Main findings

In this nationwide study, medication prescription patterns were investigated in patients under compulsory psychiatric care; with approximately one out of three (33.2%) meeting the criteria for polypharmacy and 39.2% of them meeting the criterion for prescription high-dose antipsychotics (including PRN). This is the first study in the Republic of Cyprus exploring prescription patterns of antipsychotics.

Comparison with findings from other studies

The findings are in line with the literature reporting that although efforts have been made through guidelines to avoid prescription of high-dose antipsychotics and polypharmacy,^{Reference McGorry, Killackey and Lambert20,Reference Kreyenbuhl, Buchanan, Dickerson and Dixon21} this remains a common practice. Compulsory psychiatric treatment has been associated with both polypharmacy and prescription of high-dose antipsychotics in several studies.^{Reference Mendes, Azeredo-Lopes and Cardoso2,Reference McMillan, Jacobs, Wilson, Theodoros, Robinson and Anderson5,Reference Lelliott, Paton, Harrington, Konsolaki, Sensky and Okocha22}

The present study suggest that polypharmacy is because of the use of medications PRN; mainly haloperidol, almost universally prescribed to all patients. This may suggest that the use of PRN medication is mainly targeted at sedation and management of dysfunctional behaviours in in-patients (such as agitation, violent reactions) rather than having a therapeutic effect.^{Reference Mendes, Azeredo-Lopes and Cardoso2} Similar findings were observed in the study by Mendes et al,^{Reference Mendes, Azeredo-Lopes and Cardoso2} where approximately half of the participants were prescribed with ‘as required’ antipsychotics. Nevertheless, this issue deserves more attention in future research.

In modern psychiatry, striving for patient-centred care, the practice of ‘one size fits all’ seems to be established in the case of Cyprus, a pharmacotherapy pattern that needs to be further investigated. Nevertheless, the rate of polypharmacy without a PRN schema was 5.6%, which appears to be among the lowest reported in the literature. Polypharmacy rates range widely across different countries; for instance, a rate of around 17% has been reported in the UK and Australia,^{Reference Kadra, Stewart, Shetty, MacCabe, Chang and Kesserwani9,Reference Mace and Taylor23} 32% in Italy^{Reference Santone, Bellantuono, Rucci, Picardi, Preti and De Girolamo24} and 41.6–70.5% in Portugal.^{Reference Mendes, Azeredo-Lopes and Cardoso2,Reference Martinho, Poças, Simões and Laureano4}

Interpretation of our findings

A systematic review and meta-analysis of 147 studies on polypharmacy reported that the median rate of antipsychotic polypharmacy was 19.6%.^{Reference Gallego, Bonetti, Zhang, Kane and Correll3} Polypharmacy in individuals with mental health problems has been associated with diagnoses of severe mental disorders, including psychosis.^{Reference Correll and Gallego25} As expected, this diagnosis is common in patients involving involuntary hospital admissions, along with limited response to medication treatment and inadequate insight about disease;^{Reference Sim, Su, Fujii, Yang, Chong and Ungvari26} characteristics that were also present among participants in the present study. Indeed, polypharmacy was positively associated with the PANSS positive symptoms subscale score. The present findings also demonstrated that polypharmacy was associated with male gender, positive personal psychiatric history, receiving state benefits, positive history of involuntary hospital admissions and positive history of substance use. These factors have been previously associated with a worse prognosis in patients with psychosis.^{Reference El Abdellati, De Picker and Morrens27}

Moreover, the reported low rate of polypharmacy when PRN schema was not included (5.6%), may imply that PRN schemas are not used as indicated in published guidelines (i.e. as circumstances require),^{Reference Remington, Addington, Honer, Ismail, Raedler and Teehan1,Reference Taylor, Barnes and Young18} but rather as part of the formal pharmacotherapy plan. Additionally, such an extensive use of PRN schemas may suggest that these are not primarily used for their therapeutic effect but as a form of chemical restraint towards aggressiveness during hospital admissions; a practice previously reported.^{Reference Mendes, Azeredo-Lopes and Cardoso2} Nevertheless, there is no robust evidence that PRN use in patients with psychosis is effective or even beneficial; a recent review reported lack of randomised controlled trials on the subject, or any other evidence to support this practice.^{Reference Douglas-Hall and Whicher15} In fact, the review underlined caution on PRN use and highlighted patients’ consent on this, a status compromised in compulsory psychiatric care.

Furthermore, the present findings reported a rate of 27.2% of prescription of high-dose antipsychotics–PRN not included; that increased to 39.2% when PRN schemas were included. These rates are much higher compared with the East Asia region (6.5%)^{Reference Sim, Su, Fujii, Yang, Chong and Ungvari26} and other countries, such as the UK (6.8%),^{Reference Kadra, Stewart, Shetty, MacCabe, Chang and Kesserwani9,Reference Santone, Bellantuono, Rucci, Picardi, Preti and De Girolamo24} Australia (20.5%)^{Reference McMillan, Jacobs, Wilson, Theodoros, Robinson and Anderson5} and the Netherlands (25.5%).^{Reference Barbui, Nosè, Mazzi, Thornicroft, Schene and Becker28} Only two studies, one German and one Portuguese, have reported higher rates of prescription of high-dose antipsychotics (42.5%, 51.4%, respectively),^{Reference Martinho, Poças, Simões and Laureano4,Reference Barbui, Nosè, Mazzi, Thornicroft, Schene and Becker28} as the results of the present study show; and in another study in Portugal, prescription of high-dose antipsychotics was 13.8% when PRN was not included, and 49.2% when PRN was included.^{Reference Mendes, Azeredo-Lopes and Cardoso2} Overall, a high frequency of prescription of high-dose antipsychotics may also support the hypothesis that PRN schemas are mainly used as a means of chemical restraint, unfortunately leading to high total daily doses of antipsychotics, compromising quality and patient safety.

In this study, those unemployed and receiving state benefits, and with primary education only appeared to be more likely to be prescribed with high-dose antipsychotics, as previously reported.^{Reference Mendes, Azeredo-Lopes and Cardoso2} This may imply that this group of patients is either inadequately socially adjusted or supported, and/or is presenting with more severe symptomatology, probably as a result of low adherence to therapy. Data show that those with stable social support systems and higher education report increased medication adherence; thus, they are less prone to relapse and being prescribed with high-dose antipsychotics and involuntary hospital admissions for compulsory treatment.^{Reference Barkhof, Meijer, de Sonneville, Linszen and de Haan29} Moreover, one may argue that clinicians may rely on enhanced medication schemas in the absence of a functional social network, or effective community mental health services.

Individuals with a negative history of substance use had a lower risk being prescribed high-dose antipsychotics. This might be explained by the fact that, substance-induced psychotic symptoms most of the times decline rapidly; thus, antipsychotic medication is not titrated to high dosage.^{Reference Paudel, Dahal, Mathias, Arjyal and Markovic30}

Individuals with a positive history of involuntary psychiatric hospital admissions were more frequently prescribed high-dose antipsychotics. Indeed, a history of previous compulsory psychiatric treatments suggests a longer disease duration, which usually requires higher doses of antipsychotics to manage symptoms adequately.^{Reference Baker and Aebi11} The finding that those participants with a negative family history were more frequently prescribed two or more antipsychotics may be explained by a delayed approach to mental health services because of a lack of mental health literacy and knowledge, or stigma;^{Reference Okasha, Zaki, Meguid, El-Missiry, Sabry and Ismaeil31} it is expected that those who are more familiar with psychiatric symptoms may reach mental health services earlier if they experience relevant symptoms. Nevertheless, there might be additional confounding factors not explored, such as duration of untreated psychosis or age of onset.^{Reference Okasha, Zaki, Meguid, El-Missiry, Sabry and Ismaeil31}

Findings relating to the type of antipsychotics prescribed

The most frequently prescribed antipsychotics herein were olanzapine, risperidone, quetiapine, haloperidol and zuclopenthixol. Similarly, previous studies have shown that olanzapine, risperidone and quetiapine are the most frequently prescribed second-generation antipsychotics,^{Reference Wheeler A, Hu, Profitt, McMillan and Theodoros32} whereas haloperidol^{Reference Martinho, Poças, Simões and Laureano4} and zuclopenthixol^{Reference Mendes, Azeredo-Lopes and Cardoso2} are the most frequently prescribed first-generation antipsychotics^{Reference Mendes, Azeredo-Lopes and Cardoso2,Reference Martinho, Poças, Simões and Laureano4} in those involuntarily admitted for psychiatric treatment. Moreover, in line with the present results, previous data show that the most frequent combination of antipsychotics included one first-generation antipsychotic and one second-generation antipsychotic.^{Reference Mendes, Azeredo-Lopes and Cardoso2}

Findings relating to BMI

The association of low BMI with prescription of high-dose antipsychotics in this study is in contrast with previous data showing a positive association between increased BMI and prescription of high-dose antipsychotics;^{Reference Connolly and Taylor33} this relationship was explained on the basis of weight gain as a side-effect of increased doses of antipsychotics.^{Reference Rummel-Kluge, Komossa, Schwarz, Hunger, Schmid and Lobos34} The relationship needs to be further explored.

Findings relating to LAI and clozapine

There were additional interesting findings in the present study, such as under-prescription of LAI antipsychotics (3.7%); yet this is in line with the guidelines for the prescription of LAI antipsychotics (except monthly administration of paliperidone), which are licensed for maintenance therapy instead of controlling acute psychiatric conditions. Moreover, use of LAI antipsychotics require an adequate trial of oral therapy before administration.^{Reference Pierre, Mocrane, Philippe, Sebastien, Sylvie and Ludovic35}

Only one patient was prescribed clozapine, an antipsychotic with an indication for the treatment of resistant schizophrenia. This might be explained by the fact that clozapine requires a strict initiation and titration protocol for administration, thus making it difficult to be initiated within the first 72 h after involuntary hospital admission for compulsory treatment.^{Reference Flanagan, Lally, Gee, Lyon and Every-Palmer36} In contrast, data on community treatment orders, as a form of compulsory treatment of individuals diagnosed with a mental illness in the community, show that first-generation LAI antipsychotics and clozapine are associated with antipsychotic polypharmacy, and oral second-generation antipsychotics and LAI risperidone are both associated with prescription of high-dose antipsychotics.^{Reference Gisev, Bell and Chen37}

Implications

Recommendations derived from this study may include the revision of the PRN practice in psychiatric compulsory care, especially in relation to haloperidol and its combination with anticholinergics, as well as following international guidelines.^{Reference Essock, Schooler, Stroup, McEvoy, Rojas and Jackson38} Indeed, a review regarding PRN practice revealed a lack of documentation on PRN prescription as well as relevant adverse reactions.^{Reference Vaismoradi, Vizcaya Moreno, Sletvold and Jordan39} Tools for decision-making and national guidelines for PRN administration including monitoring are also needed, as well as education and training of healthcare professionals. Furthermore, investment in depot and clozapine clinics in the community may contribute to a decrease in relapse rates, and therefore of involuntary hospital admissions in people with psychotic symptoms; individuals who are prescribed with LAI antipsychotics and clozapine seems to have lower relapse rates.^{Reference Remington, Addington, Honer, Ismail, Raedler and Teehan1,Reference McMillan, Jacobs, Wilson, Theodoros, Robinson and Anderson5,Reference Taylor, Barnes and Young18} Additionally, non-pharmacological strategies, such as verbal and non-verbal de-escalation of aggressiveness, along with relevant education, should be provided.^{Reference El Abdellati, De Picker and Morrens27,Reference Vaismoradi, Vizcaya Moreno, Sletvold and Jordan39}

Polypharmacy results in risks with adherence to therapy because of the complexity of medication schemas.^{Reference Lung, Lee, Chen, Chan, Chang and Hui6} The goal is to identify the lowest therapeutic dose of medication to minimise side-effects without compromising effectiveness in terms of relapse prevention.^{Reference Baker and Aebi11,Reference Ortiz-Orendain, Castiello-de Obeso, Colunga-Lozano, Hu, Maayan and Adams12} Finally, the universal prophylactic use of anticholinergic medication in patients treated with antipsychotics needs to be further researched.

Strengths and limitations

The most important strength of this study was the examination of prescription patterns in a relatively large census sample of patients under compulsory treatment in the referral in-patient facility in Cyprus. Moreover, this is the first study nationally to describe prescription patterns with a focus on antipsychotics and hence it provides a much-needed baseline time point for future assessment of progress. Important demographic and clinical variables including diagnostics and the assessment of psychotic symptoms were taken into consideration, which is an additional strength of the present study.

Limitations include the cross-sectional design and lack of data regarding cross-titration during switching, length of hospital admissions, the pharmacotherapy relating to the management of side-effects and other practices justifying polypharmacy such as short-term use of an oral antipsychotic when starting a LAI. Additionally, information on previous clozapine use, which would justify polypharmacy, or high doses of antipsychotics, was not available.

It should be noted that during the study several pharmaceutic agents were not available at APH, such as the relatively new LAI antipsychotics (olanzapine, paliperidone, aripiprazole), a situation that may have influenced prescription patterns. An additional limitation regards the categorisation of the main symptomatology during participants admission. However, the categorisation applied was in line with the documentation applied in acute mental health services in Cyprus. Moreover, the fact that the length of involuntary hospital admission has not been recorded is an additional limitation. Future studies need to address this variable as well.

In conclusion, the present findings demonstrate suboptimal practices relating to antipsychotic prescription in the most vulnerable psychiatric patients, i.e. those under compulsory psychiatry care. Specifically, we described a high frequency of polypharmacy and PRN schemas, which are not consistent with clinical guidelines. These provide evidence of the need to revise antipsychotic prescription practices, aiming to improve safety and quality of care. The goal is to shift from antipsychotic polypharmacy to monotherapy without affecting either psychiatric symptoms severity or the frequency of relapse.