Antipsychotic prescribing practices for outpatients with schizophrenia and reasons for non-clozapine treatment - Data from a Danish quality assessment audit

Abstract

Background

Clozapine is the gold standard for treating treatment-resistant schizophrenia (TRS) although widely underutilised. Both organisational, patient- and clinician related reasons for the underutilisation have been reported, however, the clinical impact of either in real-world settings is not fully elucidated.

Aim

This audit aimed to evaluate the local antipsychotic (AP) prescribing practices for outpatients with schizophrenia and to assess the spectrum and prevalence of journalised reasons for non-clozapine treatment amongst eligible outpatients.

Methods

Data on demographics, current and former AP treatments, as well as documented reasons for non-clozapine treatment, was extracted through chart audit.

Results

Of the 668 affiliated outpatients with schizophrenia, 43% were treated with AP polytherapy (APP) and 19.6% with clozapine. The most prevalent reason for clozapine discontinuation was related to side effects whereas the most prevalent reason for refusal or omission of clozapine treatment was related to the associated monitoring regimen.

Conclusions

This audit showed that APP prescribing is a highly prevalent practice in our services when treating outpatients with schizophrenia and that clozapine is underutilised in a ‘last resort’ manner. The blood-monitoring regimen associated with clozapine treatment was found to be an important factor in the underutilisation. It seemed, however, that the monitoring constituted a barrier for different reasons, requiring different approaches to remedy. Future studies, directly involving both patients and clinicians in the identification and management of the most clinically relevant barriers and their corresponding facilitators, are warranted.

Keywords:

• Antipsychotic polytherapy
• clozapine
• refusal
• discontinuation
• treatment withholding

1. Introduction

Schizophrenia is a severe mental illness with a worldwide prevalence of approximately 0.45% [1]. The core symptoms within the positive, negative and cognitive domains entail the risk of substantial impact on an individual’s health and socioeconomic status [2]. Early, adequate treatment reducing the duration of untreated psychosis (DUP) has shown to be crucial to the long-term prognosis [3].

Treatment-resistant schizophrenia (TRS) is, as per consensus guideline [4], defined as schizophrenia with inadequate treatment response to two different antipsychotics (APs), trialled at an effective dose and duration in monotherapy and with acceptable treatment adherence [4]. It has been estimated that approximately 1/3 (21–34%) of all patients with schizophrenia are treatment-resistant (TR) [5–8]. Some patients will be TR from the psychosis onset [7,8], while others (10–60%) will develop TR over time [9]. The prevalence of TRS in a given population will hence depend on the type of population investigated; e.g. the prevalence of TRS in populations including a higher proportion of chronically ill patients has been reported as high as 45–60% [10] versus a prevalence in the range of 14–34% for populations of patients with first-episode psychosis [10].

Clozapine, the first atypical AP [11], is considered the gold standard for the treatment of TRS [4,12,13]. Clozapine is associated with an increased risk of some rare, but severe, adverse effects, including myocardia/cardiomyopathy and agranulocytosis, and is therefore subjected to requirements of extensive monitoring [12]. Clozapine has, however, in multiple observational studies proven to be superior to all other APs in treating TRS [14,15], with reported response rates as high as 80% [16–19]. Clozapine is moreover associated with fewer relapses and hospitalisations compared to other APs [20], better compliance [21], lower risk of discontinuation [14], lower general mortality [22], fewer suicides [23], less drug- and alcohol abuse [24], and a soothing effect on aggressive [25] and self-harming behaviour [26]. The probability and degree of clozapine response do, however, appear to be dependent on timing. Thus, factors such as a longer duration of illness, a higher number of hospitalisations, and a higher number of trialled APs before commencing clozapine appear to predict a lower probability of response [20,27–30]. Considering the impact of DUP [3], it seems in everyone’s interest to give TRS patients the opportunity of a clozapine trial as soon as they fulfil the criteria thereof.

Nevertheless, numerous studies have shown that clozapine is greatly underutilised in most countries [31–37], and that high-dose and/or AP polytherapy (APP) are preferred [38] and more frequently employed [39–42] treatment strategies, over clozapine. It is estimated that approximately 20% of all patients with schizophrenia [43], and up to 65% of the patients who subsequently start clozapine treatment [41], are subjected to APP. This is in spite of the fact that evidence showing benefits of APP over AP monotherapy (APM) [43–45] is sparse.

The prevalence of clozapine utilisation among patients with schizophrenia varies significantly between countries. Assuming that 1/3 of patients with schizophrenia are TRS, the optimal clozapine use would be 200/100,000 inhabitants [31]. In a report comparing 17 countries [31], the difference between the highest and lowest prevalence (0.6–189 per 100,000 inhabitants) constituted a factor of 300. In most countries, however, the national prevalence was shown well below 100/100,000 inhabitants. The average illness duration and the number of trialled APs before clozapine commencement has been reported as 6.7–9.7 years [33,35,36] and 2.4–5.5 trialled APs [29,41,42,46], respectively. The withholding of clozapine means that a substantial proportion of patients are being kept from timely and evidence-based treatment, with expected negative consequences to their long-term prognosis.

Several studies have attempted to identify the causes of clozapine underutilisation, and reviews [47–50] of the literature have found that the most often mentioned reasons are the clinicians’ concerns about severe adverse effects, the expectation of patients’ non-compliance with monitoring and/or drug intake, and patients’ refusal of clozapine treatment due to the monitoring requirements or expected side effects. However, the synthesised literature primarily consists of surveys of clinicians’ general perceptions, which may not reflect the impact of individual causes in a clinical setting. In chart audits [37,51–58], reasons for clozapine discontinuation are usually presented as an interpretation of the main cause, whilst reasons for clozapine withholding or reasons for patients’ refusal of clozapine treatment rarely are stated. It is thus difficult to conclude which specific barriers that carry the highest impact in a clinical context and which facilitating interventions are needed the most, to successfully increase the prevalence of clozapine treatment.

The main objective of this study was to assess current clinical prescribing practices for the antipsychotic treatment of outpatients with schizophrenia within the Mental Health Services of Region Zealand East and to assess journalised reasons as to why patients are not receiving clozapine treatment.

2. Materials and methods

2.1. Setting

In Denmark, a country of 5,9 million citizens [59], approximately 29,000 people are living with schizophrenia. It is estimated that only half of these are affiliated with the psychiatric system [60]. The national prevalence of clozapine-treated patients with schizophrenia has been reported to be approximately 10% [31,34], although with substantial geographical variations (5.7–16.8%) [34]. The Mental Health Services East in Region Zealand was in 2009 reported among the four most clozapine prescribing catchment areas [34] in Denmark (prevalence 14–17%). The Mental Health Services East handles the psychiatric treatment of adult citizens in the Roskilde, Lejre, Greve, Koege, Solroed Faxe and Stevns counties. No forensic units included. In total, 310,456 citizens lived in the catchment area at the time of the audit [61]. Patients are affiliated with local outpatient facilities according to geography, and the local facilities are governed under one of two administrative upper departments, upper departments 01 and 05. If the prevalence of schizophrenia is assumed 0.45%, and if only half of all patients with schizophrenia are assumed affiliated with psychiatric services, we would expect approximately 699 patients with schizophrenia to be affiliated with our outpatient services, of whom at least 1/3 would be expected TR and thus candidates for clozapine treatment.

2.2. Materials

To assess local patterns of antipsychotic prescribing for outpatients with schizophrenia, including patterns of clozapine use, we set out to identify all patients with a diagnosis of schizophrenia and an outpatient affiliation with the Mental Health Services East. We furthermore intended to audit patients’ medical records to collect data regarding current and former antipsychotic treatment as well as documented reasons for non-clozapine treatment.

2.3. Procedures

On the 7^th of May 2020, lists of patients were extracted from the electronic patient records following these criteria:

• Outpatient affiliation with The Mental Health Services East (upper departments 01 or 05).

• ICD-10 diagnosis of schizophrenia (F20.0-F20.9).

• AP prescriptions within following categories: non-clozapine (total number of APs > 1); non-clozapine (total number of APs = 1); non-clozapine (total number of APs = 0); or AP = clozapine (+).

A subsequent chart audit was conducted from May 8^th to July 31^st, 2020. The chart audit was facilitated using a designated quality assessment template, which allowed for access to medical chart data registered within five years from the audit. Data extracted during the chart audit included: sex, age, local outpatient affiliation, upper department affiliation, and current AP treatment.

Status as of a ‘high-needs’ community sample was assigned to patients with a local outpatient affiliation with day-hospital or an institutional residence. Data regarding status was extracted as a categorical variable.

For patients in APP or clozapine treatment, the following data were also extracted:

Previously trialled APs, any naïve clozapine initiation within the past five years (e.g. from May 8^th 2015 – May 8^th 2020) and number of adequately trialled APs before clozapine initiation as well as chart-documented reasons for discontinuation, omission, or refusal of clozapine treatment.

A previously trialled AP was defined as an AP prescribed as a planned, regular treatment > 24 h. APs prescribed as necessity medications (prn) or as single-dose prescriptions were not considered trialled, nor were APs where compliance was described as questionable or lacking. Adequately trialled APs were defined as APs trialled at an effective dose according to the manufacturer’s summary of product characteristics for a minimum of six weeks for oral prescriptions and 3 months for long-acting injectables.

Data regarding reasons for non-clozapine treatment was extracted as summaries of text.

2.4. Analyses

Data was extracted to Excel (2013) and analysed with descriptive statistics utilising the statistical software R [62], version 4.2.1. Distributions of normality were tested with histograms and qq-plots. Medians and interquartile ranges (IQR = Q1-Q3), or numbers and proportions, were calculated as applicable. As a secondary outcome, differences between independent subsamples were tested with the Mann–Whitney U test for continuous variables or the chi-square test for discrete variables. Fisher’s exact test was used to compare proportions when the subsample population was < 20 in total.

Qualitative data regarding reasons for discontinuation, omission, or refusal of clozapine treatment was coded and divided into descriptive categories based on the content of the text summaries. The coding was data-driven and allowed for the coding of multiple categories per case; no main reason was defined. The data was subsequently analysed descriptively as a categorical variable.

2.5. Ethical considerations and permits

The use of study data was approved by the local Data Protection Agency under registration no. REG-071-2020 and written permission to extract data from medical charts was obtained from local authorities in advance, in accordance with Danish law. Ethical approval was, however, not required due to the nature of quality assessment.

3. Results

3.1. Sample characteristics

In total, 668 patients with a diagnosis of schizophrenia and an outpatient affiliation with the Mental Health Services East were identified. Main population characteristics related to demographics and current AP treatment are presented in Table 1.

Table 1. Characteristics of antipsychotic prescribing for outpatients with schizophrenia.

Population characteristics	Psych.East, Total^a, N = 668
Demographics
Men	404 (60.5)
Age	41.0 (27.0–56.0)
High needs status^b	163 (24.4)
Treatment
No AP	51 (7.6)
AP monotherapy (APM)	330 (49.4)
AP polytherapy (APP)	287 (43.0)
Regular APP^c	204 (71.1)
2-drug treatment	249 (86.8)
3-drug treatment	35 (12.2)
4-drug treatment	3 (1.0)
LAI treatment^d	160 (24.0)
LAI as part of APP	87 (54.4)
Clozapine treatment	131 (19.6)
Most frequently prescribed AP for regular use	Olanzapine 167 (25.0)
Most frequently prescribed oral AP for regular use	Olanzapine 150 (22.5)
Most frequently prescribed LAI AP	Paliperidone 57 (8.5)
Most frequently prescribed prn AP^e	Quetiapine 83 (12.4)

Distribution of patients with schizophrenia and an outpatients affiliation with the Mental Health Services East (Psych.East, Total) by demographics and type of prescribed antipsychotic (AP) treatment.

^a Values are shown as medians and interquartile ranges (Mdn (Q1-Q3)) or as numbers and proportions in per cent (N (%)) as applicable.

^b The proportion of patients with status of ‘High-needs’ community sample, defined as patients with a local outpatient affiliation with day-hospital or an institutional residence.

^c Regular AP polytherapy = the total number of different, regularly prescribed APs ≥ 2; regular being defined as planned daily consumption or regular injection.

^d LAI: Long-acting injectable.

^e prn AP: antipsychotic prescribed ‘as needed’.

Of the 668 East patients with schizophrenia, 404 (60.5%) were men and the median age was 41 years (IQR 27.0–56.0). One hundred and sixty-three patients (24.4%) were categorised as ‘high-needs’.

Few patients (7.6%) did not receive any AP treatment, 49.4% were treated with APM, and 43.0% with APP. The majority of the APP patients (86.8%) were prescribed two APs, whilst 12.2% received treatment with three APs, and 1% were treated with four APs.

The proportion of patients treated with long-acting injectable (LAI) APs was 24.0%. Treatment with LAI AP was often (in 54.4% of cases) augmented with another AP.

Fourteen different APs were prescribed as regular treatments, with decreasing prevalence: olanzapine, aripiprazole, clozapine, quetiapine, paliperidone, zuclopentixole, risperidone, haloperidole, perphenazine, ziprasidone, lurasidone, chlorprothixene, sertindole, and levomepromazine.

The most frequently prescribed AP was olanzapine (25.0%). Aripiprazole and clozapine shared the second place, both prescribed to 19.6% of the affiliated patients (more on clozapine in a separate section). Paliperidone was the most frequently used AP for LAI treatment (8.5%), and quetiapine was the preferred AP for prn prescriptions (12.4%).

3.2. Subgroup analyses

Subgroup analyses of patient characteristics by type of prescribed antipsychotic (AP) (clozapine, non-clozapine) are presented in Table 2. A more in-depth characterisation, divided by affiliation with upper departments 01 or 05, is presented in the Supplementary Table, S1.

Table 2. Subgroup characteristics of antipsychotic prescribing for outpatients with schizophrenia, divided by clozapine/non-clozapine treatment.

Population characteristics according to AP treatment group	Non-clozapine^a, N = 537	Clozapine^a, N = 131
All groups
Men	304 (56.6)	89 (67.9)
Age	36.0 (26.0–54.0)*	52.0 (40.5–60.5)
High needs^b	118 (22.0)*	45 (34.4)
No AP therapy	51 (9.5)	0
Men	29 (56.9)
Age	27.0 (24.0-31.5)
High needs^b	2 (3.9)
AP monotherapy (APM)	280 (52.1)	50 (38.2)
Men	199 (71.1)	20 (40.0)
Age	39.0 (26.0–59.5)*	55.0 (42.0–65.5)
High needs^b	72 (25.7)	14 (28.0)
In LAI^c treatment	73 (26.1)	0
Most frequently prescribed AP for APM	Olanzapine 87 (31.1)	Clozapine 50 (100)
AP polytherapy (APP)	206 (38.4)*	81 (61.8)
Men	105 (51.0)	51 (63.0)
Age	39.0 (26.0-55.0)*	50.0 (38.0-57.0)
High needs^b	44 (21.4)*	31 (38.3)
2-drug treatment	182 (88.4)	67 (82.7)
3-drug treatment	22 (10.7)	13 (16.1)
4-drug treatment	2 (1.0)	1 (1.2)
In LAI^c treatment	82 (39.8)	5 (7.5)
In regular APP treatment^d	137 (66.5)	67 (82.7)
Most frequently prescribed non-clozapine AP for regular^d APP treatment	Aripiprazole 70 (51.1)	Aripiprazole 13 (19.4)
		Quetiapine 13 (19.4)
		Olanzapine 13 (19.4)
Most frequently prescribed prn^e AP	Quetiapine 71 (34.5)	Quetiapine 12 (14.8)
		Chlorprothixene 12 (14.8)
Patients with naïve clozapine initiation within 5 years	5 (2.4)	20 (15.3)^f
Number of adequately^g trialled APs prior to clozapine initiation	6.0 (3.0-6.0)	4.0 (3.0-5.0)
Age at naïve clozapine initiation	30.0 (25.0-40.0)	25.0 (23.0-32.3)

Distribution of the 668 patients with schizophrenia (DF20.0-20.9) and an outpatient affiliation with the Mental Health Services East (Psych.East, Total) by type of prescribed antipsychotic (AP) (clozapine, non-clozapine) and AP treatment group (NAP = 0, 1, >1).

^a Values are shown as medians and interquartile ranges (Mdn (Q1-Q3)) or as numbers and proportions in per cent (N (%)) as applicable; tested differences between groups are marked with cursive and bold text. Differences are tested with the Mann-Whitney U test for continuous data and with the Chi-square test for discrete data. Significant differences (p < .05) are marked with an asterisk*.

^b Status as of ‘High-needs’ community sample = patients with a local outpatient affiliation with day-hospital or institutional residence.

^c LAI: Long-acting injectable.

^d Regular APP = total number of regularly prescribed APs ≥ 2; regular being defined as planned daily consumption or regular injection.

^e prn AP: antipsychotic prescribed ‘as needed’.

^f Proportion of clozapine patients in total (n = 131).

^g Adequately trialled being defined as duration ≥ 6 weeks for oral prescriptions (3 months for LAI prescriptions) at ≥ min. effective dose according to the manufacturer’s summary of product characteristics.

3.2.1. Clozapine-treated patients

One hundred and thirty-one patients were treated with clozapine. The clozapine patients were significantly older than the patients in the total non-clozapine group (52.0 vs. 36.0, p < .001), the non-clozapine APM group (52.0 vs. 39.0, p < .001), and the non-clozapine APP group (52.0 vs. 39.0, p < .001), respectively. Forty-five clozapine patients (34.4%) were considered ‘high-needs’ (Table 2), which was a significantly higher proportion than for the non-clozapine patients (34.4 vs. 22.0, p = .003), also when adjusted for the influence of the No AP patients (34.4 vs. 23.9%, p = .015).

Of all patients treated with clozapine, 61.8% were prescribed augmentative APs. Aripiprazole, olanzapine, and quetiapine were equally often prescribed as augmentations to clozapine. Twenty clozapine patients (15.3% of clozapine patients) had naïvely commenced clozapine treatment within the five-year audit period (between the 8^th of May 2015 and the 8^th of May 2020). The median age at naïve clozapine commencement was 25.0 years (IQR 23.0–32.3), and the median number of adequately trialled APs before clozapine commencement was 4.0 (IQR 3.0–5.0).

3.2.2. Non-clozapine-treated patients

3.2.2.1. No AP treatment

Of the 537 non-clozapine patients, 51 (9.5%) were not receiving any AP treatment at all (Table 2). The non-AP treated patients were significantly younger than all the other treatment groups (29.7 vs. 50.4, 40.8 and 40.1 years, p < .001). Two (3.9%) were ‘high-needs’.

3.2.2.2. AP monotherapy

Two hundred and eighty non-clozapine patients (52.1%) were treated with APM (Table 2). Olanzapine was the most frequently prescribed AP for APM (31.1%). Seventy-two patients (25.7%) were ‘high-needs’.

3.2.2.3. AP polytherapy

Two hundred and six of the non-clozapine patients (30.8% of the total sample) were treated with APP (Table 2). The most frequently prescribed AP for regular administration in the non-clozapine APP group was aripiprazole (34.0%). The most frequently prescribed prn AP was quetiapine (34.5%). Forty-four patients (21.4%) were ‘high-needs’.

At least 124 of the non-clozapine APP patients (60.2%) had trialled two or more different APs as a regular prescription before their current APP treatment and 118 APP patients (57.3%) had furthermore trialled other combinations of APP than the current one, within the five-year audit period.

Thirteen APP patients had been treated with clozapine within the last five years but were now discontinued. Five of these had naïvely commenced clozapine within the five-year audit period. For these patients, the median number of adequately trialled APs before clozapine commencement was 6.0 (IQR 3.0–6.0) and the median age at commencement was 30.0 years (IQR 25.0–40.0) (Table 2). When comparing the naïvely commenced clozapine patients who stayed in clozapine treatment (n = 20) to those who discontinued (n = 5), the continuers tended to be younger (25.0 vs. 30.0 years, p = .434) and to have trialled fewer APs prior to naïve clozapine commencement (4.0 vs. 6.0, p = .309), than the discontinuers. These differences were, however, not statistically significant.

Nine APP patients had been offered, and refused, clozapine treatment within the 5-year audit period. Of these patients, five were clozapine-naïve whilst four had been treated with clozapine before.

A further eight APP patients had a documented reason as to why they had not been offered clozapine.

3.2.3. Upper department 01 vs. 05

Differences in prescribing between departments 01 and 05 are presented in the Supplementary Table 1 (S1).

The proportions of patients receiving no-AP therapy, non-clozapine APM, or non-clozapine APP did not differ significantly between departments. The patients in upper department 01 were, however, significantly older (44.0 vs. 37.0 years, p = .003) and more frequently ‘high-needs’ (27.6% vs. 18.3%, p = .008) than patients in upper department 05.

Upper department 01 had furthermore a higher prevalence of patients treated with clozapine than department 05 (22.6% vs. 13.9%, p = .007) and the proportion of clozapine patients who were ‘high-needs’ was also significantly higher in department 01 than 05 (40.4% vs. 15.6%, p = .011).

The proportion of naïvely commenced clozapine-patients within the five-year audit period (15.2% vs. 15.6%, p = .85) and the proportion of non-clozapine APP patients who had previously been treated with clozapine at any time within the last five years (6.7% vs. 5.5%, p = .735) was, however, similar between departments.

3.3. Documented reasons for non-clozapine treatment

The chart-documented reasons as to why patients were not being treated with clozapine were often described as multifactorial (12/28 cases). The distribution of reasons are presented at case-level in Table 3 and are summarised below:

Table 3. Chart-documented reasons as to why patients are not receiving treatment with clozapine.

Case No.	Reasons for clozapine discontinuation	Supplementary reasons	Reasons for clozapine refusal	Reasons for clozapine non-candidacy
1	Insufficient response
2	Drug incompliance	Drug incompliance, Monitoring
3	Side effects, Somatic issues	Somatic issues	Unspecified reason
4	Side effects
5	Drug incompliance, Side effects
6	Side effects
7	Side effects, Stable condition
8	Insufficient response
9	Drug incompliance, Psychosis	Monitoring
10	Side effects
11	Side effects, Insufficient response
12	Side effects, Insufficient response
13	Drug incompliance, Side effects		Side effects
14			Monitoring
15			Monitoring, Psychosis
16			Monitoring, Responsibility with clozapine
17		Monitoring	Monitoring, Side effects
18			Monitoring
19			Hospitalisation
20			Side effects
21				Monitoring
22				Monitoring
23				Monitoring, Unwillingness to change medicines
24				Drug incompliance
25				Monitoring
26				Monitoring
27				Monitoring
28				Monitoring, Drug incompliance
Total	Side effects (9/13 cases); insufficient response (4/13 cases); drug incompliance (4/13 cases); stable condition (1/13 cases); psychosis (1/13 cases); somatic issues (1/13 cases).	Expected monitoring issues (3/4 cases); Expected drug incompliance (1/4 cases); somatic issues (1/4 cases).	Monitoring requirements (5/9 cases); side effects (3/9 cases); need for hospitalisation (1/9 cases); unspecified reason (1/9 cases); psychosis (1/9 cases), responsibility with clozapine management.	Expected monitoring issues (7/8 cases); expected drug incompliance (2/8 cases); unwillingness to change medicines (1/8 cases).

Distribution of case-specific reasons for non-clozapine treatment according to category of journalised reason(s).

3.3.1. Clozapine discontinuation

In total, six different reasons for discontinuation were mentioned (see Table 3). The reason for discontinuation in the majority of cases was related to side effects (9/13 cases), followed by an insufficient response (4/13 cases), issues with drug compliance (4/13 cases), stabilised condition (1/13 cases), psychosis (1/13 cases), and somatic admission with pneumonia (1/13 cases).

Overall, 10 different side effects were mentioned as reasons for clozapine discontinuation. Weight gain was the most frequently mentioned side effect (3/9 cases). The remaining side effects were blood dyscrasias (2/9 cases), constipation (1/9 cases), sedation (1/9 cases), hunger (1/9 cases), fainting (1/9 cases), hypersalivation (1/9 cases), worsening of depression (1/9 cases), nocturnal bedwetting (1/9 cases), and unspecified (1/9 cases).

3.3.2. Clozapine refusal

Nine patients had refused an offer of clozapine treatment within the five-year audit period. In most cases (6/9 cases) clozapine had been offered during hospitalisation. The chart-documented reasons for the patients’ refusal of clozapine treatment were divided into six different categories (see Table 3). Most frequently, the reason was related to different aspects of the monitoring regimen (5/9 cases). Accordingly, one case of refusal due to monitoring was due to fear of blood sampling, three due to the extensiveness of monitoring, and one was stated as due to the need for blood sampling, but not otherwise specified. The other categories were known side effects (3/9 cases), the need for hospitalisation (1/9 cases), psychosis (1/9 cases), too much responsibility related to clozapine management (1/9), and one unspecified reason (1/9 cases).

3.3.3. Clozapine non-candidacy

Eight APP patients had a chart-documented reason as to why they were not offered clozapine treatment (Table 3). Considerations about clozapine were mainly documented during hospitalisation (6/8 cases). In seven out of eight cases, these reasons were described as related to expected monitoring issues (6/7 cases due to expected incompliance with monitoring, 1/7 cases due to concerns about transport to monitoring facilities); drug incompliance was stated as the reason in two cases, and the lack of interest in changing medications in one case. Further, in four cases of clozapine discontinuation or refusal, the reasons were supplemented with a clinical argument as to why the patient was not considered a candidate for clozapine treatment after all (Table 3). In most cases (3/4), the reasoning was related to concerns about compliance with the monitoring requirements, regardless if this was part of the reason for discontinuation (case no. 2 and 9, Table 3) or if the patient requested a retrial (case no.9). Considerations regarding drug-incompliance (1/4 cases) and somatic issues (chronic mild neutropenia) (1/4) were, however, also mentioned.

4. Discussion

Approximately 1/3 of all patients with schizophrenia are assumed TR [44] and thus eligible for a trial of clozapine. Studies of outpatient populations including chronically ill patients have even reported TRS prevalences as high as 45–61% [10]. However, clozapine is widely underutilised. The objectives of this audit were to evaluate the current local prescribing practice of antipsychotics for the treatment of outpatients with schizophrenia and to assess the scope and prevalence of journalised reasons why clozapine-eligible patients are not in clozapine treatment. We expected approximately 699 affiliated outpatients with schizophrenia of whom at least 1/3 would be expected TR.

The audit showed 668 affiliated outpatients with schizophrenia corresponding an expected minimum of 223 TRS patients. We found that 131 of the 668 affiliated outpatients (19.6%) were treated with clozapine. A clozapine prevalence of 19.6% is thus higher than the 14–17% reported in a previous study [34]; however, it is still substantially lower than the expected minimum proportion of clozapine-eligible patients. The current prevalence of clozapine-treated patients is furthermore based on the proportion of patients affiliated with the psychiatric outpatient services. Considering a catchment area of 310,456 inhabitants and an assumed prevalence of schizophrenia of 0.45% (n ≈ 1397), the actual clozapine prevalence for the catchment area is probably smaller.

The utilisation of APP, on the other hand, was found to be a highly prevalent practice within the Psychiatry East, currently prescribed to 43.0% of the affiliated outpatients. The audit furthermore showed that 57% of the APP outpatients had tried other combinations of APP within the last five years – not counting combinations given only during hospitalisation. The utilisation of APP is not evidence-based, nor recommended, clinical practice [40,43,63], although commonly employed. Some systematic reviews and meta-analyses have suggested a potentially beneficial effect on negative symptoms, neuroleptic-induced hyperprolactinemia, and weight gain with the augmentation of aripiprazole to other APs [43,45,63,64]. However, guidelines advise against the use of APP before other evidence-based treatments, such as clozapine, have been trialled in monotherapy [39,43]. A mitigating circumstance, in this case, might be that aripiprazole was the most frequently utilised AP for regularly prescribed APP combinations.

Reasons for APP prescribing were not among the collected data items of this chart audit, and an analysis of these reasons are hence not possible. However, other studies have found that APP prescribing is often justified by the insufficient response of APM, and/or development of side effects when attempting to increase the dose of the primary AP [40,43,63]. Treatment with APP can thus be assumed as an expression of the inadequate effect of APM, which in this case applies to 206 of the non-clozapine patients. Which of these patients that truly are TR, and thus candidates for a trial of clozapine, is to be evaluated according to the effect, dose, and duration of the previously trialled APs. Moreover, any previous trials of clozapine treatment, refusal of clozapine treatment, or other reasons as to why the patient in question is not to be treated with clozapine, should be considered. These parameters could not adequately be elucidated within the framework of quality assessment due to the five-year limit on accessible medical chart data.

It is, however, noteworthy that 124 APP patients had trialled at least two different APs as a regular prescription before their current APP treatment, of whom only 13 had been treated with clozapine at any time within the last five years and five had been offered a trial of clozapine.

This indicates that among the 206 APP patients, there are 111 clozapine-naïve patients with possible TRS, a number that fairly corresponds the difference between the minimum expected (n=223) and the observed (n=131) number of clozapine-treated patients.

A number of 111 clozapine-naïve TRS patients would, based on the number of clozapine-treated patients (n = 131) and previously clozapine-treated patients (n = 13), furthermore constitute a clozapine-naïve TRS proportion of 43.5%. This proportion would be in line with findings from other studies, reporting that around half of the TRS patients (46–52%) in outpatient settings are clozapine-naïve [10,65].

The proportion of clozapine patients receiving APP was also considerable (62%). It was nearly the same as the previously reported proportions of 64.2–65.6% from Denmark and Iceland [34,66]; however, similar audits from e.g. the U.K. and Spain [37,67] have reported proportions of only 10 and 22%, respectively. The high proportion of clozapine patients in need of APP in our services could be the consequence of clozapine treatment being initiated too late in the treatment course for it to reach its full potential, and/or that clozapine treatment is not being sufficiently adjusted according to serum levels. Guidelines recommend trialling clozapine for a minimum of 3–6 months, and increasing serum levels to the therapeutic level, before concluding a lack of response and/or augmenting with another AP [4,44,68]. This is because clozapine responses can be delayed [5] and because APP prescribing increases the risk of high-dose treatments and a subsequent development of side effects [68]. Evidence of the benefits from AP augmentation to clozapine is furthermore sparse [44,64,69], and only for aripiprazole [44]. Additionally, guidelines advise against augmentation with drugs that share a side-effect profile with clozapine [40,68].

Within the Mental Health Services East, quetiapine, olanzapine, and aripiprazole were equally utilised as augmentations to clozapine treatment. However, both quetiapine and olanzapine have distinct antihistamine effects and share side-effect characteristics with clozapine in terms of weight gain and sedation.

Side-effects have been reported to be the most frequently mentioned reason for clozapine discontinuation [51–55], and sedation has been reported to be the most frequent, non-haematological side-effect causing discontinuation [51,52].

In this population, side effects were also the most frequently stated reason for clozapine discontinuation (9/13 cases). However, the most frequent side-effect causing clozapine to be discontinued in this population was weight gain (3/9 cases). This emphasises that augmentative drugs of further sedating and weight-gaining characteristics should be avoided to minimise the side-effect burden and unnecessary clozapine cessations.

The patients affiliated with The Mental Health Services East were treated in local outpatient services governed under two administrative departments, upper departments 01 and 05, according to location. Unexplained national geographical differences in clozapine prescribing between hospitals have previously been reported [34] and another quality assessment found evidence of differences in prescribing between departments [37]. Therefore, as a secondary outcome, we sought to assess local differences in prescribing by stratifying the patient characteristics according to local outpatient affiliation. The departments did not differ significantly in terms of the proportions of patients receiving no-AP therapy, APM or APP, respectively. There were, however, other significant differences between departments: the patients in department 01 were older (44.0 vs. 37.0 years), more frequently of ‘high-needs’ status (27.6% vs. 18.3%) and more often treated with clozapine (22.6% vs. 13.9%) than the patients in department 05.

Clozapine is often regarded as the treatment option of last resort. One simple explanation of the difference in clozapine-prescribing between departments could therefore be the fact that department 01 represents the majority of the ‘high-needs’ patients, i.e. the more severely ill patients, in the catchment area, and thereby a greater part of those patients who will eventually end up receiving clozapine. This ‘last resort’ manner of prescribing is substantiated by the high proportion of ‘high-needs’ patients in the total clozapine group (34.4% vs. 23.9% of the AP treated non-clozapine patients). When finally commenced, clozapine patients have been shown to stay in treatment for longer than patients receiving other AP treatments [70], and the number of clozapine-treated patients in department 01, as well as their mean age, can thus be assumed to accumulate over time.

Another, or an additional, explanation for both the geographical differences in prescribing and the low total clozapine-prevalence could be attributed to the local clinicians’ personal preferences and attitudes towards clozapine treatment. A previous Danish study [38] found that 64% of the surveyed clinicians preferred to combine two different APs, rather than prescribing clozapine. The study also showed that a substantial proportion of clinicians (49%) would only initiate clozapine treatment after three or more AP trials [38]. Unfortunately, it has also been reported that illness duration and the number of APs trialled before clozapine initiation are predictors of clozapine response [20,29], with a decreasing probability of response of 10% for every AP trialled before clozapine [20].

In this audit, we found that 25 previously clozapine-naïve patients had commenced clozapine treatment during the last five years. Of these, 20 patients were still in clozapine treatment, whilst five had discontinued. The median number of adequately trialled APs before clozapine commencement was, for both groups, more than double the two APs minimally required according to guidelines. For the patients who had discontinued clozapine treatment, the median number of previously trialled APs was 6.0, whereas for the patients still receiving clozapine it was 4.0. Similarly, the median age at clozapine commencement was higher amongst discontinuers than continuers (30.0 vs. 25.0 years). This could indicate that the risk of clozapine discontinuation increases with the number of previously trialled APs and/or age at commencement, maybe due to the decreased probability of response. However, the number of observations was small, and the differences did not reach statistical significance.

Data regarding illness duration was not available for extraction as per chart audit and its impact on clozapine discontinuation could therefore not be assessed.

In this audit, we also assessed the documented reasons for patients’ refusal of clozapine treatment and clinicians’ reasons for clozapine non-candidacy. Only few reasons had been journalised, however, based on the available chart-material, we found that reasons why patients are not receiving clozapine treatment are often described as multifactorial (12/28 cases), although the most frequently mentioned reason, for both non-candidacy and the patients’ refusal, was related to the monitoring regimen. Further, in two cases of clozapine-discontinuation, a supplementary reason for non-candidacy was described as related to the monitoring requirement, even though the reason for discontinuation was unrelated to the monitoring. These findings are in accordance with results from other studies reporting on barriers to clozapine treatment [48–50], in which the expected lack of compliance with, or acceptance of, the blood monitoring are some of the most frequently stated barriers. Interestingly, only one of the five monitoring-related reasons for patient refusal was described as a refusal of the blood sampling itself (here, due to fear of blood sampling). We are not aware of any other ‘refusal studies’ for comparison, however, based on these data; it seems as if the barrier of monitoring consists of several distinct barriers, related to different aspects of the monitoring, thus requiring different facilitating interventions. Several emerging questions warrants further investigation: what aspects of the monitoring constitute the most prominent barriers to clozapine treatment, and how can they be remedied? Moreover, was issues related to monitoring also the reason for non-clozapine treatment in the non-journalised cases? The audit did not provide us with an answer to these questions, but perhaps future, more case-specific, studies with direct questioning of non-clozapine patients and their clinicians could?

In summary, APP prescribing is, both currently and historically, common clinical practice when treating outpatients with schizophrenia in the Mental Health Services East. Further, the audit showed that clozapine is utilised too little and too late in our services, indicating a practice of ‘last resort’. Amongst journalised reasons for non-clozapine treatment, the blood-monitoring regimen associated with clozapine treatment was found to be an important factor in the underutilisation, seemingly, for several different reasons. Unfortunately, the impact of distinct issues related to the monitoring could not be deduced from the chart material and is yet to be elucidated. Future studies, directly involving both patients and clinicians in the identification and management of the most clinically relevant barriers and facilitators of clozapine treatment, appear essential in order to succeed with further implementation.

Author contributions

Author Jakobsen conceptualised the audit in consultancy with authors Nielsen and Simonsen. Author Jakobsen performed the data extraction, management and analyses. Author Jakobsen wrote the first draft of the manuscript in Danish. Author Schaug translated the draft and tables into the English language and author Jakobsen rewrote it into its current form. Authors Schaug, Simonsen and Nielsen critically reviewed the draft for academic and expert consensus. All authors have read and accepted the final version of the manuscript.

Disclosure statement

This work is part of author Jakobsen’s PhD project, partially funded by a bequeathed donation favouring patient-oriented research within the Mental Health Services of Region Zealand. The authors have no other competing interests to declare.

Data availability statement

Due to the nature of this audit, participants of the study did not agree for their data to be shared publicly, hence supporting data is not available.

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

This work is part of a PhD project funded by the Mental Health Services East, Region Zealand Psychiatry and the Centre for Evidence-Based Psychiatry, Psychiatric Research Unit, Region Zealand Psychiatry. No grant no. provided. The funding by the Psychiatric Research Unit derived from a bequeathed donation favouring patient-oriented research in the region.

Notes on contributors

Michelle I. Jakobsen

Michelle Iris Jakobsen, MD, PhD-student at the Department of Clinical Medicine, University of Copenhagen and the Mental Health Services East, Region Zealand Psychiatry, Roskilde, Denmark.

Julie P. Schaug

Julie Perrine Schaug, Cand.Psyc., Research Assistent at the Psychiatric Research Unit, Centre for Evidence-Based Psychiatry, Region Zealand Psychiatry, Slagelse, Denmark.

Jimmi Nielsen

Jimmi Nielsen, MD, PhD, Professor, Specialist in psychiatry, the Department of Clinical Medicine, University of Copenhagen and Psychiatric Centre Glostrup, The Mental Health Services in the Capital Region of Denmark.

Erik Simonsen

Erik Simonsen, MD, PhD, Professor, Specialist in psychiatry, the Department of Clinical Medicine, University of Copenhagen and the Mental Health Services East, Region Zealand Psychiatry, Roskilde, Denmark.