OMOP Analytics and OHDSI Tools

These flashcards cover vocabulary and key concepts related to OMOP analytics, OHDSI tools like ATLAS and HADES, and the methodology for defining cohorts and characterization.

Characterization

The use of descriptive statistics to understand a population or data source by summarizing clinical events and how they change over time.

Population-level estimation

The quantification of average causal effects of exposures, such as medical interventions or procedures, on health outcomes using observational data.

Patient-level prediction

The process of building models to predict the probability that an individual patient will experience a specific outcome in a defined future window using past data.

Custom Code

Authoring analysis code from scratch in languages like R, SAS, or Python without using OHDSI tools to achieve maximum flexibility.

HADES

The Health Analytics Data-to-Evidence Suite; a collection of OHDSI R packages (formerly called the Methods Library) used for standardized analysis against the CDM.

SqlRender

A HADES package used to translate and execute the same SQL code across different database platforms like PostgreSQL, SQL Server, and Oracle.

DatabaseConnector

A HADES package providing the infrastructure to connect to various database platforms standardized to the CDM.

CohortMethod

A HADES package that provides advanced analytics capabilities for conducting population-level estimation against CDM data.

ATLAS

A free, web-based tool that provides a graphical interface to design and execute various analyses on CDM-standardized data without writing code.

Phenotype

A reproducible set of rules used to identify patients with a specific clinical characteristic using coded healthcare data in the OMOP CDM.

Cohort

A set of persons who satisfy one or more inclusion criteria for a specific duration of time; often used interchangeably with phenotype.

Concept set

A collection of standardized clinical concepts used as building blocks to identify events in the data; also referred to as a code set.

Initial event

The specific recorded event, such as a drug exposure or procedure, that defines the time of cohort entry.

Index date

The anchor event or date that defines time zero for each person in a cohort.

Inclusion criteria

The logic applied to a cohort to restrict membership based on data domains, concept sets, attributes, and temporal logic.

Exit criteria

Rules that define when a person stops being in a cohort, such as the end of observation or a fixed time window.

Baseline (pre-index)

The time period in a patient's medical history occurring before the index date.

Post-index

The time period in a patient's medical history occurring on or after the index date.

Time at Risk (TAR)

The specific period during which people are considered susceptible to an outcome and when events are counted.

Person-time

The total accumulated time at risk across individuals, typically measured in units like days, months, or years.

Treatment pathways

Descriptive statistics detailing the sequence of interventions a person receives during the post-index time.

Incidence proportion

The proportion of people in a population who experience a specific outcome during the recorded time at risk.

Incidence rate

The number of new outcomes measured per unit of total person-time.