Abstract
Background and Hypothesis
Antipsychotics remain the main treatment for schizophrenia, but their effectiveness is challenging to compare. We aimed to assess the comparative real-world effectiveness of antipsychotics in preventing readmission among patients in Asia with early-stage schizophrenia to inform clinical decision making.
Study Design
We did a retrospective cohort study of first-admission schizophrenia patients (ICD-9-CM: 295; ICD-10-CM: F20 and F25) from January 1, 2001, to December 31, 2017. The cohort was identified from the National Health Insurance Research Database NHIRD for Psychiatric Inpatients. The exposure was any antipsychotics prescribed post-discharge. The primary outcome was the readmission risk due to psychotic disorders, which was measured by adjusted hazard ratios (aHRs). Within-individual extended Cox models were applied for analyses, where the periods of oral risperidone use served as his or her own control.
Study Results
We selected 75 986 patients (men, 53.4%; mean [SD] age, 37.6 [12.0] years; mean [SD] duration of follow-up, 8.9 [5.0]) who were first admitted to psychiatric wards with schizophrenia in Taiwan. Among them, 47 150 patients (62.05%) had at least one readmission within 4 years. Compared to the period under treatment with oral risperidone, that under monotherapy with long-acting injectable antipsychotics (LAIs) had the lowest risk for psychotic readmission, with a risk reduction of 15–20%. However, the prevalence of person-prescription prevalence of LAIs remained low (< 10%) during the follow-up period.
Conclusions
The use of LAIs after the first admission for schizophrenia has notable advantages in preventing readmission. Such formulations should be offered earlier in the course of illness.
Keywords: long-acting injectable antipsychotics, risperidone, hospitalization, mental health services, National Health Insurance
Introduction
Schizophrenia is a severe mental disorder that often entails repeated relapses and has been one of the leading causes of disability worldwide,1 with a median lifetime prevalence of 6.35 per 1000 among post-1990 studies.2 Antipsychotic drugs remain the mainstay of treatment for schizophrenia, providing the relief of symptoms and the prevention of deterioration or relapse, and effective treatment in the early stages of the disease might improve long-term outcomes.3 As numerous antipsychotics, ranging from first-generation (FGAs) to second-generation antipsychotics (SGAs), are currently available, many comparative efficacy evaluations of their formulations are based on randomized controlled trials.4,5 Despite their rigorousness for causal inference, these trials might not reflect the real-world situations. As a case in point, the beneficial efficacy in relapse prevention of long-acting injectable antipsychotics (LAIs) compared to oral antipsychotics was not detected in an earlier meta-analysis of 21 randomized controlled trials.5 It was found only in a later meta-analysis of 32 more recent trials that enrolled patients at high risk for nonadherence.6
Alternatively, the effectiveness of different antipsychotics under natural settings could be evaluated using observational studies with large sample sizes. Nevertheless, these observational studies have faced other types of limitations. Patients who received clozapine or LAIs tended to be more severely ill and have poorer prognoses. In addition, treatment decisions (eg, hospitalization) might be influenced by the knowledge of patients’ route of antipsychotic administration. These shortcomings have been linked to a lack of difference between LAIs and oral antipsychotics in relapse prevention in a meta-analysis of 42 cohort studies.7 A new method called the within-individual extended Cox model8,9 could help reduce confounding by the internal characteristics among patients in assessing effectiveness and revealed that oral clozapine and LAIs had a lower risk of readmission in Nordic populations.9,10 For first-episode schizophrenia patients, LAIs also had a lower risk of readmission than oral antipsychotics in those studies using a within-individual approach in Nordic countries9,10 and South Korea11 but not in a meta-analysis of observational studies.12 Since the use of mental health services is influenced by many factors13,14 and the prescription of antipsychotics varies across countries,15,16 more research on the real-world effectiveness of antipsychotics among patients with early-stage schizophrenia is warranted.
To fill this gap in the literature, we turned to the national cohort of first-admission schizophrenia patients established by the National Health Insurance Research Database (NHIRD)17 from 2001 to 2017. This study aims to (1) evaluate the trend of the 4-year readmission risk among first-admission schizophrenia patients over the years, (2) compare the risk of readmission for different antipsychotics among first-admission schizophrenia patients using within-individual analysis, and (3) compare the trend in prescription prevalence for different antipsychotics during the study period.
Methods
Data Source
The NHIRD covers an extremely high percentage of the Taiwanese population (97% in 2001 and >99.9% in 2014) and contains information about the registry for beneficiaries, ambulatory care claims, inpatient claims, and prescriptions dispensed at pharmacies.17 We used the NHIRD for Psychiatric Inpatients to conduct a retrospective population-based cohort study of inpatients diagnosed with schizophrenia. This project was approved by the Research Ethics Committee of the National Taiwan University Hospital (NTUH-REC no. 201911097RIND).
Study Population
Patients who were first admitted with a diagnosis of schizophrenia (ICD-9-CM: 295; ICD-10-CM: F20 and F25) to a psychiatric ward between 2001 and 2017 were selected as a dynamic incident cohort, with each first-admission patient being free of any admission for schizophrenia between 2000 and the inception of NHIRD in 1995. Using the criteria of inclusion (first-discharge diagnosis of schizophrenia from a psychiatric ward) and exclusion (patients below age 15 or over age 64 at first admission, patients of unknown sex, and patients who died during the first admission), 75 986 inpatients were included in this study (supplementary figure S1). Based on these patients, a total of 349 873 admissions were identified, of which 209 223 were psychotic readmissions.
Exposure
The exposure was the antipsychotic medication (supplementary table S1) prescribed post-discharge by extracting antipsychotic prescriptions dispensed at pharmacies in the NHIRD. We linked the prescription items in the claim data to a detailed pharmaceutical list maintained by the National Health Insurance Bureau, which included information about the ingredient name, brand name, and Anatomical Therapeutic Chemical code of antipsychotics. We combined refillable prescriptions into one record (supplementary method A1 and figure S2).
The initiated date of each antipsychotic was the prescription date, and the length of exposure time was defined as days of prescription or shorter if a new prescription of other antipsychotics was added. Treatment periods were classified as monotherapy if a specified antipsychotic was prescribed, polypharmacy if any 2 or more antipsychotics were prescribed at the same time, or no therapy with any antipsychotics if no prescription records were found within 14 days from the expected date of the next prescription. The prescription pattern for the antipsychotic medication was calculated as the person-prescription prevalence within 1 year post first discharge.
Outcomes
The primary outcome was readmission for any psychotic disorder (supplementary table S2), which reflects the treatment effects of antipsychotics. The secondary outcome was all-cause readmission, including therapeutic outcomes and side effects. The follow-up of a first-admission patient was censored by death, age over 65 years, or the study’s end.
Statistical Analysis
We pooled person-years under specific exposure (ie, antipsychotic treatment) over the whole cohort and counted the number of events (ie, readmissions) rising from the pool of person-years to estimate the incidence rate of readmission. We then applied survival analysis to estimate readmission risk up to 4 years (ie, 1460 d) after discharge from the time of first admission.
We applied a stratified within-individual extended Cox model8,9 to assess the effectiveness of different antipsychotics using risperidone, the most commonly prescribed antipsychotics in Taiwan, as the reference treatment. Each patient was treated as an independent stratum so that between-individual confounding effects were avoided. For each individual, the follow-up time in this model was reset to zero at the discharge of an admission. Different from traditional Cox proportional hazards regression, extended Cox allows exposure to be updated with time and does not require the assumption of proportional hazards. Antipsychotic use was tracked over time by treating it as a time-dependent exposure, and the exposure time of antipsychotics was defined by start time and end time using counting processes. Under the time-resetting strategy of counting processes, when the follow-up time was reset due to antipsychotic shifts, the preceding period would end as being censored, and the subsequent period would start with a left truncation. More details are provided in supplementary method A2.
To adjust for possible confounders, we applied the multivariate extended Cox model to derive adjusted hazard ratios (aHRs) to estimate the relative risk for readmission. Two time-dependent covariates were included in the model: (1) time since the first admission and (2) the number of previous treatment switches after the last discharge, with a greater number of antipsychotic switches indicating a more unstable condition of the patient.
Sensitivity analyses were conducted for the following scenarios: (1) further adjustment for carryover/delayed effects; (2) excluding the insomnia indication of quetiapine prescriptions; (3) including monotherapy using LAIs that were less commonly used; (4) further adjustment for co-medication, including antidepressants, sedatives/hypnotics, benzodiazepines, and short-acting injections; (5) using oral olanzapine as the reference to compare the results of our study with those of Nordic countries; (6) comparing the effectiveness of the pooled group of all LAIs with that of the pooled group of all oral antipsychotics; and (7) comparing the effectiveness of each LAI with its oral counterpart as the reference.
Results
The Cohort of First-Admission Schizophrenia Patients
The sociodemographic and clinical characteristics of the 75 986 patients included in this study are displayed in table 1. About 49.2% of these patients received antipsychotic treatment within 1 year prior to the first admission. During follow-up, 47 150 patients (62.1%) had at least one readmission within 4 years before censorship. The mean duration of follow-up after discharge was 8.9 years. The average number of readmissions within 2 years post-discharge was 1.4 for all-cause readmissions and 1.1 for psychotic readmissions.
Table 1.
Sociodemographic and Clinical Characteristics of Patients Who Were First Admitted for Schizophrenia Between 2001 and 2017 in Taiwan (N = 75 986)
| Variables | N (%) or Mean (SD) |
|---|---|
| Baseline | |
| Male sex, N (%) | 40 681 (53.4) |
| Age group at first admission, N (%) | |
| 15–24 | 12 337 (16.2) |
| 25–34 | 20 552 (27.1) |
| 35–44 | 20 264 (26.7) |
| 45–54 | 15 432 (20.3) |
| 55–64 | 7401 (9.7) |
| Age at first admission (y), Mean (SD) | 37.6 (12.0) |
| Duration of the first admission (d), Mean (SD) | 36.6 (30.0) |
| Diagnosis, N (%) | |
| Schizophrenia | 69 619 (91.6) |
| Schizoaffective disorders | 5811 (7.7) |
| Both | 556 (0.7) |
| First-admission year, N (%) | |
| 2001–2005 | 31 294 (41.2) |
| 2006–2010 | 21 783 (28.7) |
| 2011–2015 | 17 122 (22.5) |
| 2016–2017 | 5787 (7.6) |
| Involuntary first admission, N (%) | |
| Yes | 1592 (2.1) |
| No | 74 394 (97.9) |
| Use of antipsychotics within 1 y prior to first admissiona, N (%) | 33 956 (49.2) |
| Use of LAIs within in 1 y prior to first admissiona, N (%) | 4169 (6.1) |
| Follow-up | |
| Having ≥ one readmission within 4 y before censorship, N (%) | 47 150 (62.1) |
| The length of post-discharge follow-up (y)b,Mean (SD) | 8.9 (5.0) |
| No. of all-cause readmissions within 2 yc,Mean (SD) | 1.4 (1.9) |
| No. of psychotic readmissions within 2 yc,Mean (SD) | 1.1 (1.6) |
Note:
aOnly including the cohorts of 2002 to 2016, with the total number = 68 973.
bThe follow-up was from post-discharge of the first admission until the end of study on December 31, 2017, death, or age 65 y old.
cThe cohorts of 2016 and 2017 were not included in this part of the analysis.
Although the number of first-admission patients with schizophrenia decreased steadily over the years, the 4-year readmission risk decreased slightly, from 68.16% in 2001 to 64.35% in 2013 (figure 1). Thus, we pooled the person-years among the cohorts of different years, with a total of 628 360 person-years and 209 223 readmissions for any psychotic disorder. There were 66 032 readmission events arising from 350 511 person-years that were without any use of antipsychotics (incidence rate = 1.88 per 10 person-years; 95% CI: 1.87–1.90). Since claims data did not provide information about whether no prescription of antipsychotics was due to patients’ poor adherence or improvements in the symptoms, the person-years without antipsychotics prescription were not used as the reference exposure in this study.
Fig. 1.
The 4-y readmission risks (Kaplan–Meier survival estimates and their 95% confidence interval) among first-admission patients with schizophrenia from 2001 to 2013 (top) and the yearly number of first-admission patients with schizophrenia from 2001 to 2017 (bottom) in Taiwan.
For the remaining 277 849 person-years with the use of antipsychotics, there were 143 191 readmissions (table 2). Approximately three-fourths of the person-years were constituted by monotherapy (76.73%), and one-fourth were constituted by polypharmacy (23.27%). Among monotherapies, the greatest proportion of person-years was constituted by oral antipsychotics, including risperidone (18.53%), sulpiride (8.19%), olanzapine (8.01%), clozapine (7.19%), and quetiapine (6.31%). Hence, we chose the most commonly prescribed antipsychotic, risperidone, as the reference treatment.
Table 2.
Incidence Rates (IR) of Psychotic Readmission During the Follow-up with Specific Antipsychotics
| Antipsychotics | No. of Events | Person-years (PY; %) | IR/10-PY (95% CI) |
|---|---|---|---|
| Total with use of antipsychotics | 143 191 | 277 849 (100.00) | 5.15 (5.13–5.18) |
| (1) Monopharmacy | 98 137 | 213 190 (76.73) | 4.60 (4.57–4.63) |
| First-generation oral antipsychotics | 19 941 | 45 082 (16.23) | 4.42 (4.36–4.48) |
| Chlorpromazine | 1225 | 2281 (0.82) | 5.37 (5.07–5.67) |
| Chlorprothixene | 72 | 130 (0.05) | 5.54 (4.26–6.82) |
| Flupentixol | 2439 | 4922 (1.77) | 4.95 (4.76–5.15) |
| Haloperidol | 5271 | 10 452 (3.76) | 5.04 (4.91–5.18) |
| Levomepromazine | 16 | 27 (0.01) | 5.87 (2.99–8.75) |
| Loxapine | 225 | 614 (0.22) | 3.66 (3.18–4.14) |
| Pimozide | 310 | 298 (0.11) | 10.40 (9.24–11.55) |
| Sulpiride | 8823 | 22 765 (8.19) | 3.88 (3.79–3.96) |
| Thioridazine | 309 | 735 (0.26) | 4.20 (3.73–4.67) |
| Trifluoperazine | 871 | 2 189 (0.79) | 3.97 (3.71–4.24) |
| Zuclopenthixol | 380 | 669 (0.24) | 5.68 (5.11–6.25) |
| First-generation LAIs | 2536 | 6571 (2.36) | 3.86 (3.71–4.01) |
| Flupentixol LAI | 1373 | 3103 (1.12) | 4.42 (4.19–4.66) |
| Fluphenazine LAI | 184 | 554 (0.20) | 3.32 (2.84–3.80) |
| Haloperidol LAI | 831 | 2483 (0.89) | 3.35 (3.12–3.57) |
| (Zu)clopenthixol LAI | 148 | 431 (0.16) | 3.43 (2.88–3.99) |
| Second-generation oral antipsychotics | 73 575 | 156 343 (56.27) | 4.71 (4.67–4.74) |
| Amisulpride | 5306 | 14 077 (5.07) | 3.77 (3.67–3.87) |
| Aripiprazole | 4370 | 12 129 (4.37) | 3.60 (3.50–3.71) |
| Clotiapine | 1216 | 1981 (0.71) | 6.14 (5.79–6.48) |
| Clozapine | 10 747 | 19 965 (7.19) | 5.38 (5.28–5.48) |
| Lurasidone | 18 | 14 (0.01) | 12.80 (6.89–18.72) |
| Olanzapine | 9552 | 22 248 (8.01) | 4.29 (4.21–4.38) |
| Paliperidone | 3281 | 6553 (2.36) | 5.01 (4.84–5.18) |
| Quetiapine | 11 825 | 17 529 (6.31) | 6.75 (6.62–6.87) |
| Risperidone | 21 622 | 51 499 (18.53) | 4.20 (4.14–4.25) |
| Ziprasidone | 1173 | 2384 (0.86) | 4.92 (4.64–5.20) |
| Zotepine | 4465 | 7964 (2.87) | 5.61 (5.44–5.77) |
| Second-generation LAIs | 2078 | 5161 (1.86) | 4.03 (3.85–4.20) |
| Paliperidone LAI 28 | 483 | 823 (0.30) | 5.87 (5.34–6.39) |
| Risperidone LAI | 1595 | 4338 (1.56) | 3.68 (3.50–3.86) |
| Other | 7 | 33 (0.01) | 2.12 (0.55–3.69) |
| (2) Polypharmacy | 45 054 | 64 659 (23.27) | 6.97 (6.90–7.03) |
Comparison of Antipsychotic Effectiveness in the Prevention of Readmission
Compared to oral risperidone, the results of the within-individual extended Cox regression analysis of 18 other monotherapies with sufficient person-years are displayed in figure 2. The risk for psychotic readmission was reduced 17–22% for monotherapy with LAIs, with the lowest risk shown for haloperidol LAIs (aHR = 0.79, 95% CI 0.72–0.86), followed by flupentixol LAIs (aHR = 0.82, 95% CI 0.76–0.88) and risperidone LAIs (aHR = 0.83, 95% CI 0.77–0.89) (figure 2A); risk with paliperidone LAIs was not significantly different from that of oral risperidone. The results remain similar for the analysis of all-cause readmissions (figure 2B). For the time-dependent covariates in the model, longer time since first admission and more switches of antipsychotics was associated with a higher risk (supplementary table S3).
Fig. 2.
The hazard ratios of different antipsychotics in monotherapy compared to oral risperidone for the event of (A) psychotic readmissions; (B) all-cause readmissions in the multivariate model, with adjustment for age at first antipsychotic prescriptions and number of previous treatment switches.
When the analyses were expanded to include various combinations of 2 antipsychotics, compared to oral risperidone monotherapy, the 3 LAI monotherapies were still the most effective in preventing both psychotic and all-cause readmissions (supplementary figure S3). Nevertheless, most of the polypharmacy combinations were not significantly more effective than oral risperidone monotherapy in preventing both psychotic and all-cause readmissions.
The results remained similar in sensitivity analyses of carryover effects (reassigned exposure for the event within 7 days post switching antipsychotics), changing quetiapine indication (for the treatment of insomnia rather than psychosis), and the inclusion of 2 more LAI monotherapies (zuclopenthixol LAI and fluphenazine LAI), showing better effectiveness than the 3 initial LAIs (supplementary figure S4). Similar results showing LAIs with the best effectiveness by comparing to oral olanzapine as the reference treatment are shown in supplementary figure S5. The pooled group of LAIs had better effectiveness than the pooled group of oral antipsychotics (aHR = 0.84, 95% CI 0.81–0.88), and each LAI also showed a lower risk of readmission than its oral counterpart except paliperidone LAI (supplementary table S4).
Prescription Prevalence of Antipsychotics
The person-prescription prevalence of individual antipsychotics in each year from 2001 to 2016 is provided in supplementary table S5 and figure S6. The top 3 prescribed antipsychotics in each year from 2001 to 2016, a pool of 6, are displayed in figure 3A; all are oral formulations (the corresponding values are in supplementary table S6). Among them, oral risperidone was the most commonly prescribed antipsychotic since 2003. Meanwhile, the person-prescription prevalence of individual LAIs remained low (<10%) (figure 3B). Overall, the person-prescription prevalence for any LAI within 1 year post first-admission discharge increased from < 10% in 2001 to 23% in 2016 (figure 3C; supplementary table S7).
Fig. 3.
(A) Person-prescription prevalence of the top 3 most prescribed antipsychotics between 2001 and 2016 (each year’s top 3 resulting in a total of 6 antipsychotics during the study period); (B) person-prescription prevalence of long-acting injections (LAIs) between 2001 and 2016; and (C) person-prescription prevalence of oral antipsychotics alone, LAIs alone, and both. Person-prescription refers to the antipsychotics prescribed to a patient within 1 y post first discharge. Thus, the person-prescription prevalence was calculated as the percentage of patients who received a specific antipsychotic during the year.
The duration from discharge date to the time a patient received a first prescription for any LAIs had a mean length of 1031 days (SD = 1265) and a median length of 499 days (25th percentile = 55 days and 75th percentile = 1590 days), ie, nearly one and a half years after the first discharge.
As only patients experiencing a readmission are included in the within-individual analyses, we compared the antipsychotic formulations’ prescription pattern between patients having readmissions and patients not readmitted within 4 years. No evidence showed that patients not being readmitted tend to use oral antipsychotics more, and higher prevalence of using both oral antipsychotics was found in these 2 groups (supplementary figure S7 and table S8).
Discussion
To our knowledge, this is the largest study of first-admission schizophrenia patients in a real-world comparative effectiveness analysis of antipsychotics. We were able to estimate the 4-year risk of readmission over sequential years and compare the adjusted risk of readmission for different antipsychotics using within-individual extended Cox regression analysis. From 2001 to 2013, the patients’ 4-year readmission risk diminished slightly from 68.12% to 64.35%. Compared to the most frequently prescribed oral risperidone, monotherapy with LAIs showed a reduction of 15%–20% in the adjusted risk of readmission. Nevertheless, the prescription of LAIs during the study period remained low in both person-years percentage and person-prescription prevalence. These findings have important implications for the prevention of readmission for schizophrenia patients following first admission.
The findings of a steady drop in the yearly number of first-admission schizophrenia patients between 2001 and 2017 but a relatively stable 4-year readmission risk between 2001 and 2013 extend the earlier trends identified from 1998 to 2007.18,19 This indicates that approximately 32–36% of first-admission schizophrenia patients did not need readmission for at least 4 years post-discharge. These findings are compatible with past research showing that approximately 40% of first-episode patients would endure remission in the first 2–3 postacute years.20 Nevertheless, one caveat in using readmission as the primary outcome in effectiveness analysis is that the criteria for readmission might be changed over time, particularly in the face of decreasing bed occupancy rate. This might underestimate the effectiveness of antipsychotics in preventing readmission.
In our analysis, we did not consider patients that are not readmitted, who might be well-controlled under the use of oral antipsychotics. However, we found that the prevalence of oral formulations from 2001 to 2013 did not differ significantly between readmitted and non-readmitted patients, ranging from 45% to 75%. The higher rates of using both oral antipsychotics and LAIs in readmitted patients revealed that the treatment was likely to be changed into LAIs when patients had a disease progression in a real-world scenario.
Effectiveness analysis showed that 3 LAIs (haloperidol, flupentixol, and risperidone) exhibited better effectiveness than oral risperidone in reducing the readmission risk. Previous studies with smaller sample sizes in Taiwan did not use a within-individual approach in comparing LAIs with oral antipsychotics; however, they showed conflicting results.21,22 Our findings are in line with previous studies that apply within-individual comparisons.9–11 The main reason for the better effectiveness of LAIs is apparently due to notably decreased nonadherence, which has been associated with elevated risks for readmissions23 and accounted for approximately 40% of the costs of hospitalization.24
Intriguingly, paliperidone palmitate, which was introduced in Taiwan in 2011 and contributed only 0.3% to the follow-up person-years with antipsychotics in this study, did not show better effectiveness in preventing readmission than oral risperidone. One possibility is that the prescription of paliperidone palmitate is relatively complicated. The approved, recommended initiation regimen for paliperidone is 150 mg equivalent (paliperidone palmitate 234 mg) on Day 1 and 100 mg equivalent (paliperidone palmitate 156 mg) on Day 8 and then a monthly maintenance injection with flexible dosing (25, 50, or 100 mg equivalent).25 However, for patients already on once-biweekly LAI of risperidone, a switch to paliperidone palmitate can be started with a maintenance dose. A recent evaluation of the prescription for paliperidone palmitate in a hospital in Taiwan found that for patients originally on oral antipsychotics only 44% of them received an initiation regimen,26 which might not attain therapeutically effective plasma levels.
Among oral antipsychotics, 6 had comparable effectiveness, 2 had slightly better effectiveness, and 5 had poorer effectiveness than oral risperidone. In previous studies, clozapine showed better effectiveness than oral olanzapine in preventing psychiatric readmission among schizophrenia patients in Sweden,9 and it showed better effectiveness than no use of any antipsychotic among schizophrenia patients in both Sweden and Finland.9, 10 Because clozapine has remained the gold standard for treatment-resistant schizophrenia but requires strict hematological monitoring, it is generally not used as a first-line treatment for schizophrenia. Nevertheless, the person-prescription prevalence of clozapine in Taiwan increased from 3.83% in 2001 to 7.75% in 2016. Intriguingly, countries with greater prevalence of clozapine use, such as Finland and Sweden (in a comparison of 17 different countries15), also reported better effectiveness of clozapine in preventing readmission than the most commonly prescribed olanzapine.9,10 Whether more extensive use of clozapine in first-admission schizophrenia patients could lead to better effectiveness over time warrants further investigation.
Our results further indicate that most combinations of polypharmacy were inferior to oral risperidone in preventing readmissions. This might result from the adverse effects of concomitant use of different antipsychotics, since polypharmacy was associated with an increased global side effect burden.27 Therefore, for first-admission schizophrenia patients, polypharmacy should be the last resort treatment option after monotherapy. Polypharmacy might also imply treatment resistance of the patients.
In Taiwan, during the study period, the prescription of antipsychotics, however, did not mirror the effectiveness in preventing readmission. Despite their better effectiveness in preventing readmission and the increasing trend in the person-prescription prevalence of LAIs over the years, the overall percentage of person-years contributed by LAIs remained low (4.2%), and the time interval between discharge and the patient receiving an LAI remained long (a median length of 1.5 y). On the other hand, oral quetiapine, one of the top 3 most frequently prescribed antipsychotics between 2001 and 2016, showed poorer effectiveness in preventing readmissions than oral risperidone. Even after adjusting for the drug’s use in treating insomnia, the use of quetiapine still increased the risk for readmission by 24%. Our results support a previous suggestion that quetiapine should not be used as routine monotherapy merely due to its lower side effect profile.28
Taken together, the 4-year readmission risk for first-admission schizophrenia patients remains high, and LAIs are more effective in preventing readmission, as revealed in our real-world analysis. These results indicate that one way to further reduce readmission risk for first-admission schizophrenia patients is to expand the coverage of LAIs early in the course of inpatient care. However, as in many other Asian countries,16 LAIs remain underused in Taiwan. Barriers to the use of LAIs might stem from psychiatrists’ perspectives (eg, believing patients’ preference for oral formulations, lack of confidence in prescribing LAIs, and outdated treatment guidelines), patient perspectives (eg, perceiving LAIs as coercive, fearing needles or injection pain, and lacking knowledge about LAIs), and system barriers (eg, the higher cost of SGA LAIs).29 In the context of this study, one strategy to overcome these barriers is to start an LAI on admission, when a patient can receive peer support and have more time to absorb medication knowledge. A recent evidence-based consensus statement by Taiwanese experts on LAIs may further help overcome these barriers.30
This study has limitations. First, information related to symptoms was not provided in the claims database, while we used time since first admission and times of antipsychotics changes to indicate the severity. Second, although the severity of illness was reset to the same level when a patient was discharged and with every subsequent resetting over the follow-up period, there remains the possibility of residual confounding associated with the passing of calendar time. Third, the current analysis ignores those who were well controlled with antipsychotics and who were not readmitted. Hence, our results might underestimate the effectiveness of antipsychotics. Fourth, since 49% of our patients received antipsychotics treatment within 1 year prior to the first admission, these patients were not “truly” incident in terms of psychotic symptoms. Finally, our results on the comparative effectiveness of specific antipsychotics might be generalized only to those countries in which all antipsychotic treatments are reimbursed by the insurance system.
In conclusion, based on national cohorts of first-admission schizophrenia patients in Taiwan from 2001 to 2017, the 4-year readmission risk for these patients remained relatively stable, at approximately 68.12%–64.35%. Compared to the most commonly prescribed oral risperidone, monotherapy with 3 LAIs (haloperidol, flupentixol, and risperidone) showed a 15%–20% reduction in the adjusted risk of readmission, due to better therapy adherence with LAIs. Despite their better effectiveness in preventing readmission, LAIs remain underused. These results indicate that one way to reduce readmission risk for first-admission schizophrenia patients is to begin coverage with LAIs earlier in treatment.
Supplementary Material
Acknowledgments
W.J.C. and Y.S.L. were responsible for the concept. W.J.C. led the study, and Y.S.L. did the data analysis and validated the result. W.J.C. and Y.S.L. wrote the first draft. W.J.C., C.S.W., C.C.L., P.H.K., H.Y.C., and Y.S.L. interpreted the results. All authors had the final responsibility for the decision to submit for publication. The authors have declared that there are no conflicts of interest in relation to the subject of this study.
Contributor Information
Yi-Hsuan Lin, Institute of Epidemiology and Preventive Medicine, College of Public Health, National Taiwan University, Taipei, Taiwan.
Chi-Shin Wu, National Center for Geriatrics and Welfare Research, National Health Research Institutes, Miaoli, Taiwan.
Chen-Chung Liu, Department of Psychiatry, College of Medicine and National Taiwan University Hospital, National Taiwan University, Taipei, Taiwan.
Po-Hsiu Kuo, Institute of Epidemiology and Preventive Medicine, College of Public Health, National Taiwan University, Taipei, Taiwan; Department of Psychiatry, College of Medicine and National Taiwan University Hospital, National Taiwan University, Taipei, Taiwan; Department of Public Health, College of Public Health, National Taiwan University, Taipei, Taiwan.
Hung-Yu Chan, Department of Psychiatry, College of Medicine and National Taiwan University Hospital, National Taiwan University, Taipei, Taiwan; Office of Superintendent, Taoyuan Psychiatric Center, Ministry of Health and Welfare, Taoyuan City, Taiwan.
Wei J Chen, Institute of Epidemiology and Preventive Medicine, College of Public Health, National Taiwan University, Taipei, Taiwan; Department of Psychiatry, College of Medicine and National Taiwan University Hospital, National Taiwan University, Taipei, Taiwan; Department of Public Health, College of Public Health, National Taiwan University, Taipei, Taiwan; Center for Neuropsychiatric Research, National Health Research Institutes, Miaoli, Taiwan.
Funding
This work was supported by grants from National Health Research Institutes, Taiwan (09A1-PP10 and 10A1-PP01), and Ministry of Science and Technology, Taiwan (109-2314-B-002-172-MY3). The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.
Data Availability
Data collected for this study are proprietary of the Health and Welfare Data Science Center at Ministry of Health and Welfare, which granted researchers permission and access to the data. The data that support findings of this study are available from these authorities, but restrictions apply to the availability of these data. The code used to analyze these data is available upon request from the corresponding author, for purposes of reproducing the authors.
References
- 1. GBD. 2016 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 328 diseases and injuries for 195 countries, 1990-2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet 2017;390(10100):1211–1259. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Moreno-Küstner B, Martín C, Pastor L. Prevalence of psychotic disorders and its association with methodological issues. A systematic review and meta-analyses. PLoS One. 2018;13(4):e0195687. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Goff DC, Falkai P, Fleischhacker WW, et al. The long-term effects of antipsychotic medication on clinical course in schizophrenia. Am J Psychiatry. 2017;174(9):840–849. [DOI] [PubMed] [Google Scholar]
- 4. Leucht S, Cipriani A, Spineli L, et al. Comparative efficacy and tolerability of 15 antipsychotic drugs in schizophrenia: a multiple-treatments meta-analysis. Lancet 2013;382(9896):951–962. [DOI] [PubMed] [Google Scholar]
- 5. Kishimoto T, Robenzadeh A, Leucht C, et al. Long-acting injectable vs oral antipsychotics for relapse prevention in schizophrenia: a meta-analysis of randomized trials. Schizophr Bull. 2014;40(1):192–213. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Kishimoto T, Hagi K, Kurokawa S, Kane JM, Correll CU. Long-acting injectable versus oral antipsychotics for the maintenance treatment of schizophrenia: a systematic review and comparative meta-analysis of randomised, cohort, and pre-post studies. Lancet Psychiatry 2021;8(5):387–404. [DOI] [PubMed] [Google Scholar]
- 7. Kishimoto T, Hagi K, Nitta M, et al. Effectiveness of long-acting injectable vs oral antipsychotics in patients with schizophrenia: A meta-analysis of prospective and retrospective cohort studies. Schizophr Bull. 2018;44(3):603–619. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Lichtenstein P, Halldner L, Zetterqvist J, et al. Medication for attention deficit-hyperactivity disorder and criminality. N Engl J Med. 2012;367(21):2006–2014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Tiihonen J, Mittendorfer-Rutz E, Majak M, et al. Real-world effectiveness of antipsychotic treatments in a nationwide cohort of 29823 patients with schizophrenia. JAMA Psychiatry 2017;74(7):686–693. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Taipale H, Mehtala J, Tanskanen A, Tiihonen J. Comparative effectiveness of antipsychotic drugs for rehospitalization in schizophrenia—a nationwide study with 20-year follow-up. Schizophr Bull. 2018;44(6):1381–1387. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Joo SW, Kim H, Jo YT, Choi YJ, Ahn S, Lee J. Antipsychotic treatment and risk of discontinuation and hospitalization in first-episode schizophrenia: a nationwide population-based study. Psychol Med. 2021. doi: 10.1017/s0033291721001379. (Online ahead of print). [DOI] [PubMed] [Google Scholar]
- 12. Lian L, Kim DD, Procyshyn RM, et al. Efficacy of long-acting injectable versus oral antipsychotic drugs in early psychosis: A systematic review and meta-analysis. Early Interv Psychiatry 2021. doi: 10.1111/eip.13202. (Online ahead of print). [DOI] [PubMed] [Google Scholar]
- 13. Roberts T, Miguel Esponda G, Krupchanka D, et al. Factors associated with health service utilisation for common mental disorders: a systematic review. BMC Psychiatry 2018;18(1):262. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Molden E, Jukić MM. CYP2D6 reduced function variants and genotype/phenotype translations of CYP2D6 intermediate metabolizers: implications for personalized drug dosing in psychiatry. Front Pharmacol. 2021;12:650750. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Bachmann CJ, Aagaard L, Bernardo M, et al. International trends in clozapine use: a study in 17 countries. Acta Psychiatr Scand. 2017;136(1):37–51. [DOI] [PubMed] [Google Scholar]
- 16. Tang CT, Chua EC, Chew QH, et al. Patterns of long acting injectable antipsychotic use and associated clinical factors in schizophrenia among 15 Asian countries and region. Asia Pac Psychiatry 2020;12(4):e12393. [DOI] [PubMed] [Google Scholar]
- 17. Hsieh CY, Su CC, Shao SC, et al. Taiwan’s National Health Insurance Research Database: past and future. Clin Epidemiol 2019;11:349–358. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Chiang C-L, Chen P-C, Huang L-Y, et al. Time trends in first admission rates for schizophrenia and other psychotic disorders in Taiwan, 1998-2007: a 10-year population-based cohort study. Soc Psychiatry Psychiatr Epidemiol. 2017;52(2):163–173. [DOI] [PubMed] [Google Scholar]
- 19. Chiang C-L, Chen P-C, Huang L-Y, et al. Impact of universal health coverage on urban–rural inequity in psychiatric service utilisation for patients with first admission for psychosis: a 10-year nationwide population-based study in Taiwan. BMJ Open 2016;6(3):e010802. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Gaebel W, Riesbeck M, Wölwer W, et al. Rates and predictors of remission in first-episode schizophrenia within 1 year of antipsychotic maintenance treatment. Results of a randomized controlled trial within the German Research Network on Schizophrenia. Schizophr Res. 2014;152(2-3):478–486. [DOI] [PubMed] [Google Scholar]
- 21. Huang SS, Lin CH, Loh el W, Yang HY, Chan CH, Lan TH. Antipsychotic formulation and one-year rehospitalization of schizophrenia patients: a population-based cohort study. Psychiatr Serv. 2013;64(12):1259–1262. [DOI] [PubMed] [Google Scholar]
- 22. Ju PC, Chou FH, Lai TJ, et al. Long-acting injectables and risk for rehospitalization among patients with schizophrenia in the home care program in Taiwan. J Clin Psychopharmacol. 2014;34(1):23–29. [DOI] [PubMed] [Google Scholar]
- 23. Diaz E, Levine HB, Sullivan MC, et al. Use of the Medication Event Monitoring System to estimate medication compliance in patients with schizophrenia. J Psychiatry Neurosci. 2001;26(4):325–329. [PMC free article] [PubMed] [Google Scholar]
- 24. Weiden PJ, Olfson M. Cost of relapse in schizophrenia. Schizophr Bull. 1995;21(3):419–429. [DOI] [PubMed] [Google Scholar]
- 25. Samtani MN, Gopal S, Gassmann-Mayer C, Alphs L, Palumbo JM. Dosing and switching strategies for paliperidone palmitate: based on population pharmacokinetic modelling and clinical trial data. CNS Drugs 2011;25(10):829–845. [DOI] [PubMed] [Google Scholar]
- 26. Chang C-C, Kuo L-H, Sie T-Y. Paliperidone palmitate use evaluation. The Journal of Taiwan Pharmacy 2020;36(3):127–132. [Google Scholar]
- 27. Gallego JA, Nielsen J, De Hert M, Kane JM, Correll CU. Safety and tolerability of antipsychotic polypharmacy. Expert Opin Drug Saf. 2012;11(4):527–542. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Vanasse A, Blais L, Courteau J, et al. Comparative effectiveness and safety of antipsychotic drugs in schizophrenia treatment: a real-world observational study. Acta Psychiatr Scand. 2016;134(5):374–384. [DOI] [PubMed] [Google Scholar]
- 29. Lindenmayer JP, Glick ID, Talreja H, Underriner M. Persistent barriers to the use of long-acting injectable antipsychotics for the treatment of schizophrenia. J Clin Psychopharmacol. 2020;40(4):346–349. [DOI] [PubMed] [Google Scholar]
- 30. Yang KC, Liao YT, Yang YK, et al. Evidence-based expert consensus regarding long-acting injectable antipsychotics for schizophrenia from the Taiwanese Society of Biological Psychiatry and Neuropsychopharmacology (TSBPN). CNS Drugs 2021;35(8):893–905. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data collected for this study are proprietary of the Health and Welfare Data Science Center at Ministry of Health and Welfare, which granted researchers permission and access to the data. The data that support findings of this study are available from these authorities, but restrictions apply to the availability of these data. The code used to analyze these data is available upon request from the corresponding author, for purposes of reproducing the authors.