Introduction
Antipsychotics are effective at reducing the risk of relapse in schizophrenia at 1-year follow-up (Leucht et al. Reference Leucht, Tardy, Komossa, Heres, Kissling and Davis2012). The practice of prescribing multiple and high dose antipsychotics is common (Faries et al. Reference Faries, Ascher-Svanum, Zhu, Correll and Kane2005), not supported by the evidence base (Goren et al. Reference Goren, Parks, Ghinassi, Milton, Oldham, Hernandez, Chan and Hermann2008; Goodwin et al. Reference Goodwin, Fleischhacker, Arango, Baumann, Davidson, de Hert, Falkai, Kapur, Leucht, Licht, Naber, O’keane, Papakostas, Vieta and Zohar2009; CADTH 2012; Donnelly et al. Reference Donnelly, Rathbone and Adams2013) and is associated with increased side effects and cost (Lochmann van Bennekom et al. Reference Lochmann van Bennekom, Gijsman and Zitman2013). While several studies report a reduction in all-cause mortality in schizophrenic patients prescribed antipsychotic medication (Tiihonen et al. Reference Tiihonen, Lonnqvist, Wahlbeck, Klaukka, Niskanen, Tanskanen and Haukka2009) the rate of cardiac mortality is higher than among those not prescribed antipsychotics and increases with higher doses (Ray et al. Reference Ray, Chung, Murray, Hall and Stein2009). As required (PRN) prescribing was identified as a major cause of both high dose and combined antipsychotic prescribing (Paton et al. Reference Paton, Barnes, Cavanagh, Taylor and Lelliott2008) and occurs in up to half of psychiatric in-patients (Fishel et al. Reference Fishel, Ferreiro, Rynerson, Nickell, Jackson and Hannan1994).
Evidence-based guidelines recommend commencing monotherapy at the lower end of the licensed dose range and titrating upwards within the dose range (NICE 2009). Switching from combined antipsychotic therapy to monotherapy is a viable option for those who do not have refractory schizophrenia and can reduce side effects (Essock et al. Reference Essock, Schooler, Stroup, Mcevoy, Rojas, Jackson and Covell2011; Tani et al. Reference Tani, Uchida, Suzuki, Fujii and Mimura2013). In a study by Wunderink et al. (Reference Wunderink, Nieboer, Wiersma, Sytema and Nienhuis2013) dose reduction when in remission can result in better long term functional outcomes, including self-care, relationships, community integration and vocational functioning as measured by the Groningen Social Disability Schedule. A meta-analysis of 14 studies (N=734) investigating clozapine augmentation with antipsychotics in patients with treatment-resistant schizophrenia demonstrated a modest benefit compared with augmentation with placebo (Taylor et al. Reference Taylor, Smith, Gee and Nielsen2012).
A Prescribing Observatory for Mental Health United Kingdom (POMH-UK) audit carried out in 2006 and 2007 involving 32 services in the United Kingdom and Ireland found little change in the rates of multiple, high dose and PRN antipsychotic prescribing in the majority of services, after the introduction of a wide array of different quality improvement programmes (Paton et al. Reference Paton, Barnes, Cavanagh, Taylor and Lelliott2008). These included, but were not limited to, a benchmarked audit report, electronic slide presentation, educational work book, antipsychotic dose ready reckoner, time series chart that enabled clinical staff to chart their use of multiple antipsychotics and high dose antipsychotics over time so that trends could be identified, educational posters, workshops and various combinations thereof.
Standards set for audit
The standards set were the evidence-based guidelines underpinning the POMH-UK audit of 2006–2007 (Paton et al. Reference Paton, Barnes, Cavanagh, Taylor and Lelliott2008) that were derived from the following sources: National Institute of Health and Care Excellence guidelines (NICE 2009), British National Formulary (BNF) (Joint Formulary Committee, BMA, RPSOGB 2012) and the Royal College of Psychiatry guidelines (RCPSYCH 2006).
The POMH-UK audit standards for adherence to quality prescribing guidelines are as follows:
1. The total daily prescribed doses of antipsychotic drugs are within BNF/Summary of Product Characteristics limits. High dose is defined as a total daily dose, exceeding 100% of the maximum recommended daily dose (RCPSYCH 2006).
2. Individuals should be prescribed only one antipsychotic at a time with the exception of cross titration during switching from one antipsychotic to another and those requiring augmentation of clozapine (NICE 2009).
3. First (typical) and second-generation (atypical) antipsychotic drugs (SGAs) should not be prescribed concurrently, except during switching from one generation to another (NICE 2009).
Ethical approval
The audit was approved by the St Patrick’s and St Edmundsbury’s Hospitals’ Audit Committee.
Methods
Over a 4-day period in 2008 all adult in-patients (N=281) had their medication sheet and clinical record separately reviewed by two independent investigators. Of these, 161 patients were prescribed regular and/or PRN antipsychotic medication. The following information was gathered for any patient prescribed antipsychotic medication: age, gender, ICD 10 diagnosis code, Mental Health Act status, medication name, dose, frequency and generation of antipsychotic prescribed. If multiple and/or high dose antipsychotics were prescribed any reasons documented for this prescribing were recorded.
The protocol established by POMH-UK their 2006 and 2007 audit of antipsychotic prescribing in mental health facilities was followed (RCPSYCH 2006). This included:
∙ The maximum prescribed dose that could be administered over a 24 hours period was recorded, for both regular and PRN prescriptions, irrespective of whether they were administered or not.
∙ The dose of each antipsychotic medication was compared with the BNF maximum licensed dose.
∙ The percentages were summed to determine the total daily antipsychotic dose percentage that could be given in 24 hours (Paton et al. Reference Paton, Barnes, Cavanagh, Taylor and Lelliott2008).
Over a 2-day period in June 2012, an identical data-gathering process was repeated for all adult in-patients (N=239). Of these 131 patients were prescribed regular and/or PRN antipsychotic medication. This data was separately reviewed by two independent investigators and a consensus meeting, involving a senior colleague, held to resolve any disagreements.
In 2008 two patients prescribed clozapine were augmented by regularly prescribed antipsychotics, one first generation and one second generation. In 2012, three patients were prescribed clozapine augmented with regularly prescribed antipsychotics, two of which were second generation and one first generation. All of the patients prescribed clozapine and regular antipsychotics as augmenting agents were excluded as per the POMH-UK audit standards.
Data collected was analysed using SPSS 20.0. χ 2 testing and cross tabulation were used to determine statistical significance.
Intervention
The intervention, implemented in December 2011, consisted of a hospital-wide clinical policy titled ‘Prescribing and monitoring of antipsychotic medication, including high dose antipsychotic medication’ (SPUH 2011).
The primary objectives of the policy were that:
∙ Treatment with antipsychotic medication should be as effective, safe and as well tolerated as possible.
∙ Treatment with antipsychotic medication should consider patient preference and relative potential of antipsychotics to cause extrapyramidal side effects, metabolic side effects and other side effects.
∙ Concurrent use of two or more antipsychotics should be avoided where possible, in particular the use of first generation and second generation in combination.
Additional guidelines for high dose antipsychotic treatment (HDAT) included that:
∙ The decision to prescribe HDAT should be made by the consultant psychiatrist and documented in the clinical notes, including the clinical justification for this decision, the risks, benefits and assessment of outcomes.
∙ A sheet monitoring physical observations and baseline and ongoing investigations was introduced for patients on high dose antipsychotics.
All departments – pharmacy, nursing, consultant hospital doctors and non-consultant hospital doctors (NCHD) – were involved in creating, disseminating and implementing the policy throughout the organisation. All staff were made aware of the new policy through email contact, formalised staff training sessions, informal discussion during multi-disciplinary team (MDT) meetings, pharmacy-led interventions during ward rounds and MDT meetings and during induction training for new staff. In addition the policy was available both on the wards and via the organisation’s intranet. The policy was included in NCHD induction and presented at the NCHD weekly educational meetings. Educational sessions were also held for all MDT members.
Results
In 2008, 161 of 281 in-patients were prescribed antipsychotic medication compared with 131 of 239 in-patients in 2012 (p=0.60). There were no significant differences in age, gender, Mental Health Act status, psychiatric diagnosis or documented indications for prescribing multiple or high dose antipsychotics between the two groups. The number of patients prescribed multiple antipsychotics decreased from 45 of 161 patients prescribed antipsychotic medication in 2008 to 21 of 131 patients prescribed antipsychotic medication in 2012 (p=0.02).
The prescribing of multiple first-generation antipsychotics (FGAs) including PRN’s decreased from 5 of 45 patients prescribed multiple antipsychotics in 2008 to 0 of 21 prescribed multiple antipsychotics including PRN’s in 2012 (p=0.17). In 2008, 11 patients were prescribed multiple SGAs including PRN’s out of 45 compared with 12 out of 21 prescribed multiple SGAs including PRN’s in 2012 (p=0.01) (Fig. 1).
Fig. 1 Multiple antipsychotic prescribing 2008 and 2012. *p<0.05. FGA, first-generation antipsychotics; SGA, second generation antipsychotics.
There was a decrease in the prescribing of mixed generations of antipsychotics from 29 of 45 patients in 2008 to 9 of 21 patients in 2012 (p=0.12). The rate of high dose antipsychotic prescribing remained stable at 13 of 161 patients in 2008 and 10 of 131 in 2012 (p=1.00) (Fig. 1).
Fig. 2 PRN contribution to high dose and multiple antipsychotic prescribing (2008 and 2012).
The number of patients prescribed a PRN antipsychotic fell from 25 of 161 patients prescribed antipsychotic medication in 2008 compared with 18 of 131 patients prescribed antipsychotic medication in 2012 (p=0.74). Fourteen of the 45 patients prescribed multiple antipsychotics in 2008 were prescribed PRN antipsychotics. This fell to 6 of 21 patients prescribed multiple antipsychotics in 2012 (p=1.00). Seven of 13 patients prescribed high dose antipsychotics in 2008 were prescribed a PRN antipsychotic while 2 of 10 patients prescribed high dose antipsychotics in 2012 were prescribed a PRN antipsychotic (p=0.20) (Fig. 2).
In 2008, of the 20 patients diagnosed as having a psychotic disorder (F20–29), 12 were prescribed multiple antipsychotics while in 2012, 2 of 18 patients diagnosed with a psychotic disorder were prescribed multiple antipsychotics (p<0.01). In 2008, 5 out of 13 patients prescribed high dose antipsychotics were diagnosed with a psychotic disorder compared with none of the 10 prescribed high dose antipsychotics in 2012 (p=0.05).
Of the 97 patients diagnosed as having a mood disorder (F30–39) in 2008, 23 were prescribed multiple antipsychotics compared with 16 of 81 prescribed multiple antipsychotics in 2012 (p=0.06). In 2008, 6 of 13 patients prescribed high dose antipsychotics were diagnosed as having mood disorders (F30–39) compared with 7 of 10 patients prescribed high dose antipsychotics in 2012 (p=0.40).
Discussion
The CATIE (Lieberman et al. Reference Lieberman, Stroup, Mcevoy, Swartz, Rosenheck, Perkins, Keefe, Davis, Davis, Lebowitz, Severe and Hsiao2005) and CUtLASS (Jones et al. Reference Jones, Barnes, Davies, Dunn, Lloyd, Hayhurst, Murray, Markwick and Lewis2006) trials challenged the concept that SGAs were superior to FGAs in terms of symptom reduction and improved quality of life. The World Psychiatric Association’s review of 1600 randomised controlled trials concluded that SGA’s had a lower propensity to cause extrapyramidal side effects and tardive dyskinesia, though greater propensity to result in metabolic side effects (Tandon et al. Reference Tandon, Belmaker, Gattaz, Lopez-Ibor, Okasha, Singh, Stein, Olie, Fleischhacker and Moeller2008). Excluding clozapine, which has consistently been shown to be the most effective antipsychotic (Leucht et al. Reference Leucht, Cipriani, Spineli, Mavridis, Orey, Richter, SAMARA, Barbui, Engel, Geddes, Kissling, Stapf, Lassig, Salanti and Davis2013), there are minimal differences between FGAs and SGAs in treating positive symptoms (Abou-Setta et al. Reference Abou-Setta, Mousavi, Spooner, Schouten, Pasichnyk, Armijo-Olivo, Beaith, Seida, Dursun, Newton and Hartling2012, Rosenheck & Lin Reference Rosenheck and Lin2014).
The classification of antipsychotics into FGA and SGA groups may be an oversimplification and it has been proposed that this distinction be rejected (Tyrer & Kendall Reference Tyrer and Kendall2009). A meta-analysis of 212 trials, by Leucht et al. suggest that instead of grouping antipsychotics into FGA and SGA, hierarchies in different domains, such as all-cause discontinuation, weight gain, extrapyramidal side effects, prolactin increase, QTc prolongation and sedation, would be more useful in determining the optimal antipsychotic for an individual (Leucht et al. Reference Leucht, Cipriani, Spineli, Mavridis, Orey, Richter, SAMARA, Barbui, Engel, Geddes, Kissling, Stapf, Lassig, Salanti and Davis2013). Notwithstanding this since the POMHS-UK Guidelines (Paton et al. Reference Paton, Barnes, Cavanagh, Taylor and Lelliott2008) used this classification we also utilised it as the basis for classifying antipsychotic medications in this study.
The primary mechanism of action of FGAs and SGAs in the treatment of positive symptoms is D2/3 striatal receptor blockade, with a threshold D2 receptor occupancy of 65% needed for response, and an increased risk of side effects with D2 occupancy greater than 80% (Howes et al. Reference Howes, Egerton, Allan, Mcguire, Stokes and Kapur2009). Combining two antipsychotics with high affinity for D2/3 blockade may result in an increased risk of side effects. The combination of a low potency antagonist like clozapine or quetiapine with a high affinity D2/3 blocker may be logical however, at this point, clear evidence is lacking (Goodwin et al. Reference Goodwin, Fleischhacker, Arango, Baumann, Davidson, de Hert, Falkai, Kapur, Leucht, Licht, Naber, O’keane, Papakostas, Vieta and Zohar2009). Clozapine has less affinity for D2 receptors, blocking ~30–50%, and has less propensity to cause EPSE’s (Kane Reference Kane2008), though the precise mechanism for its superior efficacy is not fully known (Stone et al. Reference Stone, Raffin, Morrison and Mcguire2010). A meta-analysis of 14 studies investigating clozapine augmentation with a second antipsychotic demonstrated a small reduction in symptom scores compared with augmentation with a placebo (Taylor et al. Reference Taylor, Smith, Gee and Nielsen2012), though there is little evidence to demonstrate that, excluding co-prescribing with clozapine, multiple antipsychotic prescribing is an effective treatment strategy (Kane et al. Reference Kane, Correll, Goff, Kirkpatrick, Marder, Vester-Blokland, Sun, Carson, Pikalov and Assuncao-Talbott2009). Despite this, multiple antipsychotic prescribing occurs in up to 15–20% of patients with schizophrenia (Edlinger et al. Reference Edlinger, Hausmann, Kemmler, Kurz, Kurzthaler, Walch, Walpoth and Fleischhacker2005; Patel et al. Reference Patel, Bishara, Jayakumar, Zalewska, Shiers, Crawford and Cooper2014).
The principal finding of this study was the reduction in the prescribing of multiple antipsychotics after a change in local prescribing guidelines. The rate of multiple antipsychotic prescribing among all in-patients prescribed an antipsychotic medication fell from 27% pre-intervention to 16% post-intervention compared with 43% in 2006 and 39% in 2007 in the Paton et al. study. In this study, mixed generation antipsychotics were prescribed in 64% of patients prescribed multiple antipsychotics pre-intervention compared with 42% post-intervention. Paton et al. reported rates of mixed generation antipsychotic prescribing of 31% of those prescribed antipsychotics pre-intervention and 29% post-intervention. The elimination of multiple FGA prescribing contributed to the decrease in prescribing of multiple antipsychotics. While multiple antipsychotic prescribing was reduced, there was a shift to prescribing multiple SGAs that increased from 6.8% of all antipsychotics prescribed in 2008 to 9.2% in 2012. This pattern mirrors a general shift from FGA to SGA prescribing (Donohue et al. Reference Donohue, O’malley, Horvitz-Lennon, Taub, Berndt and HUSKAMP2014).
Up to 10% of schizophrenic patients are prescribed high doses, despite lack of evidence and increased risk of side effects (Patel et al. Reference Patel, Bishara, Jayakumar, Zalewska, Shiers, Crawford and Cooper2014). In the Paton et al. study high dose prescribing was 36% in 2006 and 34% in 2007 (Paton et al. Reference Paton, Barnes, Cavanagh, Taylor and Lelliott2008). In this study the rate of high dose antipsychotic prescribing was 7% at baseline and 8% post-intervention. Although the proportion of patients prescribed high dose antipsychotics was significantly less in this study, the rate of high dose prescribing remained unchanged post-intervention. This may have implications for risk of side effects and, in conjunction with multiple antipsychotic prescribing, may result in delays to commencing evidence-based clozapine treatment (Howes et al. Reference Howes, Vergunst, Gee, Mcguire, Kapur and Taylor2012, Agid et al. Reference Agid, Schulze, Arenovich, Sajeev, Mcdonald, Foussias, Fervaha and Remington2013) in the one-third of patients that may be treatment resistant (NICE 2009).
This study demonstrated a marked reduction in the prescribing of multiple and high dose antipsychotics in patients diagnosed with psychotic disorders (F20–29), however reductions in multiple and high dose prescribing were not significant in patients diagnosed with mood disorders (F30–39). This is an area where a more targeted intervention could be directed.
While PRN antipsychotic medication prescribing rates were higher in Paton et al.’s study than the 15% rate in this study, the rate of prescribing did not change significantly in either study following intervention. It continues to be an important factor for non-adherence to antipsychotic prescribing guidelines and a more targeted intervention may yield significant benefits in future.
Overall, Paton et al. found little difference in adherence to guidelines following the introduction of a wide range of interventions. They proposed three reasons for interventions failing:
∙ The standards might not have been accepted by clinicians.
∙ The interventions may not have reached or been acceptable to clinicians.
∙ Cultural or organisational factors may have impeded change in each institution (Paton et al. Reference Paton, Barnes, Cavanagh, Taylor and Lelliott2008).
Tani et al. highlighted that passive interventions were less effective compared with active educational interventions that directly cautioned doctors on the use of combined and high dose antipsychotic prescribing (Tani et al. Reference Tani, Uchida, Suzuki, Fujii and Mimura2013). The success of the intervention in this study in effecting change may be accounted for by the following factors:
1. The policy was mentioned during compulsory induction and each new NCHD received a copy of policy at induction.
2. All departments agreed to implement the intervention.
3. The policy was widely accepted and seen to be beneficial by clinicians.
4. The Pharmacy Department were active in continuing to promote adherence to the policy via reminders at ward rounds and MDT meetings.
Limitations
Information regarding which doctors prescribed the antipsychotics was not gathered in this study. This prevents us from identifying any differences in prescribing practice among teams within the hospital before and after the intervention and also prevents us from drawing any conclusions as to whether the changes in prescribing practice found occurred in all teams or were concentrated within a subset of them.
It is possible that a concomitant policy change prohibiting PRN benzodiazepine prescribing by NCHDs may have caused a shift to PRN antipsychotic prescribing instead of PRN benzodiazepine prescribing. The significant change to benzodiazepine prescribing policy at the same time within the organisation acts as a confounder to making it impossible to state that the changes were solely due to the change in antipsychotic prescribing policy. Benzodiazepine and other psychotropic medication prescribing was not recorded and this limits the study’s ability to account for their impact on the results of the study.
Data regarding diagnosis was gathered and analysed using ICD 10 coding groups, rather than the precise diagnosis code, therefore limiting more detailed analysis of diagnosis and antipsychotic prescribing.
This study took place in a non-for-profit hospital in which change may be easier to implement than within the HSE due to differences in scale, unionisation and organisation. This may limit the generalisability of the study.
Conclusion
This audit showed that implementation of a new hospital-wide antipsychotic prescribing policy, which engages with all departments and clinicians through a concerted, targeted, ongoing educational campaign can positively impact on practice improving adherence to quality prescribing guidelines.
Recommendations
It would be beneficial to re-audit on an ongoing basis to ensure there is a sustained improvement in antipsychotic prescribing adherence to guidelines. The incorporation of an educational intervention into the NCHD induction programme to improve knowledge of and adherence to prescribing guidelines may help increase quality prescribing. We would also welcome additional centres to carry out similar research in order to examine the generalisability of these findings.
Acknowledgements
We acknowledge the contribution of Ms Margaret Nichol, Clinical Audit Facilitator, and St. Patrick’s University Hospital.
