[Introduction]
-Schizophrenia has a prevalence of about 1% and is characterized by (1) a shortened average lifespan, (2) a high rate of suicide attempts, and (3) economic isolation and human rights violations due to social stigma.
-Early detection of the disease and on-going relapse management through medication are crucial, but studies show that the relapse rate is 82% after initial recovery from schizophrenia, and the risk of relapse increases 5 times if medication is discontinued.
-Medication treatment includes oral medications taken daily and long-acting injectable antipsychotics (LAI) that last for extended periods (e.g., 2, 4, 12, 24, 64 weeks). While LAI are useful for improving medication adherence, they are less preferred by patients.

[Aim]
-To compare the risk of medication discontinuation between oral medications and long-acting injectable antipsychotics (LAI) using claims data from Korea’s Health Insurance Review and Assessment Service (HIRA).
-Inferential causation, such as Propensity Score Matching (PSM) and Inverse Probability Weighting (IPW), is applied to Real-World data (claim data) to create an environment similar to that of randomized controlled trials, and the results are compared.

[Method]
-Data source
:This study utilized South Korea’s National Health Insurance (NHI) claims data from HIRA.
It is a retrospective observational study of outpatient schizophrenia patients aged 18 to under 65 who received medication treatment from 2020 to 2022.
-Outcome : Continuation or Discontinuation of medication treatment
-Covariates : Gender, Age, Residence, Medical expenses, Type of medical institution, Location of the medical institution, Characteristics of the medical institution (e.g., number of doctors and beds), Comorbidities (e.g., hypertension, diabetes, hyperlipidemia).
-Subject
: (Unadjusted Model) 9,646 patients were included(Oral n=9,467, LAIs n=179).
: (PSM) 358 patients were included using 1:1 nearest neighbor matching based on propensity scores(Oral n=179, LAIs n=179).
: (IPW) 9,454 patients were included after applying inverse probability weighting to the propensity scores and truncating weights below the 1st percentile and above the 99th percentile (Oral n=9,394, LAIs n=60).
-Statistical Analysis : Kaplan-Meier Survival Curve, Cox Proportional Hazard Model were implement with a significance level 5%.

[Results]
-Distribution of Propensity Score and SMD (Oral vs LAI)
-Comparison of Outcomes (Oral vs LAI)
  : In the Kaplan-Meier Survival Curve, the LAI group showed a lower risk of treatment discontinuation compared to the oral medication group across all models, and this was statistically significant.
  : As a result of applying the Cox Proportional Hazard Model, In the unadjusted model, the Hazard Ratio (HR) was 0.56, while in the PSM model, the HR was 0.61, and in the IPW model, it was 0.62. This indicates that the risk of treatment discontinuation in the LAI group is 0.56, 0.61, and 0.62 times lower, respectively, compared to the oral medication group.

[Conclusions]
-The analysis of the causal effect of treatment discontinuation risk (using PSM and IPW) in outpatients with schizophrenia revealed that long-acting injectable (LAI) antipsychotics are a more favorable choice than oral medications when aiming to prevent treatment discontinuation.
-Recently, South Korea has been expanding health insurance coverage for LAI antipsychotics to improve treatment adherence and continuity for schizophrenia patients.
-This study aligns with this policy direction by demonstrating that the risk of treatment discontinuation is lower with LAI compared to oral medications.

[ACKNOWLEDGEMENT]
-I would like to express my sincere gratitude to Professor Sohee Park of the Graduate School of Public Health at Yonsei University, my academic advisor.
-And I would also like to extend my heartfelt thanks to my beloved husband, our family and my in-laws for their constant support.
-Lastly, I am deeply grateful to my colleagues at HIRA who made it possible for me to participate in this ISQUA conference.