HRIP flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

Data governance vs. stewardship

Governance sets decision rights, policy, and accountability for data; stewardship executes those decisions day to day within a domain, resolving quality issues and enforcing standards.

Data owner accountability

The role accountable for a data domain's quality, access decisions, and standard compliance. Owners approve access and changes; stewards implement them operationally.

Minimum necessary standard

HIPAA rule limiting PHI use and disclosure to the least amount needed for the purpose. It does not apply to treatment, disclosures to the individual, or provider payment.

Enterprise master patient index

System that links patient records across source systems by probabilistic or deterministic matching, assigning an enterprise identifier to prevent duplicate and fragmented records.

Data dictionary required fields

Each entry documents variable name, definition, data type, permissible values, source system, responsible owner, and sensitivity classification so downstream users interpret data consistently.

CDISC SDTM purpose

Standard structure for organizing tabulated clinical trial data by observation class for regulatory submission, enabling consistent review across studies and sponsors.

LOINC vs. SNOMED CT

LOINC codes the question being asked, such as a lab test or observation. SNOMED CT codes clinical findings, procedures, and concepts as answers and clinical content.

HL7 v2 versus FHIR

HL7 v2 uses delimited segment messages like ADT and ORU. FHIR uses RESTful resources with JSON or XML, supporting modern APIs and granular queries.

Query management lifecycle

A data query is raised, answered by site staff, evaluated, then resolved or closed. Each step is timestamped and audited to preserve data integrity.

Intention-to-treat principle

Analysis includes all randomized participants in their assigned group regardless of adherence. It preserves randomization and gives a conservative, real-world effect estimate.

Confounding versus effect modification

Confounding is a distortion from a third variable to be adjusted away. Effect modification is a genuine difference in treatment effect across subgroups that should be reported.

Phase III trial purpose

Confirmatory comparison of an intervention against a control in a larger population to establish efficacy and safety for regulatory approval, typically randomized and blinded.

Sample size determinants

Required sample size depends on expected effect size, outcome variance, alpha, desired power, and anticipated attrition. Smaller effects and lower power needs increase size.

STROBE versus CONSORT

STROBE is a reporting checklist for observational studies. CONSORT guides reporting of randomized trials, including a participant flow diagram.

p-value definition

The probability of observing data at least as extreme as obtained, assuming the null hypothesis is true. It is not the probability the null is true.

Power and Type II error

Power is the probability of detecting a true effect, equal to 1 minus beta. Beta is the Type II error rate, a false negative.

Number needed to treat

NNT equals one divided by the absolute risk reduction. It states how many patients must be treated for one additional favorable outcome.

Positive predictive value dependence

PPV is the probability a positive test is a true positive. Unlike sensitivity and specificity, it changes with disease prevalence in the population tested.

Missing data mechanisms

MCAR means missingness is random; MAR means it depends on observed variables and multiple imputation applies; MNAR means it depends on unobserved values and needs sensitivity analysis.

HIPAA Safe Harbor method

De-identification by removing 18 specified identifier categories, including dates more specific than year, full ZIP codes, and biometric identifiers.

HIPAA breach notification clock

Covered entities must notify affected individuals without unreasonable delay and no later than 60 days after discovering a breach of unsecured PHI.

GDPR controller versus processor

The controller determines purposes and means of processing personal data. The processor handles data only on the controller's documented instructions and bears separate obligations.

21 CFR Part 11 requirements

Regulation requiring validated systems, secure access controls, audit trails, and attributable electronic signatures for electronic records submitted to or used by FDA.

Common Rule Subparts B and D

Subpart B adds protections for pregnant women, fetuses, and neonates. Subpart D adds protections for children, generally requiring parental permission and child assent.

Six data quality dimensions

Completeness, accuracy, consistency, timeliness, validity, and uniqueness. Each dimension needs its own metric, since a dataset can score well on one and poorly on another.

Audit trail requirements

Records must capture who changed a value, what changed, when, and why, while retaining the original entry so it can be recovered rather than overwritten.

Probabilistic record matching

Linkage method assigning agreement weights across fields such as name, date of birth, and identifier, then applying thresholds to classify pairs as matches or non-matches.

Source data verification

Comparison of transcribed CRF values against original source documents. Discrepancy rates above a preset threshold trigger retraining or increased monitoring.

Chart type selection rule

Use bars for comparison, histograms or box plots for distribution, lines for trends over time, scatter plots for relationships, and stacked bars for composition.

Truncated axis distortion

Starting a bar chart's y-axis above zero exaggerates differences between categories. Bar charts must start at zero; line charts may truncate if labeled.

Absolute versus relative risk

Relative risk expresses proportional change; absolute risk gives the actual event-rate difference. Reporting relative risk alone overstates small absolute effects.

CONSORT flow diagram

Figure tracking participants through screening, randomization, allocation, follow-up, and analysis, showing exclusions and losses at each stage of a trial.