| Adjudication |
Review of evidence, such as a genetic report, by a qualified person who records a conclusion |
| Athena |
OHDSI's free online browser for the OMOP standardized vocabularies |
| Analysis plan |
The document, agreed before the analysis, that states the question, the cohort, the methods, and how missing data will be handled |
| ATLAS |
OHDSI's free web tool for defining cohorts and running analyses on OMOP data |
| Basic, clinical, and translational research |
Basic research studies mechanisms in cells, tissues, or models; clinical research studies people; translational research connects the two |
| Biomarker |
A defined characteristic measured as an indicator of normal biological processes, disease processes, or responses to an exposure or intervention |
| Case definition |
The evidence that qualifies a person as having the condition, often in tiers |
| Chain of custody |
The record of who collected a sample, when, where it is stored, where it went, and which participant it belongs to |
| Clinical research coordinator |
The person who screens, consents, follows up with, and supports participants, and keeps the study's records |
| Clinician-scientist |
A physician who conducts research and also sees patients |
| Cohort definition |
A set of rules, run against a database, that identifies the people who meet given criteria |
| Common data element |
A standard question with defined response options, designed to be asked the same way across studies |
| Common data model |
A shared structure for organizing data so that the same query runs at every site; OMOP is one |
| Complete-case analysis |
An analysis that uses only participants with no missing values for the variables involved |
| Computable phenotype |
A definition of a condition or trait expressed as rules that can be run against data |
| Concept set |
A list of standard concepts, often a concept with its descendants and explicit exclusions, used to define a cohort or a variable |
| Consent, broad and specific |
Broad consent covers future research not yet defined; specific consent covers named uses only |
| Contact registry |
A registry that records who people are and how to reach them, with consent to be contacted about news and other studies |
| Context of use |
In FDA qualification, "the manner and purpose of use" for a drug development tool |
| Covered entity |
Under HIPAA, a health plan, health care clearinghouse, or health care provider that conducts certain electronic transactions |
| Crosswalk |
A table that maps one identifier to another, such as a registry identifier to each lab's or health system's identifier |
| Data access committee |
The group that reviews requests for a registry's data and decides whether, and on what terms, to grant them |
| Data custodian |
The person or organization legally responsible for a data set |
| Data dictionary |
The document that defines every variable: its name, meaning, type, allowed values, and source |
| Data use agreement |
A contract setting the terms under which data are shared |
| Decentralized participation |
Taking part in research without traveling to a study site |
| Descendants |
In a vocabulary hierarchy, the concepts below a given concept |
| Enrolled, confirmed, evaluable |
Enrolled: consented and baseline entered. Confirmed: meets the case definition at a stated tier. Evaluable: has the data a given analysis needs |
| Fit for purpose |
Data collected and documented to the standard their intended use requires |
| Harmonize |
Map values that mean the same thing, such as "Yes," "1," and "TRUE," to one coded value, keeping the source value and recording the rule |
| Lantern Registry |
The book's hypothetical practice registry, with synthetic data for an invented condition, Epilepsy Syndrome X, caused by variants in an invented gene, GENEX; its dates are set in the 2040s |
| Likert scale |
A rating scale of ordered responses, such as strongly disagree to strongly agree |
| Missing completely at random, at random, not at random |
Assumptions about why data are missing: unrelated to any data; related only to data that were observed; or related to the missing values themselves |
| Multiple imputation |
Replacing each missing value with several plausible values, analyzing each completed data set, and combining the results |
| Natural history study |
A study that follows people with a condition over time to describe how it progresses |
| Network study |
A study in which several sites run the same protocol against their own data in a common data model and share results rather than records |
| Phenotype |
The observable characteristics of a condition in a person |
| Placeholder value |
A value such as 999 or -1 entered to stand for missing or unknown, which a tool treats as a real number unless told otherwise |
| Provenance |
The documented history of where a value came from and every change made to it |
| Qualification |
In FDA usage, "a conclusion that within the stated context of use, the DDT can be relied upon" |
| Query |
A question about a value that may be wrong or unclear, sent to the person who can answer it; the value is not changed until the answer comes back |
| Real world data and real world evidence |
Data on health status or health care collected routinely outside clinical trials, and the evidence derived from analyzing them |
| Reference range |
The range of values a laboratory considers typical for a test, specific to the lab and assay |
| Reverse coding |
Flipping the scale of an item worded in the opposite direction so that all items score the same way |
| Secure research environment |
An access-controlled computing environment where approved researchers analyze data without downloading it |
| Semi-structured interview |
A conversation guided by set topics, from which answers can be coded |
| Standard concept |
In the OMOP vocabularies, the concept chosen to represent a meaning, to which other codes map |
| Stratification |
Dividing participants into subgroups, such as by progression or symptom, for analysis or trial enrollment |
| Sub-study |
An additional study within a registry, usually with its own questions or population |
| Table 1 |
The first table in most clinical papers, describing the people in the study |
| Tier |
A level of diagnostic evidence within a case definition, such as a genetic report, a clinician letter, or self-report |
| Validated instrument |
A tested measure with a known scoring rule, administered as its developers intended |
| Variant of uncertain significance |
A genetic variant whose effect on health is not yet known; its classification may change |
| Version control |
Keeping every change to code or documents with a history of who changed what and when |