Economic Evaluation Alongside Clinical Trials (EEACT) - Vocabulary Flashcards

Vocabulary flashcards covering key concepts, terms, and methods from the EEACT lecture notes.

EEACT

Economic Evaluation Alongside Clinical Trials; economic data are prospectively collected from patients within a controlled trial and analyzed alongside the clinical evaluation, with costs and outcomes estimated at the individual patient level.

Decision-Analytical Modelling

A modelling approach using published, grey, and unpublished information to estimate costs and effects; complementary to trials and can inform trial design and data collection.

Pragmatic (Naturalistic) Trial

A trial designed to assess effectiveness in usual clinical practice, retaining randomization but with fewer recruitment/follow-up restrictions; aims for generalisability and real-world applicability.

Explanatory (Efficacy) Trial

A trial designed to test whether an intervention can work under ideal, controlled conditions; more internal validity but less generalisability.

Internal Validity

The extent to which trial results are free from bias and confounding, ensuring that observed effects are due to the intervention.

External Validity

The generalisability of trial results to other settings, populations, and times.

Intention-to-Treat (ITT) Analysis

Analysis of all randomized participants in their original groups, regardless of adherence or protocol deviations, preserving randomization benefits.

Protocol-Driven Costs

Costs and resource use dictated by the trial protocol rather than routine practice, potentially limiting external validity.

Compliance

Adherence to the assigned intervention; differences in compliance within a trial versus usual practice can bias results.

Endoints in Pragmatic Trials

A broad set of outcomes including efficacy, feasibility, quality of life, and resource use, often using administrative data.

Skewed Data

Distributions of costs and utility data that are right-skewed; affects mean-based comparisons and uncertainty estimation.

Censored Data

Data with incomplete information due to study design or follow-up, including right, left, and interval censoring.

Bootstrap

A resampling method used to estimate uncertainty (e.g., CIs) for means and for ICERs when data are skewed or sample sizes are small.

ICER (Incremental Cost-Effectiveness Ratio

The ratio of the difference in costs to the difference in effects between two interventions; problematic when negative or when the denominator is zero.

Cost-Effectiveness Acceptability Curve (CEAC)

A curve showing the probability that an intervention is cost-effective across different willingness-to-pay thresholds.

Net Monetary Benefit (NMB)

A monetary valuation of health outcomes minus costs, offering an alternative to ICER that is easier to interpret for decision-makers.

Missing Data (MCAR, MAR, MNAR)

Types of missingness: MCAR (missing completely at random), MAR (missing at random, related to observed data), MNAR (not missing at random, related to unobserved data).

Missing Completely at Random (MCAR)

Missingness is independent of any data; complete-case analyses remain unbiased in expectation.

Missing at Random (MAR)

Missingness is related to observed data and can be predicted from them; imputation is possible with observed information.

Not Missing At Random (MNAR)

Missingness is related to unobserved data; more challenging to handle and requires sensitivity analysis.

Missing Data Imputation (LOCF, Regression, Multiple Imputation)

Techniques to fill in missing values: LOCF (single), regression-based, or multiple imputation (several datasets, then combined results) to reflect uncertainty.

Multivariate Regression Analysis

A statistical approach to model relationships between costs/effects and covariates, used to handle skewness and adjust for confounders.

Imputation versus Complete Case Analysis

Imputation fills in missing values; complete-case uses only observed data; imputation reduces bias when data are MAR.

Extrapolation

Extending trial findings beyond the trial duration or to a different population, using trial data and/or large observational data.

Life Years Gained

Additional years of life attributed to an intervention, often calculated from survival curves and life tables.

Discounting

Applying a time preference to reduce future costs and health benefits to their present value in economic evaluation.

Life Tables (qx, lx, Lx, ex)

Tables of mortality (qx) and survivors (lx) used to compute life expectancy (ex) and related metrics like years lived (Lx).

Survival Analysis Models (Cox, Weibull)

Statistical models used to analyze time-to-event data and extrapolate life expectancy for extrapolation beyond the trial.

Extrapolation Sources (Trial Data vs Observational Data)

Using either trial data or large retrospective/prospective observational datasets to inform long-term outcomes.

Multi-Centre, Multi-Country Trials

Large trials pooling data across centers/countries; advantages include larger samples and generalisability; challenges include heterogeneity and logistical coordination.

Stakeholders Buy-In

Essential cooperation among sponsors, researchers, and patients to ensure data quality and successful economic evaluation.