Target Trial Emulation

Uses

Preparation

Where it comes from

Highlighted claims

• The final real-world cohort included 6,114 adults, with 2,392 GLP-1RA initiators and 3,722 SGLT-2i initiators. — Joint Treatment Effect Estimation from Incomplete Healthcare Data: Temporal Causal Normalizing Flows with LLM-driven Evolutionary MNAR Imputation
• The real-world emulation used cohort entry as first-ever GLP-1RA or SGLT-2i prescription as first second-line glucose-lowering therapy from January 1, 2015 to December 7, 2024. — Joint Treatment Effect Estimation from Incomplete Healthcare Data: Temporal Causal Normalizing Flows with LLM-driven Evolutionary MNAR Imputation
• The real-world study followed patients per protocol until treatment deviation, comparator initiation, death, or 390 days. — Joint Treatment Effect Estimation from Incomplete Healthcare Data: Temporal Causal Normalizing Flows with LLM-driven Evolutionary MNAR Imputation
• The outcome in the real-world target trial emulation was absolute body-weight change over one year in 30-day intervals. — Joint Treatment Effect Estimation from Incomplete Healthcare Data: Temporal Causal Normalizing Flows with LLM-driven Evolutionary MNAR Imputation
• Target trial emulation is used to study causal questions with observational electronic health records when randomized trials are infeasible. — Joint Treatment Effect Estimation from Incomplete Healthcare Data: Temporal Causal Normalizing Flows with LLM-driven Evolutionary MNAR Imputation