MIMIC-IV is a publicly available electronic health record (EHR) dataset derived from the Beth Israel Deaconess Medical Center. It contains information on patient measurements, orders, diagnoses, procedures, treatments, and deidentified free-text clinical notes. The dataset is intended to support a wide range of research and educational activities, helping to reduce barriers to clinical research. MIMIC-IV is contemporary, containing data from 2008–2019, and incorporates new precise digital information sources such as the electronic medicine administration record. It is organized in a modular structure, allowing linkage to external departments and data modalities. MIMIC-IV was developed through a collaboration between Beth Israel Deaconess Medical Center (BIDMC) and the Massachusetts Institute of Technology (MIT). Data collected at BIDMC is deidentified, transformed, and made available to researchers who have completed training in human research and signed a data use agreement. The Institutional Review Board at BIDMC granted a waiver of informed consent and approved the sharing of the research resource. The creation of MIMIC-IV involved three main steps: acquisition, transformation, and deidentification. Data was acquired from BIDMC and external sources, including clinical information systems and reference systems such as Diagnosis Related Groups (DRGs), the International Classification of Diseases (ICD), and the Healthcare Common Procedure Coding System (HCPCS). Data was transformed using custom SQL scripts to align with MIMIC-III and ensure data consistency. Deidentification was performed using a combination of rule-based approaches and a neural network to remove protected health information. MIMIC-IV contains three main modules: hosp, icu, and note. The hosp module includes information on patient transfers, billing, medication, laboratory values, and provider orders. The icu module includes data from ICU bedside documentation, such as intravenous infusions, patient outputs, and procedures. The note module includes deidentified free-text clinical notes, such as discharge summaries and radiology reports. MIMIC-IV is available via PhysioNet and includes detailed documentation and code repositories for researchers. It is designed to support a wide range of research and educational activities, including clinical research, education, and technology development. The dataset is freely accessible and has been validated for data integrity, consistency, and deidentification. It is intended to complement existing publicly available critical care datasets and provide a contemporary, comprehensive resource for researchers.MIMIC-IV is a publicly available electronic health record (EHR) dataset derived from the Beth Israel Deaconess Medical Center. It contains information on patient measurements, orders, diagnoses, procedures, treatments, and deidentified free-text clinical notes. The dataset is intended to support a wide range of research and educational activities, helping to reduce barriers to clinical research. MIMIC-IV is contemporary, containing data from 2008–2019, and incorporates new precise digital information sources such as the electronic medicine administration record. It is organized in a modular structure, allowing linkage to external departments and data modalities. MIMIC-IV was developed through a collaboration between Beth Israel Deaconess Medical Center (BIDMC) and the Massachusetts Institute of Technology (MIT). Data collected at BIDMC is deidentified, transformed, and made available to researchers who have completed training in human research and signed a data use agreement. The Institutional Review Board at BIDMC granted a waiver of informed consent and approved the sharing of the research resource. The creation of MIMIC-IV involved three main steps: acquisition, transformation, and deidentification. Data was acquired from BIDMC and external sources, including clinical information systems and reference systems such as Diagnosis Related Groups (DRGs), the International Classification of Diseases (ICD), and the Healthcare Common Procedure Coding System (HCPCS). Data was transformed using custom SQL scripts to align with MIMIC-III and ensure data consistency. Deidentification was performed using a combination of rule-based approaches and a neural network to remove protected health information. MIMIC-IV contains three main modules: hosp, icu, and note. The hosp module includes information on patient transfers, billing, medication, laboratory values, and provider orders. The icu module includes data from ICU bedside documentation, such as intravenous infusions, patient outputs, and procedures. The note module includes deidentified free-text clinical notes, such as discharge summaries and radiology reports. MIMIC-IV is available via PhysioNet and includes detailed documentation and code repositories for researchers. It is designed to support a wide range of research and educational activities, including clinical research, education, and technology development. The dataset is freely accessible and has been validated for data integrity, consistency, and deidentification. It is intended to complement existing publicly available critical care datasets and provide a contemporary, comprehensive resource for researchers.