Dictionary
The Dictionary provides concise definitions of health economics terms, arranged alphabetically for quick reference. Use it to understand unfamiliar terminology or confirm the meaning of a specific term.
A
- Aggregation Bias
- A distortion arising when relationships observed in group-level aggregated data do not accurately reflect the corresponding relationships at the individual level.
- Allocation Concealment
- A safeguard in randomised trials ensuring the assignment sequence is unknown to those enrolling participants until after enrolment, preventing selection bias.
- Anchoring Bias in Elicitation
- The tendency for an expert's numerical judgement in a structured elicitation to be disproportionately swayed by an initial reference value shown to them.
- Anchoring Concept
- A cognitive bias in which an initial piece of information disproportionately influences subsequent judgements, even when arbitrary.
B
- Begg Test
- A statistical test for detecting publication bias in a meta-analysis, based on the rank correlation between effect size and its variance.
- Bias
- A systematic, rather than random, deviation of a study's results from the true underlying value, arising from flaws in design, collection, or analysis.
- Bias-Variance Tradeoff
- The tension in statistical modelling between a model's systematic error, or bias, and its sensitivity to the specific sample used to fit it, or variance.
C
- Cognitive Bias Concept
- A systematic deviation from rational judgement caused by the mental shortcuts people use to process information.
- Copas Model
- A statistical model adjusting a meta-analysis result for potential publication bias by modelling how publication likelihood depends on results.
D
- Digital Safety
- The degree to which a digital health technology avoids causing unintended harm, covering both clinical safety and data security or algorithmic bias.
E
- Ecological Bias
- A distortion arising when group-level associations are wrongly assumed to apply to individuals within those groups, potentially reversing the true relationship.
- Effect Modification
- A situation where an exposure's effect on an outcome genuinely differs across levels of a third variable, unlike confounding, which is a bias.
- Egger Test
- A statistical test for funnel plot asymmetry, and by extension potential publication bias, using regression of effect estimates against standard errors.
F
- Funnel Plot
- A scatter plot assessing potential publication bias by plotting each study's effect estimate against its precision, expecting symmetry if unbiased.
- Funnel Plot Analysis
- The formal examination of a funnel plot's symmetry, often with a statistical test such as Egger's, to assess likely publication bias.
H
- Hedges' G
- A standardised effect size measure similar to Cohen's d but corrected for small sample size bias.
- Hypothetical Bias
- The tendency for willingness-to-pay values stated in a hypothetical survey to overstate what respondents would actually pay in a real transaction.
I
- Immortal Time Bias
- A bias where a period of follow-up is misclassified so a participant cannot, by design, experience the outcome during it, inflating apparent benefit.
- Information Bias
- A category of bias from errors in how exposure, outcome, or covariate data are measured or recorded, including recall bias and measurement error.
L
- Lead Time Bias
- A bias in screening evaluations where earlier diagnosis appears to extend survival but only lengthens awareness of diagnosis, without delaying underlying disease progression.
- Length Bias
- A bias in screening evaluations arising because screening preferentially detects slower-growing cases that stay detectable longer, while aggressive cases are more often missed.
- Level of Evidence
- A classification ranking a study design's strength based on its susceptibility to bias, placing randomised trials above case reports or opinion.
M
- Measurement Error Methods
- Statistical techniques correcting for bias arising when a model variable is measured with error rather than reflecting its true value exactly.
N
- Non-Response Error
- A survey error arising when selected participants fail to respond, and respondents differ systematically from non-respondents, potentially biasing the results.
O
- Outcome Reporting Bias
- Bias arising when the outcomes reported in a published study are selected based on the direction or significance of their results, not as originally planned.
- Overconfidence
- A cognitive bias in which individuals, including expert forecasters, express excessive certainty, producing judgements narrower than their actual accuracy would justify.
P
- Present Bias Concept
- The tendency to place disproportionate weight on immediate rewards or costs relative to future ones, beyond what constant discounting predicts.
- Pro-Rich Bias
- A pattern in which the benefits of a healthcare subsidy or programme disproportionately accrue to wealthier groups relative to their share of need.
- Publication Bias
- A systematic distortion in published evidence arising because studies with significant or favourable results are more likely to be published.
- Publication Bias Assessment
- An evaluation of whether published evidence is systematically distorted because significant or favourable results are more likely to be published than null ones.
Q
- Quality Assessment
- The general process of systematically evaluating a study's methodological rigour and trustworthiness, covering design, risk of bias, and reporting completeness.
- Quality Assessment Model
- A structured tool or checklist systematically evaluating research methodological quality against predefined criteria, such as the Cochrane risk of bias tool.
- Quality of Evidence
- A judgement about how much confidence can be placed in an effect estimate, based on study design, risk of bias, and precision.
R
- Random Error
- Variation in results attributable to chance alone, from the statistical variability of sampling, unlike systematic error, which reflects a consistent directional bias.
- Recall Bias
- A form of information bias where participants with a particular outcome recall past exposures differently, often more thoroughly, than those without it.
- Reporting Bias
- A broad category of bias arising when selective reporting of research findings, at study or outcome level, systematically distorts available evidence.
- Reporting Standards
- The requirements governing how research findings, positive and negative, must be documented and disclosed, to prevent selective reporting bias.
- Risk of Bias
- The degree to which a study's design, conduct, or analysis may have introduced systematic errors distorting its findings from the truth.
- Risk of Bias Assessment
- A structured evaluation of a study's methodological quality, examining domains such as randomisation and blinding, to judge how much confidence its results deserve.
S
- Selection Bias
- Bias arising when the individuals included in a study differ systematically from the broader population of interest, in ways linked to exposure and outcome.
- Small Study Effects
- A pattern in meta-analysis where smaller studies tend to report systematically larger treatment effects than larger studies, possibly reflecting publication bias.
- Status Quo Bias Concept
- The tendency to prefer the current state of affairs over available alternatives, even when switching would be beneficial.
T
- Trim And Fill
- A method detecting and adjusting for potential publication bias by imputing hypothetical missing studies that would restore funnel plot symmetry.
U
- Unbiasedness
- A property of an estimator whose expected value, across repeated samples, equals the true underlying value of the parameter being estimated.