Impact of risperidone long-acting injectable on hospitalisation and medication use in Canadian patients with schizophrenia

Summary

A retrospective chart review was used to assess hospitalisation and medication switching in patients with schizophrenia initiated on either intramuscular risperidone long-acting injectable (RLAI) (n=69) or oral atypical antipsychotics (n=93) in Canada.

In the RLAI-treated patients, compared with an identical period prior to RLAI use (40.8 months), there were significant decreases in hospitalisation (50.7 vs. 4.3%; p<0.0001) and duration of hospitalisation (23.5 vs. 0.3 days per patient; p<0.0001) when patients were switched to RLAI (mean treatment period 41.5 months). Compared with patients receiving oral atypicals for 57.2 months, RLAI patients had a reduced risk of hospitalisation (95% confidence interval 1.8–16.5% vs. 54.7–76.4%) and medication switching (95% confidence interval 34.6–58.4% vs. 55.7–76.4%).

By virtue of its periodic intramuscular administration, RLAI offers the efficacy and tolerability of an atypical medication without the compliance issues associated with oral drugs, and leads to significant decreases in hospitalisation.

Keywords:

• risperidone long-acting injectable
• schizophrenia-related hospitalisation
• oral atypical
• medication switching

Introduction

Therapeutic outcomes for patients with schizophrenia have improved to varying degrees with the introduction of oral atypical antipsychotic drugs such as risperidone and olanzapine1–3. These improvements include an impact on a broader range of symptoms as well as evidence of marginal increases in patient adherence to therapy as a consequence of the improved atypical safety profile4–6. Patient adherence to therapy in schizophrenia is one of the most significant barriers to successful control of disease symptoms4,7,8. The consequences of poor adherence are increased risk of patient relapse, re-hospitalisation and increased clinic and emergency room visits, all of which contribute significantly to the economic burden of schizophrenia6,9. In Canada, Goeree et al10 estimated that the direct healthcare and non-healthcare cost of schizophrenia in 2004 was $2.02 billion Canadian dollars (CAD), 61% of which was due to acute and non-acute hospital care, whilst the indirect cost through lost productivity was $4.83 billion CAD10.

However, adherence to therapy remains a common problem with oral atypical antipsychotics4. Dolder et al4 reported that after 6 months of treatment only 57% of outpatients with schizophrenia were adherent to oral atypical therapies. Although this was significantly higher than rates for typical agents (50%; p=0.05), 43% of patients were nevertheless non-adherent to oral atypical therapy and therefore at higher risk for relapse4,9. One way to improve adherence is through depot formulations of antipsychotics, where periodic intramuscular (im) injections of the drug replace daily oral therapy11. These formulations have been available for conventional antipsychotics for many years. Compared with oral administration, depot formulations of conventional antipsychotics reduce relapse and hospitalisation rates, particularly in patients who have difficulties with medication compliance11–13. Depot formulations also provide continuous delivery of medication, avoiding clinical problems associated with fluctuating serum drug levels due to first-pass metabolism of oral drugs14,15. Despite these advantages, depot formulations of conventional antipsychotics are associated with rates of adverse events similar to those observed with their oral counterparts, presenting a significant barrier to long-term patient adherence and control of schizophrenia symptoms11,12.

Risperidone long-acting injectable (RLAI) is the first atypical agent available in a depot formulation offering the improved safety and efficacy profile of oral risperidone as a depot therapy11,16. Clinical trials have demonstrated the efficacy and tolerability of RLAI administered at doses of 25–75 mg every 2 weeks5,17–21. Additional evidence suggests that RLAI therapy over 12 months may also reduce hospitalisation14,22,23. Considering the large economic burden of schizophrenia, any therapy that lowers the need for hospitalisation clearly has significant implications for healthcare providers24. However, the data available on these parameters were based on international clinical trials with a relatively short follow-up period. The present study assesses and compares the long-term risk of schizophrenia-related hospitalisation and medication switching in Canadian patients with schizophrenia and schizoaffective disorder initiated on RLAI or oral atypical agents.

Patients and methods

Study design

This was a multicentre, retrospective, chart audit of long-term follow-up data on Canadian patients with schizophrenia or schizoaffective disorder who originally participated in the RLAI clinical trials in Canada5,17. These trials had a duration of 12 weeks5 to 12 months17, and safety and efficacy data from these studies have been published5,17. At the conclusion of the trials, patients entered into an open-label follow-up study with the objective of assessing the long-term safety of 25-, 50- and 75-mg RLAI administered every 2 weeks. Subjects entered the follow-up study no more than 7 days after having completed the last assessment/endpoint visits of the clinical trials. Adverse events and concomitant medication use were recorded at each injection visit for the first 3 months, then every 3 months thereafter for the first year of follow-up, followed by every 6 months. Investigators were required to use their own discretion to withdraw patients from the follow-up study as a consequence of a serious adverse event or if the patient withdrew consent.

This study describes a retrospective chart review of these patients using a ‘mirror image’ methodology to compare hospitalisation before and after initiation of RLAI therapy. In a second component of the study, time to first hospitalisation and time to first medication switch for patients initiating therapy with RLAI were compared with values determined for similar patients with schizophrenia who were initiated on a new oral atypical medication over the same period. Ethical approval was obtained for the study at each participating site.

Patients

All 27 clinical sites participating in the RLAI clinical trials in Canada were invited to participate in the study. Eight sites agreeing to participate were asked to identify the charts of patients with schizophrenia or schizoaffective disorder who were enrolled in the pivotal RLAI clinical trials. Enrolment for these trials occurred between the 1st June 1999 and the 30th November 2000. Eligibility for chart review was defined by the inclusion criteria of the original RLAI clinical trials5,17. Patients in the clinical trials were at least 18 years of age with a diagnosis of schizophrenia or schizoaffective disorder according to the Diagnostic & Statistical Manual for Mental Disorders criteria25. Eligible patients entered a 2-week run-in period during which all antipsychotic medications were discontinued and all patients received oral risperidone 1–6 mg/day, reflecting standard medical practice when patients are switched to RLAI. Patients were then initiated on 25-, 50- or 75-mg RLAI im injection every 2 weeks, the actual depot dose depending on the dose of oral risperidone at the end of the run-in period5,17. Oral risperidone was discontinued 2–3 weeks after initiation of RLAI. Eligibility for chart audit in the present study was defined by participation in the clinical trials and RLAI therapy for at least 1 month. The latency period reflecting co-administration of oral risperidone with first injections of RLAI was, therefore, excluded from the study5,17. All RLAI-treated patients at each participating site were included in the analysis.

Participating sites were also asked to select charts of patients initiating a new oral atypical antipsychotic (i.e. patient was prescribed an atypical antipsychotic that had not previously been prescribed) between the 1st June 1999 and the 30th November 2000. Chart selection continued in a consecutive manner from the 1st June 1999 until the target number of subjects was reached at each site. Patient inclusion criteria were identical to those of the RLAI clinical trials5,17 and eligible subjects received the oral atypical for at least 1 month. Charts were excluded if the patient had a diagnosis of substance abuse or dependence within 3 months of the initiation of a new atypical therapy, had documented disease of the central nervous system, had an acute unstable and/or significant or untreated somatic illness or had a clinically significant electrocardiographic abnormality. There were no criteria defining concomitant medication use in selected patients.

Where appropriate, all patients, or their legal representatives, provided written informed consent for their charts to be reviewed.

Data collection

Data were collected by review of each selected patient chart using a specifically designed case report form. For RLAI patients, charts were reviewed for similar periods before and after initiation of RLAI therapy. The actual chart review duration was defined by the RLAI treatment period (i.e. from initiation of RLAI to therapy stop date). In the event that the patient was still receiving RLAI at the time of the chart audit, the RLAI treatment period was defined by the date of the chart review. Charts were retrospectively reviewed for an identical period prior to the RLAI index date. For patients lacking medical records over the complete pre-RLAI period, available data and duration of chart review were both recorded. For those patients who received RLAI for at least 1 month but less than 1 year, the pre-RLAI chart review period was fixed at 12 months to simplify data collection. Data collected included psychiatric-related hospitalisations, date of first medication switch in the post-RLAI treatment period and use of anxiolytics, anticholinergics and sedatives prior to and during RLAI therapy. Additional demographic data were obtained by review of the clinical trial data.

For the oral atypical antipsychotic patient group, demographic data were collected in addition to information on oral atypical medication, time to first hospitalisation and antipsychotic medication switch. In this case, chart review continued from initiation of the new oral atypical medication to the first medication switch. In the event that patients did not switch medication over the assessment period, the date of chart review defined the duration of the chart audit.

Statistical analysis

For the target sample size assumptions, it was estimated that to detect a pre-post RLAI difference of 0.3 hospitalisations per patient, over similar assessment periods, with 80% power using a paired two-sided t-test with a 5% significance level and assuming a within-patient standard deviation (sd) of 1.0 hospitalisation per patient for the pre-post difference, would require a sample of 88 patients. A similar sample size was required to detect with 80% power a pre-post RLAI difference of 6 days in duration of hospitalisation, over similar assessment periods, using a paired two-sided t-test with a 5% significance and assuming a within-patient sd of 20 days in the pre-post difference.

At each participating site, the number of oral atypical charts reviewed was targeted to be 50% larger than the number of RLAI eligible patient charts. Since the former group may be receiving any one of a number of different atypical drugs, with a potential variation in patient outcomes, the 50% increase in patients ensured a more accurate determination of relevant comparative variables.

Descriptive statistical summaries were reported for demographic and response variables. For the pre-post RLAI assessments, the number of hospitalisations and duration of hospitalisation were compared using the within-subject differences normalised by the observation period. In the case of patients who were hospitalised at the time of the switch to RLAI, hospitalisation was assigned to the pre-RLAI period. The analysis also investigated possible effects of subject age, gender and time since diagnosis on these pre- and post-RLAI comparisons.

The Kaplan–Meier estimator26 was used to compare time to first psychiatric-related hospitalisation or medication switch between the RLAI- and oral atypical-treated patients. Significance values for these comparisons were based on the log-rank test27. The proportion of patients hospitalised or with a medication switch in both treatment groups was obtained from the Kaplan–Meier survival estimators. These proportions at different time points were expressed as the 95% confidence interval (CI). The Cox proportional hazard model28 was used to investigate the effect of age, gender and time since diagnosis on the time to hospitalisation or medication switch. RLAI pre-post comparisons were based on the paired t-test for continuous variables and McNemar's test for discrete values29. RLAI/atypical comparisons were based on the t-test for independent samples and the Fisher's Exact Test for continuous and discrete variables, respectively. Statistical analyses were carried out using SAS PC Release 8.2.

Results

Eight of the RLAI clinical trial sites contributed data on 69 RLAI and 93 oral atypical antipsychotic patients. The majority of patients were from four sites in Quebec (83 patients; 51%) with additional patients from Alberta (48 patients; 30%), Ontario (23 patients; 14%) and Saskatchewan (8 patients; 5%). As shown in Table 1, there were no significant differences between the RLAI and oral atypical subjects in age at diagnosis of schizophrenia or in gender distribution. However, the RLAI-treated patients were marginally, but not significantly, older than those receiving oral atypicals, with mean ages of 45.7 and 41.4 years, respectively (p=0.052). At the time of the chart review, RLAI-treated patients had received depot therapy for a mean period of 41.5 months, 55.1% were still receiving therapy and 24.6% had received treatment for <12 months. The RLAI doses at the initiation of the last trial in which patients participated were 25 mg (18.8% of patients), 50 mg (31.9%) and 75 mg (49.3%) administered every 2 weeks.

Table 1. Baseline characteristics of patients included in the analysis

Characteristic	RLAI patients (n=69)	Oral atypical patients (n=93)	p-value
Age at time of chart audit (years)
Mean (sd)	45.7 (11.6)	41.4 (14.82)	0.052
Median	45.3	40.8
Range	24.6–75.4	18.29–88.96
Age at diagnosis of schizophrenia (years)
Mean (sd)	27.8 (9.37)	26.5 (9.03)	0.4015
Median	26.0	24.0
Range	14.0–58.0	15.0–61.0
Gender, n (%)
Women	21 (30.4)	26 (28.0)	0.7303
Men	48 (69.6)	67 (72.0)

RLAI, risperidone long-acting injectable; sd, standard deviation.

To assess the impact of RLAI therapy on hospitalisation, patient charts were reviewed for similar periods prior to and post initiation of RLAI therapy. As shown in Table 2, these periods were identical at 40.8 and 41.5 months for the pre- and post-assessments, respectively (p=0.5841). Prior to RLAI, 50.7% of subjects had been hospitalised and 21.7% had more than one hospitalisation during the review period. Following initiation of RLAI, only 4.3% of patients were hospitalised, an 12-fold decrease, and no patients were hospitalised multiple times. Furthermore, with RLAI there was a highly significant decrease in the number of hospitalisations per patient (0.81 hospitalisations per patient pre RLAI vs. 0.04 during treatment; p<0.0001), a 99% decrease in total duration of hospitalisation (1,620 days vs. 23 days; p<0.0001) and duration of hospitalisation per patient (23.5 vs. 0.3, p<0.0001; 46.3 vs. 7.7 days per patient hospitalised). Two (2.9%) of the 69 patients initiated RLAI as inpatients and were discharged 3 and 7 days, respectively, following the initiation of depot therapy. Both hospitalisations were assigned to the pre-RLAI period. Prior to RLAI, 75.4% and 53.6% of patients were taking anticholinergics and anxiolytics, respectively, whilst after RLAI this decreased to 58.0% (p=0.0005) and 33.3% (p=0.0015). However, sedative use was not significantly different over the two assessment periods.

Table 2. Comparison of pre- and post-RLAI treatment periods.

Outcome	Pre-RLAI (n=69)	RLAI treatment (n=69)	p-value*
Duration of chart review (months)
Mean (sd)	40.8 (23.16)	41.5 (25.58)	0.5841
Median	54.2	56.8
Range	5.4–70.0	0.9–69.7
Patients hospitalised, n (%)	35 (50.7%)	3 (4.3%)	<0.0001
Patients with >1 hospitalisation, n (%)	15 (21.7%)	0 (0.0%)
No. of hospitalisations/patient^†
Mean (sd)	0.81 (1.05)	0.04 (0.21)	<0.0001
Range	0–5	0–1
^‡	1620.0	23.0	<0.0001
Duration of hospitalisation (days/patient)^†
Mean (sd)	23.5 (42.65)	0.3 (1.95)	<0.0001
Range	0–279	0–15

RLAI, risperidone long-acting injectable; sd, standard deviation.

* p-values based on paired t-test for continuous variables and McNemar's test for proportions.

^† Mean value based on all patients.

^‡ Total number of days of hospitalisation recorded for all patients over each assessment period.

For a wider assessment of the impact of RLAI, outcomes with this depot formulation were compared with those of patients initiated on a new oral atypical antipsychotic (Table 1). In the 1-month period prior to the initiation of the new oral atypical, 54 (58.1%) of the recruited subjects were receiving an antipsychotic, and for 23 (43%) of these patients this was a depot antipsychotic. The major new oral atypicals prescribed were olanzapine (41.9%), risperidone (43%) and quetiapine (8.6%), with mean daily doses of 12.2, 3.4 and 643.8 mg/day, respectively; the remaining subjects (6.5% of total) were prescribed other oral atypicals. As shown in Table 3, data on oral atypical patients were obtained over a mean chart review period of 57.2 months compared with 41.5 months in the RLAI-treated group (p<0.0001). Time to first hospitalisation, assessed only in terms of those patients actually hospitalised, was not significantly different for the two patient groups (13.8 vs. 17.8 months (p=0.7187) for RLAI and oral atypical patients, respectively), a result that was maintained after adjustment for the marginal age difference at baseline (Table 1). However, the overall time to first hospitalisation was significantly lower (p<0.0001) for the oral atypical patients, as can be seen in the Kaplan–Meier plot shown in Figure 1. The 95% CI for the risk of hospitalisation for oral atypical patients obtained from the Kaplan–Meier survival product estimates was 54.7–76.4% compared with 1.8–16.5% for RLAI-treated patients. The Cox proportional hazard analysis based on all patients indicated that treatment group and patient age were significantly associated (p=0.0001 and p=0.0018 by χ2, respectively) with time to first hospitalisation.

Figure 1. Kaplan–Meier survival analysis of time to first hospitalisation for patients receiving RLAI or a new oral atypical medication (p<0.0001 based on log-rank test).

Table 3. Hospitalisation and medication switches in RLAI- and oral atypical-treated patients

Outcome	RLAI patients (n=69)	Oral atypical patients (n=93)	p-value
Duration of chart review period (months)*
Mean (sd)	41.5 (25.58)	57.2 (12.46)	<0.0001
Median	56.8	59.7
Range	0.9–69.7	4.8–70.7
Risk of hospitalisation (95% CI)^†
6 months	0.2–9.8	15.5–32.5
1 year	0.2–9.8	27.7–47.2
2 years	1.8–16.5	32.8–52.7
3 years	1.8–16.5	54.7–76.4
Overall	1.8–16.5	54.7–76.4
Time to first hospitalisation (months)^‡
Mean (sd)	13.8 (11.9)	17.8 (18.82)	0.7187
Median	17.9	10.5
Range	0.4–23.2	0.0–64.7
Patients with an antipsychotic medication switch (95% CI)^$
6 months	10.3–28.6	9.2–24.1
1 year	16.1–36.6	21.9–40.5
2 years	22.3–44.3	34.7–54.8
3 years	28.8–51.7	41.9–62.1
Overall	34.6–58.4	55.7–76.4	0.0349
Time to first medication switch (months)^¶
Mean (sd)	16.5 (16.98)	18.9 (16.66)	0.5249
Median	8.3	14.1
Range	0.9–60.4	0.9–64.0

CI, confidence interval; RLAI, risperidone long-acting injectable; sd, standard deviation.

* From the initiation of RLAI or oral atypical therapy.

^† Values based on the Kaplan–Meier product estimate for the proportion of patients without hospitalisations. These estimates take into account differences in the exposure time of the patients in both treatment groups.

^‡ Based on patients hospitalised.$ Values based on the Kaplan–Meier product estimate for the proportion of patients with antipsychotic medication switch.

^¶ Based on patients who switched medication.

Patients taking RLAI also had a significantly lower risk of switching medication than those on an oral atypical (p=0.0349) (Figure 2; Table 3). However, in the two patient groups time to first medication switch was not significantly different for those patients who switched medication (16.5 and 18.9 months in RLAI- and oral atypical-treated patients, respectively; p=0.5249). From the Kaplan–Meier analysis (Figure 2), median time to first medication switch was somewhat greater than 60 months for RLAI patients compared with 35.2 months for oral atypical patients. At 3 years post initiation of an oral atypical, 50.5% (95% CI 41.9–62.1%) of patients had switched medication compared with 39.1% (95% CI 28.8–51.7%) for RLAI patients. In a covariate analysis, the most significant parameters influencing time to first medication switch were patient age (p=0.054) and gender, with younger male patients having a higher risk of switching.

Figure 2. Kaplan–Meier survival analysis of time to first medication switch for patients receving RLAI or a new oral atypical medication (p=0.0349 based on log-rank test).

The benefits of RLAI were further emphasised in a Kaplan–Meier survival analysis of combined endpoints of time to medication switch or time to first hospitalisation. Time to this combined endpoint occurred significantly faster (p<0.0001 by log-rank test) for patients receiving an oral atypical (results not shown). In the latter group, 50% of patients had reached this endpoint at approximately 15 months compared with 60 months for the RLAI patients, amounting to a four-fold delay in reaching the combined endpoint with RLAI.

Discussion

The safety and efficacy of RLAI in the treatment of schizophrenia have been confirmed in several clinical trials5,17,18. The present study demonstrates that for Canadian patients who initiated RLAI during these studies, there were dramatic changes in hospitalisation over a long-term follow-up period. Over a period of 41.5 months following initiation of RLAI, only 4.3% of patients were hospitalised with a mean duration of 7.7 days per hospitalised patient. In contrast, in the same patients over an identical period prior to initiation of RLAI, 50.7% were hospitalised for a period of 46.3 days per hospitalised patient, reflecting the established pattern of extensive hospitalisation associated with standard oral antipsychotic therapy in schizophrenia30,31. These Canadian data confirm previous studies demonstrating a similar impact of RLAI on schizophrenia patient hospitalisation14,22,23,32. In an assessment of direct medical resource utilisation during a 12-month RLAI clinical trial in several countries, including Canada, patients requiring hospitalisation decreased from 38% over a 12-week period prior to the initiation of RLAI therapy to 12% during the last 12 weeks of the treatment period, a 68% reduction (p<0.0001)14,23. In addition, for patients hospitalised at baseline (24% of assessed patients), 71% were discharged from hospital at some point during the clinical trial14,23.

However, the present data suggest that the impact of RLAI on hospitalisation in Canadian patients was more dramatic than that observed in previous studies14,22,23. Although Leal et al23 reported decreased hospitalisation with RLAI, subjects were still hospitalised for 30.5 days per year, a mean value that included 7% of patients who remained in hospital for the duration of the study. In contrast, in the present study, after RLAI treatment the mean hospitalisation rate was 0.3 days per patient over the total chart review period (3.5 years), based on all patients, or 7.7 days per hospitalised patient. These differences may reflect varied country-specific treatment patterns33, different patient populations or varying definitions of study-related hospitalisation. In the present study, only psychiatric-related hospitalisation was assessed and 97% of subjects were outpatients when initiated on RLAI (compared with 76% in the study by Leal et al23). In addition, Leal et al23 only assessed outcomes in patients receiving a modal RLAI dose of 25 or 50 mg every 2 weeks, whereas in the current study patients received 25, 50 and 75 mg. Of the patients selected for chart review 49% were receiving 75-mg RLAI, reflecting dose instructions in the initial Canadian clinical trials, as defined in the clinical trial protocol, and for the majority of patients this dose was unchanged over the course of the study. However, 75-mg RLAI is not available as a marketed dose strength as it was not shown to produce incremental benefits over the 25- or 50-mg dose in clinical trials5.

Varying data on the impact of RLAI on patient outcomes and health resource utilisation have recently been published34–36. In a UK mirror-image study, Young and Taylor35 showed that schizophrenia patients with severe disease who were switched to RLAI had a high rate of discontinuation at 12 months (67.6%) and healthcare resource utilisation increased significantly compared with the pre-RLAI period. In contrast, in a UK patient population with a similar clinical profile to that defined by Young and Taylor35, Niaz and Haddad32 demonstrated net financial savings after a switch to RLAI and suggested that the different outcomes may be due to variations in RLAI dosing, duration of RLAI therapy, method of switching to RLAI and general study methodology differences. These variable data suggest that schizophrenia is a complex disorder and underscores the need for an individualised approach to pharmacotherapy.

There have been no head-to-head studies of RLAI and oral atypicals in matched patient populations. However, chart review of similar patients initiated on an oral atypical antipsychotic at the same time as the RLAI patients did provide a basis for direct comparison of outcomes. Compared with an oral atypical antipsychotic, RLAI treatment led to a decreased need for hospitalisation (95% CI 1.8–16.5% for RLAI vs. 54.7–76.4% for oral atypical patients), fewer medication switches (95% CI 34.6–58.4% vs. 55.7–76.4%) and significant decreases in the time to first medication switch (p=0.0349) and hospitalisation (p<0.0001). The significant decrease in the number of RLAI patients switching medication suggests a considerable improvement in patient adherence to therapy, reflecting the improved efficacy and tolerability of RLAI compared with oral atypical drugs. Literature values on rates of medication switching or discontinuation for conventional and oral atypical medications in schizophrenia are highly variable4. A recent study by Lieberman et al37 determined that 74% of patients initiated on an oral atypical or a conventional antipsychotic discontinued medication before 18 months and that rates were similar for both drug classes. In the present study, approximately 40% of oral atypical patients had switched medication at 18 months and overall median time to this endpoint was 35.2 months, dramatically different from the results described by Lieberman et al37. However, the latter study, based on randomised clinical trials, excluded patients with schizoaffective disorder and the primary outcome was discontinuation of treatment for any cause, including the patient's decision to stop treatment independently. The present study assessed only the rate and timing of documented medication switching in a population of patients receiving one of three oral atypicals in a real-world clinical setting and this may account for the study differences.

Since up to 79% of the direct costs of schizophrenia result from institutional care, any decrease in hospitalisation would have a large economic impact38. Although the current study was not designed to measure the cost impact specifically, based on a per diem cost of $624 for psychiatric-related hospitalisation39, the estimated annual cost savings achieved through the switch to RLAI in hospitalisation alone (excluding drug costs) would amount to $4,249 per patient. There would also be additional savings through decreased anticholinergic and anxiolytic use after the switch to RLAI, although this would be minor compared with hospitalisation since medication represents only approximately 1–6% of the cost of schizophrenia23,38,40. These savings in hospitalisation with a switch to RLAI ($4,249) are virtually identical to the predicted savings generated in a recent Canadian study where a discrete event simulation model estimated annual savings in hospitalisation alone (excluding long-term hospitalisation) with RLAI of $4,541 (CAD, 2003) per patient41. Using this model, Chue et al41 estimated that annual overall discounted savings with RLAI compared with oral risperidone were $2,626 per patient when all relevant direct medical resources were included in the simulation.

There are a number of limitations to the present study. The RLAI patients selected for chart review participated in clinical trials in Canada and at least part of their treatment was delivered in the context of a structured clinical trial, although the long-term follow-up component was less protocol driven. Nevertheless, outcomes for these patients were compared with those assessed before initiation of RLAI and with outcomes observed in patients initiating an oral atypical. Both comparisons were, therefore, based on usual clinical care scenarios and this may have biased the study to more favourable outcomes with RLAI. However, the RLAI patient population in Canada provided a valuable source of long-term hospitalisation/switch data, despite the potential limitations. In addition, not all participating sites in the risperidone clinical trials (n=27) agreed to participate in the study and the patients selected for chart review may not reflect outcomes for the total RLAI-treated population. The reasons provided for non-participation included inaccessibility to medical charts, concerns about recruiting suitable oral atypical patients and inability to comply with study timelines. There may also have been some selection bias in the subjects recruited for the new oral atypical chart review at each site, and the chart review period was significantly longer than that of the RLAI patients. However, regarding the latter, the Kaplan–Meier analysis methods used to compare outcomes adjusted for this variation in assessment periods.

Conclusions

By virtue of its periodic im administration, RLAI offers the efficacy and tolerability profile of an atypical medication without the serious compliance issues associated with an oral drug. The significant impact on hospitalisation and long-term patient adherence to treatment suggest that in a Canadian healthcare setting RLAI, apart from the clear clinical benefits, also offers the potential for substantial cost savings in the care of patients with schizophrenia.

Acknowledgements

The authors would like to thank the additional investigators who contributed patient data for this study: Stephen Boucher (Calgary, AB); Angelo Fallu (Sherbrooke, QC); Javad Moami (Hull, QC); Jean-Pierre Mottard (Montreal, QC); MS Renuka-Prasad (Saskatoon, SK); and Brian Ticoll (Markham, ON).

This study was supported by a grant from Janssen Ortho Canada Inc.

Linda Beauclair currently receives grant support from Janssen-Ortho, Eli Lilly, Pfizer and Organon. Pierre Chue receives grant support from Janssen Ortho, Eli Lilly, AstraZeneca, Pfizer, GlaxoSmithKline, Wyeth-Ayerst, Boehringer Ingelheim, Lundbeck and Novartis. Fernando Camacho is a consultant for Janssen Ortho Canada Inc. McKesson Canada was contracted by Janssen Ortho Canada Inc. to design and administer the chart audit study and to prepare the manuscript.

Notes