterized by an acute phase with prominent positive, cognitive, and psychomotor symptoms, followed by a subacute phase dominated by negative, cognitive, and depressive symptoms, after the remission of positive and psychomotor symptoms has been achieved with pharmacological treatment. Thereafter, alternating periods of acute psychotic episodes with periods of full or partial remission occur. Each phase of the illness poses a different set of symptoms and treatment challenges, best described by the staging model of schizophrenia [Reference Martínez-Cao, de la Fuente-Tomás, García-Fernández, González-Blanco, Sáiz and Garcia-Portilla12]. Several staging models have been developed [Reference Martínez-Cao, de la Fuente-Tomás, García-Fernández, González-Blanco, Sáiz and Garcia-Portilla12, Reference McGorry, Nelson, Goldstone and Yung13] but only one is validated [Reference Berendsen, van der Paardt, van Bruggen, Nusselder, Jalink and Peen14].
Treatment for schizophrenia is challenging, and non-response and non-remission rates for pharmacotherapy in schizophrenia can be very high [Reference Samara, Nikolakopoulou, Salanti and Leucht15]. Apart from pharmacological therapy, which is the first-line treatment for schizophrenia, treatment should include psychosocial interventions along with various modalities of psychotherapy, such as cognitive-behavioural therapy (CBT) and family therapy [16, 17] applied during different stages of the illness to achieve the functional recovery of patients. Finally, schizophrenia itself, as well as the use of antipsychotic medication, is associated with a multitude of somatic conditions, including obesity, hypertension, hyperlipidaemia, and metabolic syndrome, all of which can cause cardiovascular and cerebrovascular disease, reduce life expectancy [Reference Correll, Solmi, Veronese, Bortolato, Rosson and Santonastaso18], and also contribute to the complexity of treatment [19].
There are several existing clinical guidelines to support clinicians’ treatment decisions in the management of schizophrenia, developed based on a thorough evaluation of existing scientific literature, with clinical recommendations graded according to the level of evidence-based scientific data [16, 17]. As the process of guideline development usually takes years and, therefore, often may not include the most recent advances, specific clinical topics may best be covered by additional clinical guidance documents [Reference Falkai, Wagner, John, Yakimov, Galderisi and Bitter20]. In everyday clinical practice, however, there are many situations where guidelines/guidance are not fully used, despite numerous efforts to promote their adoption. This contributes to a significant divergence between recommended clinical practice and that actually followed in the “real world” [Reference Grol and Grimshaw21]. The reasons may include guideline complexity, a lack of unambiguous and clinically practical recommendations, and clinicians’ beliefs regarding evidence-based practice, among others [Reference McIntyre22, Reference Forsner, Hansson, Brommels, Wistedt and Forsell23]. Non-adherence to clinical guidelines may further increase the considerable variability in treatment approaches for the same psychiatric condition across Europe, as we previously identified in the case of post-traumatic stress disorder (PTSD) [Reference Rojnic Kuzman, Padberg, Amann, Schouler-Ocak, Bajic and Melartin24].
Therefore, we conducted a survey among psychiatrists trying to capture clinical decision-making for schizophrenia treatment in the “real world.” By using clinical case vignettes that depicted various phases and clinical scenarios of schizophrenia, we aimed to gather treatment preferences in managing “real-world” patients with schizophrenia from practicing clinicians in Europe and to compare their treatment choices with recommendations from selected clinical guidelines/guidance.
Methods
Study design, setting, and participants
This study followed the design of previous European Psychiatric Association (EPA) Ambassador Programme studies [Reference Rojnic Kuzman, Padberg, Amann, Schouler-Ocak, Bajic and Melartin24, Reference Rojnic Kuzman, Slade, Puschner, Scanferla, Bajic and Courtet25]. We approached psychiatrists working in Europe, who were associated with the EPA community, including individual members of the EPA and its member associations and attendees of the EPA congresses. In 2020, they were offered the opportunity to become “EPA Ambassadors” and to participate in EPA surveys, and the same recruitment process was repeated for each Ambassador study. Initially, we launched the survey during the EPA Forum of the EPA Congress 2023, where the invitation to participate was open to EPA Congress attendees. In the second wave, all representatives from the EPA Council of National Psychiatric Associations (Council of NPAs including 44 NPAs) were asked to recruit members of their NPAs, and in case of non-response from the NPAs representatives, a third wave of recruitment included members of the Board and representatives of the EPA Sections who were then asked to distribute the invitations to their members. Responses were collected from April to December 2023, using an online questionnaire, which took about 20–25 min to complete. The study was open to psychiatrists and psychiatry specialist trainees working in Europe. The authors declare that all procedures contributing to this work complied with the ethical standards of the relevant national and institutional committees for human experimentation and with the Helsinki Declaration of 1975, as revised in 2008 and 2013 [26]. The study was approved by the Ethical Committee of the Zagreb University Hospital Centre (number 02/013AG).
Case vignettes
Treatment attitudes were assessed using vignettes describing six cases of schizophrenia (the prodromal phase, first-episode psychosis, cognitive and negative symptoms, pregnancy, treatment resistance, and antipsychotic-induced metabolic side effects). Case vignettes were used as a tool to examine clinical judgments by health professionals, which can be highly generalizable to “real-life” behaviour [Reference Evans, Roberts, Keeley, Blossom, Amaro and Garcia27]. The cases were constructed using experimental design and were systematically manipulated across vignettes to assess their effect on the dependent variables while maintaining response consistency through standardized response options for all participants to ensure that the data gathered are interpretable in a consistent manner. First, one expert designed 10 cases following a set of recommendations (short cases derived from clinical experience, balanced in age and sex, and neutral with respect to cultural and socio-economic factors, following a narrative story-like progression, with a similar format for all cases and highlighting key variables of interest) [Reference Falkai, Wagner, John, Yakimov, Galderisi and Bitter20]. The initial cases were then reviewed and revised by three independent experts in the field of schizophrenia, including Chairs of the respective EPA Sections (Section of Schizophrenia, Section of Prevention and Early Intervention, and Section of Psychopharmacology) who read each vignette, provided a diagnosis, and confirmed the presence of key diagnostic features to improve the vignettes’ clarity, cultural neutrality, and validity. In the next step, the draft questionnaire containing six selected and revised case vignettes was submitted to a pilot study among the remaining 17 members of the EPA Board, including representatives of GAMIAN [Reference Grol and Grimshaw21] and EUFAMI [Reference McIntyre22] to address issues of internal and external validity. They were asked to respond to the questions, which included multiple choices for diagnoses, assessment tools, and standardized treatment options across all vignettes, and to give separate feedback on the difficulty, clarity, and length of the survey. The final questionnaire was reviewed and revised according to the pilot study analysis. All versions, as well as the final questionnaire, were developed and administered in the English language. The vignettes are described in Supplementary Table 1. Socio-demographic data (age, gender, and country of work), the nature of clinicians’ expertise, training, and practice (time since qualifying as consultant psychiatrists, subspecialty, work position, type of practice, clinical setting, and the use of clinical guidelines in everyday clinical practice were recorded.
Study outcomes
The primary outcome of the study was the level of consensus between clinicians’ treatment preferences in the six clinical vignettes. The secondary study outcome was the concordance of clinicians’ treatment preferences with recommendations from selected clinical guidelines/guidance.
Statistical analysis
We used descriptive statistics. Analyses were performed using StataCorp. 2019. Following the analysis of the primary outcome, in assessing the level of consensus among clinicians with respect to specific treatment preferences, we categorized responses as “low consensus” with <25% of participants choosing a specific treatment option, “partial consensus” with 25–50% of participants choosing a specific treatment option, “moderate consensus” with 50–75% of participants choosing a specific treatment option, and “high consensus” with more than 75% of participants choosing a specific treatment option (Stata Statistical Software: Release 16. College Station, TX: StataCorp LLC). The manuscript was written according to the STROBE guidelines for reporting cross-sectional studies [Reference Vandenbroucke, von Elm, Altman, Gøtzsche, Mulrow and Pocock28].
Selection of guidelines/guidance
We conducted a review of key guidelines regarding general treatment for schizophrenia, including those seen as long-standing and considered to have the most impact on the clinical practice of psychiatrists from Europe [16, 17, Reference Keepers, Fochtmann, Anzia, Benjamin, Lyness and Mojtabai29] European guidance papers for the treatment of schizophrenia focus on specific phases/clinical scenarios in schizophrenia, such as the EPA Guidance for Pharmacological Treatment of Schizophrenia, supported by all 44 NPAs [Reference Falkai, Wagner, John, Yakimov, Galderisi and Bitter20]; EPA Guidance on Treatment of Cognitive Impairment in Schizophrenia [Reference Vita, Gaebel, Mucci, Sachs, Barlati and Giordano30]; EPA Guidance on Treatment of Negative Symptoms in Schizophrenia [Reference Galderisi, Kaiser, Bitter, Nordentoft, Mucci and Sabé31]; the EPA Position Statement on Cardiovascular Disease and Diabetes in People with Severe Mental Illness – supported by the European Association for the Study of Diabetes and the European Society of Cardiology [Reference De Hert, Dekker, Wood, Kahl, Holt and H-J32]; EPA Guidance on Physical Activity as a Treatment for Severe Mental Illness: a Meta-review of the Evidence and Position Statement from the EPA – supported by the International Organization of Physical Therapists in Mental Health (IOPTMH) [Reference Stubbs, Vancampfort, Hallgren, Firth, Veronese and Solmi33]; the EPA Guidance on Lifestyle Interventions for Adults with Severe Mental Illness: A Meta-review of the Evidence [Reference Maurus, Wagner, Spaeth, Vogel, Muenz and Seitz34], and the EPA Guidance on Early Intervention in Clinical High-Risk States for Psychoses [Reference Schmidt, Schultze-Lutter, Schimmelmann, Maric, Salokangas and Riecher-Rössler35]. In addition, we included guidelines by the British Association for Psychopharmacology, as one of the few dealing specifically with the topic of pregnancy in patients with schizophrenia and at ultra-high-risk for schizophrenia [Reference Barnes, Drake, Paton, Cooper, Deakin and Ferrier36, Reference McAllister-Williams, Baldwin, Cantwell, Easter, Gilvarry and Glover37]. Following our secondary outcome in assessing concordance of clinicians’ treatment preferences with the guideline recommendations, we divided the responses into the following: (i) “not concordant,” with <25% of participants following a specific recommendation; (ii) “partial concordance,” with 25–50% of participants following a specific recommendation; (iii) “moderate concordance,” with 50–75% of participants following a specific recommendation; and (iv) “full concordance,” with more than 75% of participants following a specific recommendation.
Results
Participant characteristics
The online survey was completed by 855 mental health professionals worldwide. However, as this analysis focused on psychiatrists and psychiatry trainees from Europe only, the final sample consisted of 774 participants, including 637 psychiatrists and 137 psychiatry specialist trainees working in 39 European countries. The regional distribution of the final sample notably deviated from the regional distribution of the target population of European psychiatrists, and therefore, we did not perform the regional analysis (Supplementary Table 2). There were 320 missing data points (missing answers) for the clinical vignettes. As imputation would not have been valid, the missing data were excluded, and further analysis was conducted on the 454 complete responses.
The sample comprised 454 participants, nearly evenly split between males (N = 223, 49.1%) and females (N = 231, 50.9%). Most participants were psychiatrists (N = 392, 86.3%), while the remaining 62 (13.7%) were trainees. Nearly half held a PhD (N = 196, 43.2%). About half worked in university hospitals (N = 218, 48.0%). Participants had a median of 14 years of professional experience (interquartile interval [iqi] = 7–25) and a median of 13 years (iqi = 6–23) of experience specifically in schizophrenia care. A significant portion (N = 200, 44.1%) worked in early intervention services (specialized mental health programmes designed to provide timely and comprehensive care to individuals experiencing their first episode of psychosis or those at high risk), for a median of 4 years (iqi = 2–9).
More than a third were certified psychotherapists (N = 160, 35.2%), with the majority specializing in CBT (N = 83, 51.9%), psychodynamic psychotherapy (N = 48, 30.0%), or other psychotherapy modalities (N = 36, 22.5%). Participants represented all four European regions: Central and Eastern Europe (N = 262), Southern Europe (N = 58), Northern Europe (N = 46), and Western Europe (N = 88) (Table 1 and Supplementary Table 2).
Table 1. Description of participants, raw, unweighted data, and total sample (n = 454)
Abbreviations: iqi, interquartile interval; SD, standard deviation.
Note: Data are presented as number (percentage) of participants if not stated otherwise.
The majority of participants frequently or consistently followed national guidelines (N = 333, 73.3%). Nearly half adhered to the National Institute for Health and Care Excellence (NICE) guidelines (N = 204, 44.9%), followed by the EPA guidance papers (N = 164, 36.1%) and the American Psychiatric Association (APA) guidelines (N = 125, 27.5%). Other guidelines were used by 91 participants (20%) (Supplementary Table 3).
Treatment preferences
In Case A describing a patient with an acute first episode of schizophrenia, more than 80% of participants correctly identified the diagnosis as a psychotic disorder, including schizophrenia (N = 270, 59.5%) and transient and acute psychotic episode (N = 99, 21.8%), with catatonia and other diagnoses comprising 12.6% (N = 57) and 6.2% (N = 28), respectively. The majority would use scales for psychosis (N = 264, 58.2%), scales for mood and anxiety (N = 159, 35%), neurocognitive assessment (N = 134, 29.5%), and diagnostic interview (N = 152, 33.5%) (Supplementary Table 4).
In Case B describing a patient with attenuated psychosis syndrome, the majority of participants did not identify the diagnosis in this patient, choosing the response “I don’t know” (N = 146, 32.2%, followed by “other” (N = 91, 20%) and “attenuated psychosis syndrome” (N = 69, 15.2%). The majority of participants would apply the scales for mood and anxiety (N = 250, 55.7%), followed by scales for psychosis (N = 138, 30.4%) and neurocognitive assessment (N = 105, 15.6%). Additionally, 71 participants (15.6%) would use risk assessment scales, that is, double the proportion compared to Case A (N = 36, 7.9%; Supplementary Table 4).
The use of pharmacotherapy, psychotherapy, psychosocial interventions and other inteventions for all cases are shown in Table 2. The level of consensus between clinicians’ treatment preferences in the six clinical vignettes and their concordance with recommendations from selected clinical guidelines/ guidance are shown in Table 3. Additionally, doses of antipsychotics were analysed in two cases, Case B and case E/E1 because of specific recommendations in guidelines/ guidances. In case B, low dose of new antipsychotic would prescribe 21/33 (63%) for amisuprid, 11/141 (7%) for aripiprazole, 9/48 (18%) for cariprazine, 9/23 (39%) for lurasidone, 18/80 (23%) for olanzapine, 3/15 (20%) for paliperidone, 41/98 (42%) for quetiapine, 8/56 (14%) for risperidone 2/9 (22%) for ziprasidone, and 5/11 (45%) for sulpride. Low dose was defined as the lower than standard doses, as follows: amisulprid and sulpiride 200mg, aripiprazole 5 mg, cariprazine 1,5mg, lurasidone 40mg, olanzapine 5mg, paliperidone 3mg, risperidone 1mg, ziprasidone 60mg, quetiapine 150mg. In cases E and E1, most would consider clozapine in doses up to 350mg and 400mg, respectively (N=127/162, 78% for case E and 168/211, 80% for case E1). Screening and monitoring for metabolic side effects initially and within 3 months of initiating new treatment is shown in Supplementary Table 5.
Table 2. Treatment preferences for the six described cases
Abbreviations: LAI - long-acting injectable.
Table 3. Summary of reached consensus among clinicians and their concordance with guidelines/guidance papers
Note: The level of consensus among clinicians is categorized as “low consensus” with <25% of participants choosing a specific treatment option (red), “partial consensuses with 25–50% of participants choosing a specific treatment option (orange), “moderate consensuses” with 50–75% of participants choosing a specific treatment option (yellow), and “high consensuses” with more the 75% of participants choosing a specific treatment option (green). In assessing the concordance of clinicians’ choices with the guidelines’ recommendation, we divided the response into “not concordant” with <25% or participants following a specific recommendation, “partial concordance” with 25–50% of participants following a specific recommendation, “moderate concordance” with 50–75% of participants following a specific recommendation, and “good concordance” with more the 75% of participants following a specific recommendation. Grey boxes with no text indicate “no specific recommendation” in the guidelines/guidance papers.
* Low dose was defined as the lower than standard doses, as follows: amisulprid and sulpiride 200mg, aripiprazole 5 mg, cariprazine 1,5mg, lurasidone 40mg, olanzapine 5mg, paliperidone 3mg, risperidone 1mg, ziprasidone 60mg, quetiapine 150mg.
Discussion
Pharmacotherapy
Case A: First episode of schizophrenia
With more than 90% of participants recommending antipsychotics and choosing one of the three second-generation antipsychotics, results indicate high consistency in preferences for antipsychotics during the acute phase of schizophrenia. While guidelines, in general, do not prefer one antipsychotic over another but suggest consideration of various factors when choosing antipsychotics for first-episode psychosis, these results are partially comparable with the data from Eastern [Reference Szkultecka-Dębek, Miernik, Stelmachowski, Jakovljević, Jukić and Aadamsoo40] and Northern Europe [Reference Hamina, Taipale, Lieslehto, Lähteenvuo, Tanskanen and Mittendorfer-Rutz41, Reference Taipale, Puranen, Mittendorfer-Rutz, Tiihonen, Tanskanen and Cervenka42], where olanzapine was the most commonly prescribed, followed by clozapine and risperidone, but also with the data from the United Kingdom and the United States, where aripiprazole and quetiapine were the most prescribed drugs (including for bipolar disorders) [Reference Richards-Belle, Launders, Hardoon, KKC, Bramon and DPJ43, 44]. The majority would consider using more than one antipsychotic, which is not in line with any of the guidelines [Reference Højlund, Köhler-Forsberg, Gregersen, Rohde, Mellentin and Anhøj45].
About a third of participants would prescribe benzodiazepines, primarily lorazepam. While the use of lorazepam may be indicated with the clinical presentation of catatonia symptoms [Reference Rogers, Oldham, Fricchione, Northoff, Ellen Wilson and Mann46] identified by 13% of participants, the use of benzodiazepines is still high, especially considering the lack of evidence for their efficacy in treating psychotic symptoms and existing evidence of side effects [Reference Zaman, Sampson, Beck, Sharma, Clay and Spyridi47]. The high use of benzodiazepines is consistent with studies from Eastern Europe, indicating high numbers of patients with schizophrenia treated with benzodiazepines for lengthy periods of time [Reference Maric, Andric Petrovic, Russo, Jerotic, Ristic and Savić48]. However, the high use of benzodiazepines seems to be a widespread clinical practice, unrelated to a specific disorder, as was also previously identified as a treatment preference for PTSD in Central European and Southern European countries [Reference Rojnic Kuzman, Padberg, Amann, Schouler-Ocak, Bajic and Melartin24]. While some studies have suggested that benzodiazepine use is more widespread in Western countries [Reference Bushnell, Stürmer, Gaynes, Pate and Miller49, 50], there may be a lack of data from Central European countries, as pointed out by some authors. [Reference Winkler, Krupchanka, Roberts, Kondratova, Machů and Höschl51]
Almost two-thirds would consider therapeutic drug monitoring (TDM), even when there is no indication of resistance to treatment, non-compliance, or adverse reactions, which are the major indications for TDM according to guidelines [Reference Falkai, Wagner, John, Yakimov, Galderisi and Bitter20, Reference Hiemke, Bergemann, Clement, Conca, Deckert and Domschke39] Interestingly, about half would consider long-acting injectables (LAIs), which is in line with the EPA guidance on the pharmacological treatment of schizophrenia [Reference Falkai, Wagner, John, Yakimov, Galderisi and Bitter20] and recent clinical consensus [Reference Arango, Fagiolini, Gorwood, Kane, Diaz-Mendoza and Sahota52]. This may reflect a trend towards increasing prescription of LAIs, as they were found to be more efficacious in relapse prevention compared to oral formulations [Reference Barnes, Shingleton-Smith and Paton53, Reference Arango, Baeza, Bernardo, Cañas, de Dios and Díaz-Marsá54]. However, this percentage seems rather low compared to the United States [Reference Bunting, Chalmers, Yohanna and Lee55] where up to 60% of patients with first-episode psychosis are prescribed LAIs, but with considerable variation – in community-based settings with specialized programmes, the use of LAIs is higher compared to hospital-based care.
Case B: Prodromal stage/high-risk states
The relatively low number of participants correctly identifying “attenuated psychosis syndrome” may reflect the combination of two interrelated issues: (1) the currently undefined diagnostic status of prodromal schizophrenia in major classification systems, and (2) the broader challenge of relatively modest diagnostic precision when assessing psychosis-risk syndromes in real-world clinical vignettes, even by international leading experts (as evidenced by recent research [Reference Urkin, Parnas, Raballo and Koren56]).
Concordantly with the low number of participants who correctly identified the “attenuated psychosis syndrome,” only about 15% of participants would use risk assessment scales, reflecting that take-up of appropriate assessment tools was far lower than the EPA recommendation, which states that any psychosis-preventive intervention requires full assessment of the clinical high risk (CHR) status in accordance with the EPA guidance on early detection of psychosis [Reference Schmidt, Schultze-Lutter, Schimmelmann, Maric, Salokangas and Riecher-Rössler35].
Interestingly, disregarding confusion about the concept or diagnostic scales, the use of antipsychotics was, in general, in line with the guidelines, as about a third stated they would not use any antipsychotics or did not mark any, and those who prescribed would use second-generation antipsychotics. However, the “low-dose formulation” may be particularly difficult to grasp in clinical practice, especially for some of the medications. The use of other medications, especially antidepressants, was in line with the guidelines.
Case C: Negative and cognitive symptoms
Overall, these results follow the recommendation of the EPA guidance on the treatment of negative symptoms to use antipsychotics with antidepressant properties and add-on antidepressant treatment. To avoid polypharmacy, however, the add-on antidepressant should be discontinued if it does not lead to improvement in negative symptoms and/or depression [Reference Galderisi, Kaiser, Bitter, Nordentoft, Mucci and Sabé31]. In the case of specific antipsychotics, most participants would choose aripiprazole, cariprazine, and, more rarely, risperidone. This can be considered in line with the EPA guidance on the treatment of negative symptoms in schizophrenia, which suggests the use of amisulpride and cariprazine for negative symptoms, but with the notion that further research is necessary before a specific recommendation can be provided [Reference Galderisi, Kaiser, Bitter, Nordentoft, Mucci and Sabé31]. This is also in line with the head-to-head comparison of antipsychotics, where second-generation antipsychotics were associated with fewer negative symptoms than first-generation antipsychotics [Reference Leucht, Corves, Arbter, Engel, Li and Davis57], although both haloperidol and risperidone can induce negative symptoms in healthy individuals [Reference Artaloytia, Arango, Lahti, Sanz, Pascual and Cubero58].
There was very low consensus among participants in the choice of specific add-on antidepressants, and concordantly, there is no specific recommendation on the use of antidepressants in the EPA guidance on the treatment of negative symptoms in schizophrenia [Reference Galderisi, Kaiser, Bitter, Nordentoft, Mucci and Sabé31], nor in SIGN recommendations [16]. Therefore, considering the multitude of antidepressants, a call for more specific add-on strategy recommendations may be appropriate. Possible beneficial effects were found by augmentation with antidepressants [Reference Galling, Roldán, Hagi, Rietschel, Walyzada and Zheng59] where 12 combinations outperformed monotherapy, with Serotonine and Norepinephrine Reuptake Inhibitors (SNRIs) associated with the largest effect size. In a report encompassing 82 add-on studies, Selective Serotonin Reuptake Inhibitors (SSRIs) and tetracyclic antidepressants were more efficacious than control, but individually, selegiline, duloxetine, citalopram, fluvoxamine, and mirtazapine appeared significantly more efficacious than monotherapy for negative symptoms [Reference Falkai, Wagner, John, Yakimov, Galderisi and Bitter20, Reference Correll, Rubio, Inczedy-Farkas, Birnbaum, Kane and Leucht60, Reference Helfer, Samara, Huhn, Klupp, Leucht and Zhu61], although not for the primary negative symptoms [Reference Arango, Garibaldi and Marder62, Reference Galderisi, Mucci, Buchanan and Arango63].
Contrary to guidelines and the results of our study, in the European guidance papers [Reference Falkai, Wagner, John, Yakimov, Galderisi and Bitter20], concordance with the two similar recommendations on the treatment for negative symptoms, “we suggest offering amisulpride (at a low dose) or olanzapine; we suggest avoiding the use of strong D2 receptor blockers by using antipsychotics with a suitable profile or avoiding high-dose treatments,” and “In case of inadequate response to antipsychotic monotherapy, we suggest offering additional treatment with antidepressants to people with schizophrenia and predominant negative symptoms,” reached only 30 and 52.2%, respectively, further indicating variable clinical practice.
Finally, these results are in line with the existing EPA guidance on treatment of cognitive impairment, where it is stated that benzodiazepines have a clearly negative effect on cognition and should be avoided [Reference Vita, Gaebel, Mucci, Sachs, Barlati and Giordano30, Reference Fond, Berna, Boyer, Godin, Brunel and Andrianarisoa64].
Case D: Pregnancy
Our results indicate that clinical practice differs from the international guidelines, as more than a third stated they would not prescribe any antipsychotics. This is contrary to guideline recommendations against stopping antipsychotic medication and stating that antipsychotics do not pose significant risks for pregnancy, supported by studies where patients with treated mental illness are compared to patients with untreated mental illness [Reference Betcher, Montiel and Clark65]. On the other hand, the majority of participants who would continue antipsychotic medication recommend the continuation of the current antipsychotic (in this case, aripiprazole), which proved effective and is in line with the guidelines. Concordant with our results, in the United States, aripiprazole is one of the most frequently prescribed antipsychotics in pregnancy, while in older studies, haloperidol, olanzapine, and risperidone were the most prescribed [Reference Park, Huybrechts, Cohen, Bateman, Desai and Patorno66].
About a third would consider TDM, as indicated by the Consensus Guidelines for Therapeutic Drug Monitoring in Neuropsychopharmacology [Reference Hiemke, Bergemann, Clement, Conca, Deckert and Domschke39], indicating partial concordance with the guideline’s recommendations.
Case E: Treatment resistance
There is considerable variation in definitions of treatment resistance in both research and clinical settings. According to APA guidelines [Reference Keepers, Fochtmann, Anzia, Benjamin, Lyness and Mojtabai29], in terms of treatment adequacy, at least two trials with two different antipsychotics for 6 weeks and adherence of at least 80%, with at least one antipsychotic blood level to assess adherence, are required, including evaluating adherence from at least two sources and obtaining information on prior treatment as well. In clinical (not research) settings, at least two trials of 6 weeks each in adequate clinical doses are required. However, if there is no significant improvement after a few weeks of treatment (e.g., 20% reduction in symptoms), the likelihood of adequate treatment response is minimal, and, often in clinical practice, a faster switch of antipsychotics may occur [Reference Howes, Vergunst, Gee, McGuire, Kapur and Taylor67, Reference Agid, Arenovich, Sajeev, Zipursky, Kapur and Foussias68]. Concordantly, in our study, the median (iqi) switch would happen after 4 (3–6) weeks. Interestingly, one of the statements in the European guidance of pharmacological treatment of schizophrenia [Reference Falkai, Wagner, John, Yakimov, Galderisi and Bitter20] that did not reach clinical consensus at the level of 75% (agreement was 67.5%) was “We recommend evaluating the response status after two weeks (at the latest after four weeks) by using a suitable scale (ideal: PANSS, BPRS; easier: CGI). In case of lack of response (CGI unchanged or worse [CGI < 3]) despite adequate dosing and after excluding secondary causes, we recommend offering the patient a switch to an antipsychotic with a different receptor binding profile, with the aim to achieve response,” indicating variability when it comes to the clinical definition of resistance to treatment.
Guidelines consistently stress the importance of excluding pseudo-resistance and non-adherence. In line with this recommendation, most would probably or certainly consider TDM and screening for substance drug use, and the use of LAIs in the majority of cases.
While the use of clozapine is indicated in treatment-resistant schizophrenia in several guidelines, fewer than half of participants would recommend clozapine, and more often would try a combination of two or even three further antipsychotics. It was noted that barriers to the use of clozapine include the overestimation of its side effects and a lack of knowledge and experience in prescribing [Reference Howes, Vergunst, Gee, McGuire, Kapur and Taylor67, Reference Farooq, Choudry, Cohen, Naeem and Ayub69], although its use in first episode psychosis has been suggested even before two 4-week trials have failed, due to significant improvement and lower rates of re-admittance with clozapine compared with other antipsychotics [Reference Tiihonen, Haukka, Taylor, Haddad, Patel and Korhonen70].
Augmentation with other medications was recommended by half the participants, but distributed evenly between medications, probably indicating a lack of (clinical) evidence for effective augmentation strategies.
Case F: Metabolic syndrome
More than 75% of participants would choose a strategy to replace the antipsychotic associated with metabolic changes with one with a lower propensity to gain weight and metabolic disturbances [Reference Pillinger, McCutcheon, Vano, Mizuno, Arumuham and Hindley71]. This is not the main strategy recommended in the guidelines. While the use of metformin is recommended only for very-high-risk individuals (Case F would qualify for this recommendation), only 16% of participants would recommend metformin, possibly indicating an unwillingness or lack of expertise among psychiatrists to recommend non-psychiatric medication. Indeed, in the European guidance on the pharmacological treatment of schizophrenia [Reference Falkai, Wagner, John, Yakimov, Galderisi and Bitter20], the statement “If there is strong weight gain and it is necessary to continue the current antipsychotic medication, after performing the specified psychotherapeutic and psychosocial interventions, we recommend offering treatment with metformin (first choice) or topiramate (second choice) for weight reduction, taking into account the risks of an additional drug treatment” did not reach the 75% of consensus, despite the good quality scientific evidence.
The use of ECT and TMS
The use of Electroconvulsive Therapy (ECT) was recommended by <25% in the case of treatment resistance, which is not according to the guidelines [16, 38]. The relatively high percentage of participants (20%) who would probably or certainly consider ECT in Case A (first-episode psychosis), compared to other cases can be at least partially explained by the clinical presentation of the first case (catatonic symptoms). Transcranial Magnetic Stimulation (TMS) is considered infrequently (up to around 10%, most commonly in the treatment of prodromal, negative, and cognitive symptoms). This possibly indicates a lack of availability but also a lack of evidence in the use of TMS on negative symptoms [Reference Lorentzen, Nguyen, McGirr, Hieronymus and Østergaard72], which is also reflected by the lack of clinical consensus in the EPA guidance of pharmacological treatment of schizophrenia on the use of rTMS in the treatment of negative symptoms and use of ECT or rTMS as an augmentation treatment for improving treatment-resistant symptoms [Reference Falkai, Wagner, John, Yakimov, Galderisi and Bitter20].
Psychotherapy, sociotherapy, and recovery-oriented care
Despite psychotherapy (predominantly CBT and family therapy) being consistently recommended in the guidelines during all phases of schizophrenia, it was recommended relatively infrequently in this study. It can be considered at best as only partially in line with the guidelines. Other psychotherapies were recommended even less frequently, making psychological interventions an underused treatment modality.
Psychoeducation and lifestyle recommendations (diet, sleep, smoking, alcohol, drugs, and physical activities) were most often suggested, reaching about 50–60% depending on the case. However, a structured physical health training programme (exercise), which is strongly recommended in the guidelines for the management of cognitive and negative symptoms, as well as being a core part of preventing and managing poor physical health in people with Severe Mental Illness (SMI) [Reference Vita, Gaebel, Mucci, Sachs, Barlati and Giordano30, Reference Stubbs, Vancampfort, Hallgren, Firth, Veronese and Solmi33, Reference Maurus, Wagner, Spaeth, Vogel, Muenz and Seitz34] was endorsed in only about 30% of cases involving antipsychotic-induced metabolic side effects and in 24% of cases describing cognitive and negative symptoms.
Creative therapies were suggested in 20% of cases. These therapies are generally supported by the guidelines, but quite non-specifically or not universally. For example, while the use of art therapy in the treatment of negative symptoms is supported by NICE guidelines [17], it is not supported in other guidelines, leaving the evidence on its efficacy as inconclusive [Reference Laws and Conway73].
Cognitive remediation, which is largely supported for the treatment of cognitive symptoms, was recommended by fewer than 15%, possibly indicating a relative unavailability or lack of experience [Reference Keepers, Fochtmann, Anzia, Benjamin, Lyness and Mojtabai29, Reference Vita, Gaebel, Mucci, Sachs, Barlati and Giordano30].
Support from psychiatric services
The low use of non-pharmacological treatment in all cases cannot be considered in line with guidelines, as these consistently recommend psychotherapy and interventions aimed at recovery within services for persons with schizophrenia during all phases of the illness [17, Reference Keepers, Fochtmann, Anzia, Benjamin, Lyness and Mojtabai29].
Considering that only up to 50% would recommend the use of recovery-oriented care, it seems that recovery-oriented care is not supported within the organizational framework of psychiatric services across Europe. A community mental health model providing these services should be fostered across Europe. At the moment, community-based services are still developing in many European countries, while hospital-based services remain the predominant model of care [Reference Winkler, Krupchanka, Roberts, Kondratova, Machů and Höschl51, Reference Semrau, Barley, Law and Thornicroft74].
A relative lack of psychotherapists across Europe, possibly indicating the need to improve the provision of psychotherapy training as part of psychiatric training across Europe [Reference Fiorillo, Luciano, Giacco, Del Vecchio, Baldass and De Vriendt75], may be reflected in only up to 25% of participants recommending psychotherapy, except in the case of describing a person in prodrome. Alternatively, it is possible that psychotherapy is more often offered to younger patients within specialized services, such as early intervention services. Indeed, guidelines specifically recommend the use of other sociotherapy and recovery-oriented care as part of early intervention services for first-episode psychosis and persons experiencing prodromal symptoms [17]. These are offered across Europe and worldwide [Reference Maric, Raballo, Rojnic Kuzman, Andric Petrovic, Klosterkötter and Riecher-Rössler76, Reference Kotlicka-Antczak, Podgórski, Oliver, Maric, Valmaggia and Fusar-Poli77] and include a wide and variable range of services, from psychopharmacology to psychosocial interventions. However, it seems that the pathway to care for people at risk is not straightforward, even in countries where these services exist. Persons at clinically high risk may be treated within other services (e.g., in services for children and adolescents, youth services, general practitioners, or general psychiatric services) [Reference Kotlicka-Antczak, Podgórski, Oliver, Maric, Valmaggia and Fusar-Poli77, 78], suggesting that better provision of intervention services should be encouraged across the European Union countries.
Finally, during pregnancy, referral to perinatal services is suggested [79]. While some European countries have established perinatal psychiatry services, there remains a significant need for the development and implementation of comprehensive perinatal mental health care across the continent [Reference Horakova, Nemcova, Hrdlickova, Kalli, Davletova and Duarte80].
Limitations of the study
The study had several limitations, which have been described elsewhere [Reference Rojnic Kuzman, Padberg, Amann, Schouler-Ocak, Bajic and Melartin24]. Briefly, we cannot claim that the sample is fully representative at the national level because of low response rates in individual countries and a possible association of non-response with specific preferred treatment approaches. The overall number of participants comprised a small proportion of all psychiatrists, and the analysis was done at the whole sample level, not taking potential regional differences into account. Second, considering that about half of the participants worked in university hospitals and have a PhD, the sample could be biased towards clinicians focused on research and more likely to follow guidelines (that they produce) compared to clinicians working outside university centres. Third, in the sample, there was a high proportion of psychotherapists trained in CBT, which could have biased some of the responses and may reflect the variability in availability and training across Europe [Reference Fiorillo, Luciano, Giacco, Del Vecchio, Baldass and De Vriendt75]. Fourth, this study used a non-standardized tool to measure treatment attitudes based on clinical cases. Although case vignettes are frequently used to assess clinical judgments made by health professionals on “real-world” phenomena, and were developed following a set of recommendations to ensure both internal and external validity [Reference Falkai, Wagner, John, Yakimov, Galderisi and Bitter20], the responses were standardized for treatment options only, and not for assessing diagnostic precision and differential diagnostic competencies. Additionally, case vignettes were administered in English, which may have reduced clarity due to potential language barriers.
Furthermore, the selection of analysed guidelines assumed that European guidance papers, along with selected international guidelines (e.g., APA guidelines), have a greater influence on psychiatric practice in Europe, which was confirmed by the results summarized in Supplementary Table 3. Second, we assumed that national guidelines, where available, are largely aligned with these broader guidelines. Finally, the selection was also limited by language constraints, as only English-language guidelines were considered. However, it is interesting to note that 5–20% of participants reported that they do not use guidelines in clinical practice, although this percentage was much lower compared to the results obtained from the Ambassador study on PTSD, where non-use of guidance reached about 50% [Reference Rojnic Kuzman, Padberg, Amann, Schouler-Ocak, Bajic and Melartin24]. Possibly, this correctly reflects practice in the field of schizophrenia compared to other fields.
Conclusions
We analysed treatment preferences across six different clinical scenarios and situations in schizophrenia in a sample of European clinicians and compared these preferences with relevant clinical guidelines/guidance papers. Our findings demonstrated the following:
-
• A moderate to high level of concordance among European psychiatrists and high to moderate compliance with guidelines in the following cases:
-
○ Pharmacological treatment of first-episode psychosis – The most frequently selected antipsychotics were risperidone, olanzapine, aripiprazole, and quetiapine, with consideration of LAIs. However, monotherapy was used less frequently than expected, and this was not in accordance with the guidelines, while the use of sedatives was higher than anticipated.
-
○ Prodromal phase – Most clinicians opted for antidepressant treatment, and about one-third prescribed antipsychotics, but predominantly at standard rather than low doses. Most clinicians recommended CBT.
-
○ Cognitive and negative symptoms – The majority of clinicians selected newer antipsychotics (predominantly aripiprazole and cariprazine) and augmentation with antidepressants
-
○ Pregnancy – Most participants continued with the initial antipsychotic, but approximately one-third of participants discontinued antipsychotic treatment.
-
○ Treatment-resistant schizophrenia – Aligned with guidelines related to the exclusion of pseudo-resistance (TDM, screening for substance use, and LAIs).
-
-
• While treatment preferences among clinicians were similar, compliance with guidelines was poor in the following cases:
-
○ Antipsychotic-induced metabolic side effects – most participants discontinued the antipsychotic associated with weight gain and switched to aripiprazole, but only 16% selected metformin as recommended by guidance papers.
-
○ Treatment resistance – only 40% of participants chose clozapine, although there was high agreement on using clozapine in doses up to 400 mg daily, and most participants chose antipsychotic polypharmacy.
-
-
• Despite their strong representation in clinical guidelines, non-pharmacological treatments and recovery-oriented care were generally underutilized and thus reflected poor compliance with guidelines, except in the use of psychoeducation and lifestyle modification. Psychotherapy was also underused and non-compliant with guidelines, with CBT reaching ~25% and family therapy around 20%.
-
• Poor compliance was also found in using structured physical training, as it was endorsed in only about 30% of participants involving antipsychotic-induced metabolic side effects and in 24% of participants describing cognitive and negative symptoms. Cognitive remediation was recommended in only 13% of participants, even when treating cognitive symptoms, which contradicts guideline recommendations.
-
• Finally, complete initial and follow-up assessments (e.g., laboratory tests, electrocardiograms, and other monitoring after initiating antipsychotic therapy) were performed by fewer than half of the participants, highlighting the need to promote a more holistic approach in psychiatric practice.
Implications of the findings for future practice
The study has several important implications. First and foremost, it highlights a general convergence in treatment preferences across various phases of schizophrenia, although only partial alignment with existing clinical guidelines/guidance papers is observed. Notably, more than 70% of participants in this sample reported consulting national guidelines, while 30–45% referred to other international guidelines. Only about 11% of respondents rarely or never consulted guidelines. These findings suggest that the approach used in developing the EPA guidance on the pharmacological treatment of schizophrenia – which involves endorsing existing national guidelines and complementing them with clinical consensus – may be an effective strategy for improving guideline implementation and clinician adherence.
To further support this goal, the EPA should actively promote and propose educational initiatives that enhance the implementation of best practices at different levels. These initiatives should include: (1) Continuing professional development programmes for psychiatrists, with a focus on specific clinical stages of schizophrenia; (2) Integration of best practice standards into the European Board Examination in Psychiatry to reinforce evidence-based treatment approaches; and (3) Encouragement of real-world clinical trials and the incorporation of their findings into updated clinical guidelines.
Clinicians rely on a combination of clinical knowledge, accumulated experience, and research data summarized in guidelines, which they then apply and refine through their clinical practice. This positions them uniquely to contribute to the clinical consensus level of evidence, which should play a role in the development of future guidelines – particularly in areas where well-designed studies replicating real-world clinical practice are lacking but where extensive clinical experience exists. This need is particularly evident in cases where strong consensus among clinicians diverges from current guidelines. Such discrepancies should be carefully considered in the guideline revision process to ensure that recommendations align not only with research evidence but also with the realities of clinical practice.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1192/j.eurpsy.2025.10072.
Data availability statement
The data that support the findings of this study are available as open source.
Acknowledgements
We thank the EPA staff for the technical support in developing and distributing the survey.
Competing interests
Celso Arango has been a consultant to or has received honoraria or grants from Abbot, Acadia, Ambrosetti, Angelini, Biogen, BMS, Boehringer, Carnot, Gedeon Richter, Janssen Cilag, Lundbeck, Medscape, Menarini, Minerva, Otsuka, Pfizer, Roche, Sage, Servier, Shire, Schering Plow, Sumitomo Dainippon Pharma, Sunovion, Takeda, and Teva.
Open access
Comments
No Comments have been published for this article.