思覺失調症首次住院率、再住院率、與抗精神病藥物效果比較：2001至2017年的臺灣全國世代研究

Abstract

背景及目的 在台灣，思覺失調症從1998到2007年首次住院率顯著下降，再住院卻相對穩定，近期鮮有研究討論隨著精神醫療資源可近性提高，住院率是否有近一步的改善趨勢或趨於穩定。過去雖有許多抗精神病藥物相關療效研究，卻受限於個體間干擾因子而難以反映藥物實際效果，而個體內Cox擴充風險迴歸分析能夠有效控制個體間藥物服從性不一及藥物反應異質性等問題。因此，本研究目的為，利用高覆蓋率的全國性健保資料，了解思覺失調症在2001至2017年間的首次住院率與再住院率趨勢、抗精神病藥物開立趨勢、並利用個體內比較的方法分析抗精神病藥物預防再住院的長期效果。 方法 利用全國精神疾病住院患者健康保險資料庫，篩選15-64歲首次精神科住院思覺失調症患者，共75,986位。計算2001至2017年間思覺失調症年齡調整後的首次住院率，並利用Kaplan-Meier存活分析計算不同年度的2年內、4年內再住院率。抗精神病藥物預防再住院效果的分析中，利用分層個體內Cox擴展風險模型 (stratified with-in individual extended Cox model)，隨著追蹤時間更新患者的藥物暴露，符合實際中藥物會隨患者狀況作調整的治療方式。並且，再住院為反覆發生的事件 (recurrent event)，本方法將每一次再住院納入分析。在每次出院嚴重度相近的假設下，疾病嚴重度會受到校正。每位患者為獨立的分層，使用理思必妥口服藥物(oral risperidone)的時期作為的治療對照組，計算個體內藥物間的再住院風險，合併各分層後，比較不同藥物對於預防再住院的平均療效。 結果 從2001到2017年間，思覺失調症首次住院率從每十萬人每年41.7名患者下降到15.6名患者，下降約62.5 %，而2年內、4年內再住院率則相對穩定，分別維持在50-55%及60-70%。在抗精神病藥物對於預防再住院的效果比較中，以理思必妥口服藥為基準，長效注射單一藥物治療的再住院風險下降15-20%。不論是因精神病再住院或因任何原因再住院，皆以氟哌啶醇長效注射劑(haloperidol LAIs, HR = 0.80; 95% CI: 0.73-0.88)、氟哌噻噸長效注射劑(flupentixol LAIs, 0.83; 95% CI: 0.77-0.89)及理思必妥長效注射劑(risperidone LAIs, HR = 0.84; 95% CI: 0.79-0.90)的預防再住院效果最佳。此外，在年紀較大及換藥次數較多時，抗精神病藥物預防再住院效果較差。 結論 在2001-2017年間，首次住院率持續下降，再住院率相對穩定。以分層個體內Cox擴展風險模型分析新的藥物效果，長效注射劑有效降低思覺失調症首次住院患者再住院的風險。

Background. In past decade, the mental healthcare service has been improved in Taiwan, whereas little research examining the changes in first-admission rates and readmission rates for schizophrenia. Antipsychotics are the gold standard treatment for schizophrenia; however, comparing effectiveness among antipsychotics might be biased by several between-individual confounders. A new method called within-individual extended Cox model could help reduce the biases. Hence, this study is aimed to (1) understand the trends in first admission rates and readmission rates from 2001 to 2017, (2) examine the pattern of antipsychotics prescription, and (3) evaluate the comparative effectiveness of antipsychotics for schizophrenia in reducing the risk for readmission with within-individual extended Cox regression analysis. Methods. We adopted a retrospective cohort study to examine the trends in both first admission rate and readmission rate from 2001 to 2017 with the National Psychiatric Inpatient Medical Claims Database (NPIMCD). Patients first admitted to psychiatric wards with the diagnosis of schizophrenia during the study period were selected as a dynamic incident cohort. To compare the risk for readmission between different antipsychotics, we applied the stratified within-individual extended Cox model. Antipsychotics exposure was updated with time by treating it as a time-dependent exposure, and the model integrate the effects on all readmissions. Each individual was an independent stratum, where the periods of oral risperidone use served as his/her own control to assess the comparing effectiveness on preventing readmission. Results and Discussion. During the period from 2001 to 2017, the rate of first admission to psychiatric wards (per 100,000 person-years) for schizophrenia decreased from 41.7 cases in 2001 to 15.6 cases in 2017, a reduction of 62.5%. The 2-year and 4-year readmission risk declined from 53.25% in 2001 to 50.33% in 2015 and from 68.16% in 2001 to 64.35% in 2013 respectively; both the readmission risks remained relatively stable. For the antipsychotics effectiveness, the risk for psychotic readmission was the lowest during LAI monotherapy, with a risk reduction of 15-20%. Among LAIs, haloperidol LAIs (LAIs, HR = 0.80; 95% CI: 0.73-0.88), flupentixol LAIs (HR = 0.83; 95% CI: 0.77-0.89), and risperidone LAIs (HR = 0.84; 95% CI: 0.79-0.90) had the lowest psychotic readmission risk. The results remained similar for the analysis of all-cause readmissions. Since patients became more adherent while using LAIs, the effectiveness of antipsychotics was significantly increased. Conclusions. First admission rates for schizophrenia have decreased steadily from 2001 to 2017, while the readmission rates remained stable during the study period. For antipsychotics effectiveness on preventing readmissions, we concluded that treatment with LAIs provide better prevention for readmission for schizophrenia patient who are admitted for the first time.