Long-term outcomes in patients with schizophrenia treated with risperidone long-acting injection or oral antipsychotics in Spain: Results from the electronic Schizophrenia Treatment Adherence Registry (e-STAR)

Methods Parameters were assessed at baseline and at 3 month intervals for 2 years in patients initiated on risperidone long-acting injection (RLAI) ( n = 1345) or a new oral antipsychotic (AP) ( n = 277; 35.7% and 36.5% on risperidone and olanzapine, respectively) in Spain. Hospitalization prior to therapy was assessed by a retrospective chart review. Results At 24 months, treatment retention (81.8% for RLAI versus 63.4% for oral APs, p < 0.0001) and reduction in Clinical Global Impression Severity scores (−1.14 for RLAI versus −0.94 for APs, p = 0.0165) were significantly higher with RLAI. Compared to the pre-switch period, RLAI patients had greater reductions in the number (reduction of 0.37 stays per patient versus 0.2, p < 0.05) and days (18.74 versus 13.02, p < 0.01) of hospitalizations at 24 months than oral AP patients. Conclusions This 2 year, prospective, observational study showed that, compared to oral antipsychotics, RLAI was associated with better treatment retention, greater improvement in clinical symptoms and functioning, and greater reduction in hospital stays and days in hospital in patients with schizophrenia. Improved treatment adherence, increased efficacy and reduced hospitalization with RLAI offer the opportunity of substantial therapeutic improvement in schizophrenia. Keywords Risperidone long-acting injection Antipsychotic medication Schizophrenia Treatment retention Observational study 1 Introduction Of significant clinical and economic importance in schizophrenia is the occurrence of multiple episodes of relapse over the course of the disease [9,34] . Nonadherence to antipsychotic therapy is a common reason for relapse and rehospitalization of patients with schizophrenia and contributes to the high cost of treating the illness and to the poor prognosis for nonadherent patients [1,9] . The identification of factors associated with noncompliance is therefore a priority in psychiatry for the development of strategies aimed at improving compliance levels [1] . A number of studies have associated low patient insight with poor medication compliance [19,20] . Recently Acosta et al. [1] have demonstrated that poorer insight and increased conceptual disorganization were independently associated with noncompliance in schizophrenia patients and the authors suggest that interventions aimed at improving insight and cognitive deficits in patients, as well as using long-acting or depot injectable antipsychotics, may lead to significant increases in compliance. Patient adherence to therapy is particularly poor with oral medications [33,35] . However, a major advance in the treatment of schizophrenia was heralded in the 1960s with the introduction of conventional depot antipsychotics. When compared to conventional oral medications, depot formulations offered improved compliance, better control over dose adjustments, more predictable and consistent plasma drug concentrations and lower rates of patient relapse [2,10,17] . Nevertheless, when compared to their oral counterparts, depot formulations of conventional antipsychotics also exhibit a similar adverse event profile, which presents a significant barrier to long-term compliance and control of schizophrenia symptoms [2,4] . Until recently long-acting formulations of antipsychotics in the atypical class were unavailable to clinicians. The need for a single agent in which the clinical advantages of a long-acting formulation are combined with the efficacy and safety of an atypical antipsychotic led to the development of the first long-acting injectable atypical antipsychotic, risperidone long-acting injection (RLAI). Clinical trials have demonstrated the efficacy and tolerability of RLAI [7,12,18,31] . In addition, studies assessing outcomes prior to and post an antipsychotic medication switch show that a switch from conventional oral antipsychotics to RLAI can lead to significant reductions in hospitalization, a possible consequence of improved adherence to therapy and lower relapse rates with RLAI [3,8] . However, the majority of data published on atypical antipsychotics are derived from randomized controlled trials (RCTs). By virtue of its rigid design features, the RCT provides the highest standard of evidence to show variation in the efficacy of different interventions. However, the short duration of the majority of RCTs, and the very defined patient inclusion criteria, suggest that such studies do not reflect either the duration of the disease in real clinical practice or patients with schizophrenia in the community [16,28,30] . One way to address the limitations of RCTs is through complementary large naturalistic observational studies [5] . The advantages of such studies in schizophrenia include the ability to evaluate larger numbers of patients over longer periods of time in real world clinical practice settings, with less restrictive inclusion and exclusion criteria. In fact the National Institute for Health and Clinical Excellence (NICE) in the United Kingdom has called for the development of well-designed observational studies in schizophrenia that reflect normal clinical practice [25] . In response to this call, the electronic Schizophrenia Treatment Adherence Registry (e-STAR) was initiated. e-STAR is an international, prospective, observational study which aims to establish a large database to evaluate long-term clinical and economic outcomes in patients with schizophrenia who have commenced a new oral antipsychotic or RLAI, as part of their routine continuing clinical management. Spain is the first country to complete the e-STAR study and the 24 month results from this country are presented here. 2 Methods 2.1 Study design e-STAR is a large-scale, international, prospective, observational study of clinical and economic outcomes in patients with schizophrenia. The primary objective of the study is to evaluate drug use patterns and the effectiveness of antipsychotics as used in daily practice from a representative sample of patients seeking treatment for schizophrenia. Data collection in e-STAR is accomplished mainly through a secure web based system, although a traditional paper based system is also available to investigators, allowing the inclusion of a broad range of centers and participants [27] . In addition, in order to capture real-life treatment of patients with schizophrenia, e-STAR has no formalized diagnostic procedures, randomization or treatment decisions and physicians are encouraged to base treatment decisions on their best medical judgement. The present paper describes results from the e-STAR study in Spain which is the first country to complete the study and also has two treatment arms, RLAI and oral antipsychotics. 2.2 Study population Participating physicians (150 physicians with a primary specialty in psychiatry from all 17 autonomous communities in the country) throughout Spain enrolled patients in e-STAR after they initiated treatment with, or changed antipsychotic medication to, RLAI or an oral antipsychotic (conventional or atypical oral therapy). Each physician was asked to include data from at least four patients and in order to balance the proportion of patients receiving RLAI or oral antipsychotic treatment, for every four patients initiated on the former, the next patient starting on any oral antipsychotic was enrolled in e-STAR. Any inpatient or outpatient prescribed RLAI or any oral antipsychotic was eligible for inclusion. The patient or the patient's authorized legal representative provided written informed consent. The study protocol was approved by the ethics committee of Complejo Hospitalario Xeral Cies, Vigo (Spain). To capture information from a normal clinical practice setting and to ensure that the study design did not interfere with the normal course of health care, the clinical management of patients was determined solely by the treating physician. After a physician and a patient (or the patient's representative) had reached an agreement about the need to start treatment with a new antipsychotic medication, physicians had the opportunity to document the patient's clinical history and progress in the e-STAR database. During the study special mechanisms were instituted to maintain patient confidentiality during data collection and transfer to the e-STAR registry [27] . 2.3 Data collection Data for each patient enrolled in e-STAR were recorded at baseline and at site visits every 3 months up to a total of 24 months following initiation of RLAI or oral antipsychotic therapy. Data collected at baseline included patient demographics, principal diagnosis and duration of illness, type and dose of new treatment initiated, Clinical Global Impression Severity (CGI-S) [14] and Global Assessment of Functioning (GAF) [11] scores, and reason for initiating new treatment. Data on hospitalization (number of hospitalizations and duration of each stay) over the 12 months prior to the baseline assessment were collected by a retrospective chart review. The use of concomitant medications was also recorded. The latter included an assessment of the use of antipsychotics, anticholinergics, antidepressants, anxiolytics and mood stabilizers. Prospective outcomes, measured every 3 months, included CGI-S and GAF scores, hospitalization, concomitant medication utilization and treatment discontinuation. Treatment discontinuation and patient drop outs, with the associated reasons for drop out, were also recorded over the course of the study. Both the 12 month retrospective and the 24 month prospective assessments were defined by the date of initiation of therapy with RLAI or an oral antipsychotic (index date). The study design is summarized in Fig. 1 . 2.4 Statistical analysis Kaplan–Meier survival curves were used to estimate time to all cause treatment discontinuation for both patient groups. Cox proportional hazards models were used to compare time to discontinuation between the groups and to identify baseline clinical and demographic parameters correlated to treatment discontinuation. All models were fitted using a stepwise selection procedure with inclusion and exclusion significance levels of 5%. CGI-S and GAF score changes over time were analyzed using a full mixed model controlling for patient differences in baseline demographic and clinical characteristics and also taking differences in treatment discontinuation and drop outs between the two treatment groups into consideration. Hospitalization patterns, defined in terms of the percent of patients hospitalized, number and days of hospitalizations per patient, were compared over equal periods prior to and post initiation of the new antipsychotic and expressed as the change in each parameter between the retrospective and prospective periods. A full mixed model was fitted to the hospitalization parameters to identify clinical and demographic characteristics impacting changes after initiating RLAI or oral antipsychotics, as well as correcting for differences in treatment discontinuation and drop outs between the two treatment groups over time. In this analysis an autoregressive correlation structure was fitted to the within-patient residuals. For inpatients at the time of initiation of the new antipsychotic, the baseline hospitalization was included in the retrospective assessment, as the admission was due to failure of the previous treatment; all subsequent hospitalizations were included in the prospective assessment. 3 Results 3.1 Baseline demographics Of the 1622 patients enrolled in e-STAR from Spain, 1345 (83%) were prescribed RLAI at study initiation while the remaining patients ( n = 277) received oral antipsychotics. The baseline characteristics of both treatment groups are shown in Table 1 . Patients receiving RLAI or oral antipsychotics had similar mean ages (38.4 and 37.0 years for RLAI and oral antipsychotic patients, respectively, p = 0.059), were mostly male (63% male) and in both groups the majority of participants were outpatients at initiation of therapy (8.8% and 6.5% of the participants were inpatients in the RLAI and oral antipsychotic groups, respectively, p = 0.24). The two patient cohorts did significantly differ, however, with respect to the duration of illness and type of diagnosis. Patients receiving RLAI had a longer duration of illness than those on oral agents (12.6 years ± 9.5 versus 10.9 years ± 9.7, respectively, p = 0.008), and a greater proportion had a diagnosis of schizophrenia (85.1% versus 79.4% for the oral antipsychotic treated group, p = 0.024). Although there was a tendency toward greater severity of illness in the RLAI cohort (mean CGI-S scores were 4.62 ± 0.92 for RLAI versus 4.51 ± 0.91 for oral patients), this difference was not statistically significant ( p = 0.088). The majority of patients had moderate to marked disease when the CGI-S scores were categorized by severity ( Table 2 ). Mean GAF scores at baseline were 46.9 and 48.4 for the RLAI and oral antipsychotic treatment groups, respectively ( p = 0.29). 3.2 Antipsychotic medication at baseline There were significant differences ( p < 0.001) in the way the two patient groups were treated prior to the medication switch ( Table 1 ). While most patients had been treated with oral atypicals, a higher percentage of patients in the RLAI group received conventional depot antipsychotics (8.5% versus 2.9%) and combination therapy with either oral antipsychotics (11.7% versus 4.3%) or oral and conventional depot antipsychotics (9.7% versus 4.0%). Oral risperidone was utilized more frequently in the RLAI group (32.9% versus 14.1% of the oral antipsychotic patients). In addition, 11.9% of the RLAI treated patients had not received prior antipsychotic therapy compared to 19.9% of the oral atypical group. 3.3 Reasons for initiating treatment Significant differences were observed between the RLAI and oral antipsychotic treated groups in the reasons for initiating therapy ( p < 0.0001). In both groups ( Table 1 ), the most common reason was insufficient response to previous therapy (32.3% versus 42.6% of patients, respectively). For a notably higher proportion of patients in the RLAI treated group (30.6% versus 13.4% in the oral antipsychotic group), compliance with the previous antipsychotic was cited as the major factor in initiating new therapy. Conversely, in comparison to the RLAI treated patients, a greater proportion of the oral antipsychotic patients (29.2% versus 15.9%) indicated unacceptable tolerability as the major reason for initiating a new therapy. 3.4 RLAI dosing The mean dose of RLAI at initiation of therapy was 37.2 mg intramuscularly (im) every 2 weeks with 43%, 26% and 28% of the patients on 25 mg, 37.5 mg and 50 mg, respectively. The remaining patients (3%) received 75–100 mg doses. Mean administration dose increased over the course of the study, reaching 42.9 mg at 6 months, 45.1 mg at 12 months and 47.6 mg at 24 months. 3.5 Oral antipsychotic therapy At baseline the oral antipsychotic group was prescribed oral atypicals (94.2% of patients) either as monotherapy (66.1%) or with conventional oral (24.9%) or depot antipsychotics (3.3%); 5.4% of the patients were prescribed monotherapy with oral conventional antipsychotics. The most common oral atypicals utilized were risperidone (35.7%) and olanzapine (36.5%). 3.6 Treatment retention rates As shown in Fig. 2 , the Kaplan–Meier analysis revealed that after 24 months, 81.8% of RLAI treated patients remained on initial therapy (RLAI), compared to only 63.4% of patients taking oral antipsychotics ( p < 0.001). A proportional hazards model showed that patients initiated on RLAI were twice as likely to remain on initial therapy after 2 years (adjusted hazards ratio = 2.3; 95% confidence interval (CI) = 1.79–2.97; p < 0.0001). The proportional hazards model also showed a positive link between the baseline GAF score and treatment retention ( p = 0.0005), indicating that patients with high GAF scores had longer treatment retention time. Other factors which tended to increase the retention time included age ( p = 0.0135) and self-injury ( p = 0.0204) while previous use of conventional depot medications ( p = 0.0365) and suicidal ideation ( p = 0.0751) decreased treatment retention. Both the type of antipsychotic used prior to the switch, and the first reason for the switch, had no impact on retention times in either the RLAI or oral atypical patient groups. 3.7 Use of non-antipsychotic concomitant medications Changes in the use of concomitant medications within each treatment group were evaluated over the 2 year study period. Two major categories were considered: use of non-antipsychotic concomitant medications in general and anticholinergics in particular. In both treatment groups there was a significant reduction in the use of concomitant non-antipsychotic medication. In the RLAI treated group, the proportion of patients taking non-antipsychotic concomitant medications decreased from 74.4% at baseline to 54.4% by study end (a reduction of 20%) compared to a 12.8% reduction in the oral antipsychotic group (64.8% of patients at baseline to 52.0%). As shown in Fig. 3 , anticholinergic medication use decreased from 33.8% of patients at baseline to 14.4% by study end (a reduction of 19.4%, p < 0.001) in RLAI patients. This reduction was greater than that of the oral antipsychotic treated group where the decrease in anticholinergic medication use was 10.4% (from 20.0% at baseline to 9.6% at study end). There was no significant change in either group in the utilization of mood stabilizers over the study period, which was 15% and 14.4% ( p = 0.72) at baseline and 2 years, respectively, in the RLAI group and 15.2% and 16.8% ( p = 0.59) in the oral antipsychotic group. 3.8 Efficacy Over the 2 year study period both treatment groups demonstrated statistically significant improvements in overall clinical severity, as measured by the CGI-S ( p < 0.0001) ( Table 2 ). In the RLAI treated cohort, the mean CGI-S score decreased from 4.62 ± 0.92 to 3.48 ± 1.00, an improvement of 1.14 points, while the group receiving oral antipsychotics had a reduction of 0.94 from 4.51 ± 0.91 to 3.56 ± 0.87 ( Table 2 ). A full mixed model fitted to the CGI-S scores revealed that the 20% greater reduction in mean CGI-S score with RLAI compared to oral antipsychotics was statistically significant ( p = 0.0165). Other factors with a positive impact on the CGI-S score changes included patient age ( p = 0.0016), time since diagnosis ( p < 0.0001), patient employment ( p < 0.0001) and GAF score ( p < 0.0001) at baseline. As shown in Table 2 there were also changes in the distribution of patients in more severe disease categories defined in terms of the CGI-S score. After treatment with RLAI the proportion of patients with severe/extremely severe disease decreased from 17.9% at baseline to 0.2% at 24 months while the proportion defined as being not ill or with very mild disease increased from 1% to 15%. Similar changes were observed in the oral antipsychotic treatment group, although in this case 8.8% of patients were classified as not ill/having very mild disease at 24 months (versus 15% in the RLAI group). Both groups also demonstrated an improvement in global functioning over the study period. Mean GAF scores increased by 17.3 (from 46.9 ( n = 1306) to 64.2 ( n = 535)) in the RLAI group and by 14.9 (from 48.4 ( n = 264) to 63.3 ( n = 125)) in the oral antipsychotic patients. A full mixed model was also fitted to the GAF scores to compare the increase over time between the two treatment groups by controlling for baseline patient differences. Again, the RLAI treatment interaction term was statistically significant ( p < 0.03) indicating that, compared to the oral antipsychotic treated group, patients treated with RLAI had a 16% greater improvement in GAF scores over the course of the study. 3.9 Hospitalization For the analysis of hospitalization the mean number of stays per patient, the mean number of days hospitalized and the proportion of patients hospitalized over the 12 month period prior to the initiation of RLAI or an oral antipsychotic medication were compared to the same parameters in the first and second year post the initiation. Over the retrospective period, RLAI patients had a significantly greater intensity of hospitalization than the oral antipsychotic group in terms of the percent of patients hospitalized (35.1% versus 27.1%, p = 0.025) and number of hospital stays (0.49 versus 0.32 per patient, p = 0.019) but not for the mean duration of stay (21.21 versus 15.27 days, p = 0.357). As shown in Table 3 , for both the RLAI and oral antipsychotic treated patients there were statistically significant ( p < 0.0001) changes in hospitalization after the medication switch. For RLAI treated patients in the first 12 months after the therapy switch, 13.7% were hospitalized at least once (versus 35.1% over the pre-switch period), the population had a mean number of hospital stays of 0.21 per patient (versus 0.49) and the mean number of days hospitalized was 3.98 days per patient (versus 21.21), changes which were all statistically significant ( p < 0.0001). The decrease in hospitalization with respect to the pre-switch period continued into the second year of therapy with further decreases of 6.2% in the percent of patients hospitalized at least once, 0.09 stays per patient, and 1.52 days hospitalized per patient ( Table 3 ). The oral antipsychotic treatment group also showed significant changes ( p < 0.0001) in hospitalization over the first 12 months after the therapy switch with a lower percentage of patients hospitalized (10.0% versus 27.1% in the retrospective period) and reductions in the number of stays (0.14 versus 0.32) and days hospitalized (2.31 versus 15.3 days) ( Table 3 ). These changes also continued into the second year for these patients, but were less intense than with RLAI. In terms of the number of hospital stays per patient, the reductions with RLAI treatment (−0.28 and −0.37 over the first and second year of therapy, respectively) were 56% (year 1) and 85% (year 2) greater than those observed for patients receiving oral antipsychotics (−0.18 and −0.20, respectively); the full mixed model indicated that this difference over time was statistically significant ( p < 0.05) ( Fig. 4 ). Similarly, differences in decreases in the mean days hospitalized with RLAI therapy (−17.2 and −18.7 days per patient in post years 1 and 2, respectively) were 32% (year 1) and 44% (year 2) significantly greater ( p < 0.01) than the reductions in the oral antipsychotic treated patients (−13.0 days in both years) ( Fig. 4 ). There were also greater decreases in the percent of patients hospitalized with RLAI, which were 25% and 49% higher over the first and second year of therapy, respectively, than the decreases observed in the oral antipsychotic group ( Fig. 4 ). However, these differences were not statistically significant ( p = 0.4) based on a full mixed model that assessed decreases in the percent of patients hospitalized over time. 4 Discussion Although the safety and efficacy of RLAI have been demonstrated in controlled clinical trials, there are limited data on RLAI use in real clinical practice, compared to oral antipsychotic therapy [6] . The e-STAR study provides, to the best of our knowledge, the first analysis from a prospective observational study that directly compares the effectiveness of RLAI to that of oral antipsychotics in the routine clinical management of schizophrenia. The data from Spain demonstrate that, compared with oral antipsychotic therapy, patients initiated on RLAI stayed on therapy longer, had greater improvements in clinical and functional severity and greater reduction in hospital services. After 24 months of therapy 81.8% of the RLAI treated patients remained on treatment compared to 63.4% of the oral antipsychotic patients, a 29% increase ( p < 0.0001) in the likelihood of adherence to therapy with RLAI. This increase in adherence may reflect the improved efficacy of RLAI compared to oral antipsychotics [6] . The low RLAI discontinuation rates reported from Spain are supported by combined e-STAR interim data from six European countries where 89.6% of the patients ( n = 1847) were still on RLAI after 12 months [22,26] . The superior clinical and functional properties of RLAI were apparent in the reduction in CGI-S scores of 1.14 points and increase in GAF scores of 17.3. The results are consistent with previous short-term RLAI trials where decreases in CGI-S scores ranged from 0.6 to 0.87 [13,24,29] while GAF scores improved by approximately seven points following treatment initiation [23,24] . The marginally greater improvement in GAF score observed in the present study may be related to longer patient follow-up times. These data indicate that the clinical and functional efficacy of RLAI observed in previous studies [13,23,24,29] is maintained in patients receiving this therapy in Spain under real clinical practice conditions. The increased adherence to therapy with RLAI in this study confirms data from a recent study by Beauclair et al. [3] where patients with schizophrenia who were switched to RLAI had a significantly reduced risk of medication switching compared to patients switched to a new oral atypical (95% CI: 34.6%–58.4% for RLAI versus 55.7%–76.4% for oral atypicals). Haro et al. [15] in a study (the Schizophrenia Outpatient Health Outcomes (SOHO) study) with a similar observational design to the e-STAR study, found that after 3 years the proportion of schizophrenia patients ( n = 7728) remaining on oral atypical treatment ranged from 66% for clozapine to 34% for quetiapine; the rate for oral risperidone was 57%. The medication adherence rates reported in SOHO for oral atypicals are similar to the overall rate for oral therapy of 63% reported in the present observational study. However, in contrast, in the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) study [21] , oral antipsychotic medication discontinuation rates at 18 months ranged from 64% to 82% depending on the medication prescribed at baseline, rates considerably higher than those reported here or in the SOHO study [15] . However, the CATIE study was based on a blinded randomized design which may have contributed to the large differences in all cause discontinuation rates compared to the observational studies. In so far as poor adherence to therapy is correlated with a higher risk of rehospitalization [32] , our results confirm a relatively lower use of hospital resources in the more treatment-adherent RLAI group. Treatment with both RLAI and oral antipsychotics was associated with significant decreases in hospitalization compared to the pre-switch period, but the reductions were significantly greater in patients receiving RLAI for the number and days of hospitalizations over both the first and second 12 months of therapy. Since the RLAI treatment group had more intense hospitalization prior to the medication switch, the greater reduction in hospitalization days per patient as a result of the switch compared to the oral antipsychotic group has particular clinical significance. Furthermore, the data indicate a continuing long-term impact of RLAI on hospitalization between 12 and 24 months, an impact not apparent in the oral antipsychotic treated patients where first year decreases were relatively unchanged into the second year. These data on RLAI confirm previous studies which, for the most part, have shown that RLAI treatment consistently leads to a decrease in hospitalization, compared to prior therapy, or to a similar population receiving oral antipsychotic therapy [3,8,29] . The data also suggest that switching patients with schizophrenia to RLAI would lead to considerable cost savings over 24 months through reduced hospitalization. This has been confirmed by Olivares et al. [27] who have shown, using preliminary Spanish e-STAR data, that the contribution of hospitalization to the overall per patient treatment costs decreased from 76% pre-switch (24 month cost) to only 11% after the switch to RLAI. As a result total monthly treatment costs (including drug costs) over a 24 month period post switch to RLAI were €459.95 per patient (€ 2005), 22% lower than over a similar pre-switch period (€589.02 per patient per month). There are a number of limitations with the current study. This was an observational study and as such patients were not formally randomized to treatment, and in the recruitment process a proportionally greater number of patients receiving RLAI versus oral antipsychotics were enrolled (approximately a 4:1 ratio). However, the study protocol instructed physicians that any decision about a change in the type of antipsychotic medication should be made prior to, and independently of, any decision to enrol a specific patient in the study. While this ensures that patients enrolled in the study were representative of patients in the real world treated with RLAI or oral antipsychotics, it might also introduce imbalance between the RLAI and oral antipsychotic groups. Indeed, the RLAI cohort at baseline tended toward greater clinical severity than the oral antipsychotic group, with a longer duration of illness, a higher proportion of patients with schizophrenia and a higher prior utilization rate of conventional depot antipsychotics, concomitant non-antipsychotic medications, and hospitalization services. However, full mixed models were used to control for differences in demographic and clinical characteristics as well as study drop outs between the two treatment groups when treatment outcomes were compared. Furthermore, for hospital outcomes, the emphasis was on a comparison of pre-post changes; this difference-in-difference design, where each patient effectively serves as their own control, also helps to minimize this limitation. Another limitation was the high drop out rates over the 24 month study period (41% of the patients in both groups dropped out of the study) and the difference in treatment discontinuation. In many cases the reason for drop out was unclear and this may have had an impact on the results, particularly for the oral antipsychotic group where the number of participating patients was low to begin with. However, loss of patients to follow-up is a common phenomenon in real clinical practice, particularly in patients with schizophrenia, so in some respects the observational nature of the study is an appropriate design to quantify this phenomenon. In any case the full mixed model statistical analysis was used to mitigate the impact of differences in discontinuation and drop out rates between treatment groups. Finally, because of differences in culture and health care systems between countries, findings from this study in Spain may not be generalizable to other countries. However, interim results from pooled e-STAR data from six European countries have shown that low discontinuation rates with RLAI in real clinical practice were consistently observed [26] . 5 Conclusions The present large-scale study demonstrates that in a real clinical practice setting in Spain, RLAI was associated with significantly better adherence compared to oral antipsychotic treatment and was more effective in improving clinical and functional severity and in reducing the use of concomitant medications and the number and days of hospitalizations. This was despite the fact that RLAI may have been used in more severe and difficult to treat patients than the oral antipsychotic population. The data provide evidence that RLAI may offer more effective care and cost savings for patients with schizophrenia in a real world clinical setting than oral antipsychotics. Acknowledgements The study was supported by a grant from Janssen Cilag. e-STAR Spanish study group: J.C. Villalobos Vega, J. Alonso Cuéllar, F.J. Alberca de Castro, C. Morillo-Velarde Quintero, J.F. Román Martíin, P. Tabares Domínguez, J.L. Prados Ojeda, S. Sanz Cortés, F.I. Mata Cala, C. Gutiérrez Marín, L. Moyano Castro, M.A. Haza Duaso, J. Requena Albarracín, G. Narbona Vergara, J.A. Fernández Benítez, F. Mayoral Cleries, J.M. García-Herrera Pérez-Brian, A. Bordallo Aragón, J.C. Rodríguez Navarro, J.A. Algarra Biedma, R. Bravo de Pedro, J.F. Delgado González, M.E. Jaén López, H. Díaz Moreno, J.A. Soto López, E. Ojeda Rodríguez, C. Martínez de Hoyos, M. Pardilla Sacristán, M.D. Molina Martín, E. Martín Ballesteros, P.A. Sopelana Rodríguez, L. Fernández Menéndez, R. Santos Rivas, P. del Pino Cuadrado, J. Correas Lauffer, J.J. Rodríguez Solano, J.M. Fernández Martínez, F. García Solano, P. García-Lamberde Rodríguez, J.A. Romero Rodríguez, T. Rodríguez Cano, M. Ducaju Fortacin, J.M. Blanco Lobeiras, J.M. Piñeiro Sampedro, A. Pérez Bravo, A. Fernández Pellicer, M.D. Alonso López, J. Fraga Liste, M. Riobo Fernández, A. Casas Losada, R. Vazquez-Noguerol Mendez, S. Agra Romero, J.J. Blanco Blanco, I. Tortajada Bonaselt, M.C. García Mahia, E. Ferrer Gómez del Valle, P. Quiroga Yañez, M. Gelabert Camarasa, J.A. Barbado Alonso, G. Florez Mendez, F. Doce Feliz, M.A. López Lamela, M. Vega Piñero, P. Fuentes Alvarado, I. López Gómez, P. Fadon Martín, J.L. Santos Gómez, A. García López, R. Rodríguez Jiménez, A. Escudero Nafs, N. Casas Barquero, R. Fernández-Villamor Ortiz, J.L. Velez Noguera, P. Ruiz Carrasco, J. Martín Muñoz, M. Masegoza Palma, C. Marín Hortelano, L. Sánchez Bonome, J. Sánchez Sevilla, J.M. Mongil San Juan, J.M. García Ramos, J.L. Vallejo Muñoz, J. Elorza Guisasola, L. Santamaria Vazquez, F. Campo Guerras, F.J. Arrufat Nebot, F.J. Baron Fernández, A.L. Palomo Nicolau, R. Catala Subirats, M. Messays Kidias, V. Fabregat Navarro, B. Frades García, F. Mejias del Rosal, T. de Vicente Muñoz, J. Año Ballester, P. Malabia Lieb, A. Delgado Martel, E. Roca Bea, I. Grau Joaquim, F. Boatas Enjuanes, M. Bañuelos Piñol, E. Fontova i Carbonell, R. Martín Muñoz, C. Argila Giribets, L. Albages Sans, A. Serrano Blanco, M. Arcega Felipe, P. González Muñoz, A. Pons Villanueva, M. Bernardo Arroyo, R. Coronas Borri, S. Miret Fallada, M. Celma Merola, E. Parellada Rodon, J.R. Pigem Palmes, E. Pérez Martínez, J. Matarredona Catala, A. Sandoval Coca, F. Pascual Ferrandiz, E. Ferrandiz Paya, G. Iturri Caballero, A. Franco Bonet, J. Fluvia Figueras, P. Moreno Pagador, M. Medina Garibo, V. Pérez Camo, C. Sanz Carrillo, C. Pelegrin Valero, F.J. Caro Rebollo, J. García Campayo, J.M. Sala Ayma, M. Martínez Roig, M.A. de Uña Mateos, R. García Bertolin, A. Martín García, F. Jiménez Mazo, J.L. Galvez Velasco, L. Santa Maria Pérez, C. Jiménez Casado, J.J. Mancheño Barba, M. Conde Diaz, J.P. Alcon Rubio, A. Soler Mandoli, A. Uson Herrero, A. Rodríguez Martínez, P. Salgado Serrano, E. Nieto Rodríguez, J. Segui Montesinos, J. Ferragud Macia, A. Mateos Marcos, J.V. Pérez-Fuster Soto, M. Verdaguer Dumont, J. Parra Pagan, V. Balanza Martínez, M. Santiuste de Pablos, C. Espinosa Delgado, M.D. Martínez Quiles, F.J. Manzanera López, P. Pozo Navarro, A. Micol Torres, F.J. Martínez Ingles, J.M. Salmeron Arias-Camison, J.C. López Manzano, R. Villanueva Peña, G. Petersen Guitarte, H. Blasco Fontecilla, J. Barjau Romero, R. Sanz Gil, J. Marín Lozano, L. Donaire Adanez, I. Zarranz Herrera-Oria, J. Pérez Jiménez, F. Carrato Vaz, O. Sanz García, C. Contreras Anton, R. Reixach Casula, M.C. Natividad Hernandez, F. Teba Escabias, J. Rodríguez Torresano, A. Huidobro Pérez-Villamil, L. Estevez, M. Aragües Figuero, A. Muñoz de Morales, J.L. Rodríguez Calvin, M. Delgado Criado, V. Molina Rodríguez, E. Balbo Ambrosolio, P.M. Holgado Madera, G. Ponce Alfaro, M.M. Rojas Vidal, A. García Valtuille, O. Ruiz, G. Lucas Cabornero, M. Echevarria Martínez de Bujo, M.J. Maicas Mallen, J. Santandreu Puigros, A. Liñana Martorell, A. Clar Forteza, E. Rodríguez Arrebola, M. Rodríguez de la Torre, C.G. Anton Saiz, C. Bardolet i Casas, E. Rodríguez Linde, R. De Arce Cordon, E.M. Padial Molina, F.J. Ruiz Carazo, J.J. Muro Romero, D. Vico Cano, M. Soria Dorado, S. Campos Velazquez, A.J. Rodríguez Sánchez, S. Ocio Leon, H. Pachas Sánchez, M. Henry Benitez, A. Intxausti Zugarramurai, M.A. Contreras, M. De la Varga González, P. Barreiro Marín, F. Gómez Robina, M. Sánchez García, F.J. Otero Pérez, P. Cubero Bros, A. Carrillo Gómez, J. de Dios Molina Martín, J.L. Carrasco Perera, M.C. Averbach, J.L. Carrasco Perera, E. Goenaga Palancares, M.T. Gallego de Dios, C. Fernández Rojo, S. Sánchez Iglesias, M.I. Rubio Merino, N. Prieto Mestre, A. Pérez Urdaniz, J.M. Martínez Sánchez, R. Gordo Seco, J. Franco Muñoz, M. Mateos Agut, M.L. Blanco Lozano, F. Martín Herguedas, A. Torcal Pena, J. Vicente García, A. Varona Martínez, O. Sanz Granado, M.A. Medina Fernández, J.M. Moran Canseco, P.A. Megia López, M.A. Franco Martín, J.A. Espina Barrio, J. Giner Ubago, M. Roca Bennassar, J.M. Olivares Díez, J.L. Hernandez Fleta, F. Porras Fortes, C. Arango López, O. Medina, D. Figuera Alvarez, J.M. Peña Roca, G. Rubio Valladolid, J.A. Furquet Tavera, J.A. García-Castrillon Sales, I. Batalla Llordes, C. Anchuistegui Melgarejo, F. Cañas de la Paz, V. Vallés Callol, M. Bousoño García, J. Bobes García, F.J. Vaz Leal, E. Cáceres Corrales, E. Sánchez Iglesias, M.A. Carreiras Gómez, G. García Serrano, E.G. Román Chillarón, F.J. Samino Aguado, J.J. Molina Castillo, A. González González, J. Gallardo Vázquez, M. Bolivar Peralvarez, M. Rios Diaz, M. Ybarzabal Mesa, F.J. Acosta Artiles, M. Ajoy Chao, M. Ybarzabal Mesa, P. del Rosario Santana, M.A. García Escudero, M. Molla Berenguer, J.M. Bonete Llacer, J.A. Juan Berna, J. Barragán Ortiz, L. Tost Pardell, C. Hernández-Alvarez de Sotomayor, M.R. Cejas Méndez, R. Cabrera Garate, B. Díaz Múgica, M. Caballero González, J. Pujol Domingo, C. Sáez Navarro, G. Selva Vera, M.A. Cuquerella, J. Lonjedo Monzo, P. Cervera Boada, M.F. Martín Pérez, E. Carrasco Parrado, J.J. Yañez Sánchez, and J. Calvo Fernández. References [1] F.J. Acosta E. Bosch G. Sarmiento N. Juanes A. Caballero-Hidalgo T. Mayans Evaluation of noncompliance in schizophrenia patients using electronic monitoring (MEMS ® ) and its relationship to sociodemographic, clinical and psychopathological variables Schizophr Res 2008 E-pub [2] C.E. Adams M.K. Fenton S. Quraishi A.S. David Systematic meta-analysis of depot antipsychotic drugs for people with schizophrenia Br J Psychiatry 179 2001 290 299 [3] L. Beauclair P. Chue J. McCormick F. Camacho A. Lam D. Luong Impact of risperidone long-acting injectable on hospitalization and medication use in Canadian patients with schizophrenia J Med Econ 10 2007 427 442 [4] N.H. Bhanji G. Chouinard H.C. Margolese A review of compliance, depot intramuscular antipsychotics and the new long-acting injectable atypical antipsychotic risperidone in schizophrenia Eur Neuropsychopharmacol 14 2004 87 92 [5] N. Black Why we need observational studies to evaluate the effectiveness of health care Br Med J 312 1996 1215 1218 [6] P. Chue Long-acting risperidone injection. Efficacy, safety, and cost-effectiveness of the first long-acting atypical antipsychotic Neuropsychiatr Dis Treat 2007 E-pub [7] P. Chue M. Eerdekens I. Augustyns B. Lauchaux P. Molcan L. Eriksson Comparative efficacy and safety of long-acting risperidone and risperidone oral tablets Eur Neuropsychopharmacol 15 2005 111 117 [8] P. Chue P. Llorca I. Duchesne A. Leal D. Rosillon A. Mehnert Hospitalization rates in patients during long-term treatment with long-acting risperidone injection J Appl Res 5 2005 266 274 [9] L.M. Davies M.F. Drummond Economics and schizophrenia: the real cost Br J Psychiatry 25 Suppl. 1994 18 21 [10] J.M. Davis L. Matalon M.D. Watanabe L. Blake L. Matalon Depot antipsychotic drugs: place in therapy Drugs 47 1994 741 743 [11] J. Endicott J. Spitzer J. Fleiss J. Cohen The global assessment scale: a procedure for measuring overall severity of psychiatric disturbance Arch Gen Psychiatry 33 1976 766 771 [12] W.W. Fleischhacker M. Eerdekens K. Karcher G. Remington P.M. Llorca W. Chrzanowski Treatment of schizophrenia with long-acting injectable risperidone: a 12-month open-label trial of the first long-acting second-generation antipsychotic J Clin Psychiatry 64 2003 1250 1257 [13] G.M. Gharabawi N.C. Gearhart R.A. Lasser R.A. Mahmoud Y. Zhu E. Mannaert Maintenance therapy with once-monthly administration of long-acting injectable risperidone in patients with schizophrenia or schizoaffective disorder: a pilot study of an extended dosing interval Ann Gen Psychiatry 6 2007 3 [14] W. Guy Clinical global impression A.J. Rush Handbook of psychiatric measures 2000 American Psychiatric Association Washington, DC 100 102 [15] J.M. Haro D. Suarez D. Novick J. Brown J. Usall D. Naber Three-year antipsychotic effectiveness in the outpatient care of schizophrenia: observational versus randomized studies results Eur Neuropsychopharmacol 17 2007 235 244 [16] A. Hofer M. Hummer R. Huber M. Kurz T. Walch W.W. Fleischhacker Selection bias in clinical trials with antipsychotics J Clin Psychopharmacol 20 2000 699 702 [17] J.M. Kane E. Aguglia A.C. Altamura J.L. Ayuso Gutierrez N. Brunello W.W. Fleischhacker Guidelines for depot antipsychotic treatment in schizophrenia European Neuropsychopharmacology Consensus Conference in Siena, Italy Eur Neuropsychopharmacol 8 1998 55 66 [18] J.M. Kane M. Eerdekens J.P. Lindenmayer S.J. Keith M. Lesem K. Karcher Long-acting injectable risperidone: efficacy and safety of the first long-acting atypical antipsychotic Am J Psychiatry 160 2003 1125 1132 [19] M.J. Kikkert A.H. Schene M.W. Koeter D. Robson A. Born H. Helm Medication adherence in schizophrenia: exploring patients', carers' and professionals' views Schizophr Bull 32 2006 786 794 [20] J.P. Lacro L.B. Dunn C.R. Dolder S.G. Leckband D.V. Jeste Prevalence of and risk factors for medication nonadherence in patients with schizophrenia: a comprehensive review of recent literature J Clin Psychiatry 63 2002 892 909 [21] J.A. Lieberman T.S. Stroup J.P. McEvoy M.S. Swartz R.A. Rosenheck D.O. Perkins Effectiveness of antipsychotic drugs in patients with chronic schizophrenia N Engl J Med 353 2005 1209 1223 [22] Medori R, Wapenaar R, de Arce R, Rouillon F, Gaebel W, Cordes J, et al. Relapse prevention and effectiveness in schizophrenia with risperidone long-acting injectable (RLAI) versus quetiapine. Presented at the annual meeting of the American Psychiatric Association, 3–7 May 2008, Washington DC, USA. [23] A. Mohl K. Westlye S. Opjordsmoen A. Lex A. Schreiner M. Benoit Long-acting risperidone in stable patients with schizoaffective disorder J Psychopharmacol 19 5 Suppl. 2005 22 31 [24] H.J. Möller P.M. Llorca E. Sacchetti S.D. Martin R. Medori E. Parellada Efficacy and safety of direct transition to risperidone long-acting injectable in patients treated with various antipsychotic therapies Int Clin Psychopharmacol 20 2005 121 130 [25] NICE (National Institute for Clinical Excellence, UK). Schizophrenia: full national clinical guideline on core interventions in primary and secondary care. Available at < http://www.nice.org.uk/guidance/index.jsp?action=download&;o=29033 >. Schizophrenia full guideline; 2003. p. 68 and 228. [26] Olivares JM, Peuskens J, Pecenak J, Tuma I, Van Kooten M, Eriksson L, et al. Treatment retention with risperidone long acting injection (RLAI) in European patients with schizophrenia: 12-month interim results from electronic Schizophrenia Treatment Adherence Registry (e-STAR). International Society for Pharmacoeconomics and Outcomes Research, 10th annual European Congress, 20–23 October 2007, Dublin, Ireland. Abstract PMH9. [27] J.M. Olivares A. Rodriguez-Martinez J.A. Buron D. Alonso-Escolano A. Rodriguez-Morales Cost-effectiveness analysis of switching antipsychotic medication to long-acting injectable risperidone in patients with schizophrenia. A 12- and 24-month follow-up from the e-STAR database in Spain Appl Health Econ Health Policy 6 2008 41 53 [28] D. Robinson M.G. Woerner S. Pollack G. Lerner Subject selection biases in clinical trials: data from a multicenter schizophrenia treatment study J Clin Psychopharmacol 16 1996 170 176 [29] D.M. Taylor C. Young M. Patel Prospective 6-month follow-up of patients prescribed risperidone long-acting injection: factors predicting favourable outcome Int J Neuropsychopharmacol 9 2006 685 694 [30] B. Thornly C. Adams Content and quality of 2000 controlled trials in schizophrenia over 50 years Br Med J 317 1998 1181 1184 [31] M. Turner E. Eerdekens M. Jacko M. Eerdekens Long-acting injectable risperidone: safety and efficacy in stable patients switched from conventional depot antipsychotics Int Clin Psychopharmacol 19 2004 241 249 [32] M. Valenstein L.A. Copeland F.C. Blow J.F. McCarthy J.E. Zeber L. Gillon Pharmacy data identify poorly adherent patients with schizophrenia at increased risk for admission Med Care 40 2002 630 639 [33] M. Valenstein F.C. Blow L.A. Copeland J.F. McCarthy J.E. Zeber L. Gillon Poor antipsychotic adherence among patients with schizophrenia: medication and patient factors Schizophr Bull 30 2004 255 264 [34] D.C. Watt K. Katz M. Shepherd The natural history of schizophrenia: a 5-year prospective follow-up of a representative sample of schizophrenics by means of a standardized clinical and social assessment Psychol Med 13 1983 663 670 [35] J.L. Young H.V. Zonana L. Shepler Medication noncompliance in schizophrenia: codification and update Bull Am Acad Psychiatry Law 14 1986 105 122