Missing Data
Unobserved or unavailable values required for a specified analysis, whose consequences depend on the pattern, mechanism and target estimand.
Encyclopaedia
Explore comprehensive, evidence-informed explanations of key health economics concepts, including their development, application and relationships to other concepts. Published entries are validated through human expert review.
Concepts, entries and ontology terms
A concept page is a published entry on the wiki. Some concepts also have a full encyclopaedia article. The Health Economics Ontology (HEO) holds every concept, including dictionary terms without an article, so it is larger than the list of published pages. The knowledge graph shows how concepts relate to each other.
Missing Not at Random
Missing not at random describes a missing-data mechanism in which the probability that a value is unobserved depends on that unobserved value even after conditioning on the available observed information.
Mixed Effects Model
A mixed effects model is a regression with fixed and random effects, used in health economics for repeated or clustered data such as costs and utilities.
Modified Rankin Scale
Modified Rankin Scale (mRS) is a clinician-rated 0 to 6 grade of disability after stroke, used as a trial outcome and to define stroke model health states.
Multicollinearity
Multicollinearity is near or exact linear dependence among the predictors in a cost, utility or risk regression, making their separate effects imprecise.
Multiple Imputation
Multiple imputation fills missing costs or QALYs several times with plausible draws, analyses each completed dataset and pools results by Rubin's rules.
Multiple Regression
Multiple regression models one outcome, such as patient cost or QALYs, on two or more explanatory variables, adjusting each coefficient for the others.
Multivariate Meta-Analysis
Multivariate meta-analysis pools study results on two or more correlated outcomes, such as progression-free and overall survival, in a single joint model.
N-of-1 Trial
An N-of-1 trial is a randomised multiple crossover trial in one patient, comparing treatments over repeated periods to guide that patient's treatment.
Narrative Review
A narrative review is a literature review that summarises research on a topic without a prespecified, reproducible search, selection and appraisal method.
Negative Binomial
The negative binomial distribution models counts, such as hospital admissions per patient, whose variance exceeds the mean because rates vary by patient.
Negative Binomial Regression
Negative binomial regression models counts of health care use, such as admissions or exacerbations, whose variance is larger than a Poisson model allows.
Negative Likelihood Ratio
Negative Likelihood Ratio (LR-): (1 minus sensitivity) over specificity, the factor by which a negative test multiplies disease odds in decision models.
Negative Predictive Value
Negative predictive value (NPV) is the share of people with a negative test result who truly lack the condition, which models use to count missed cases.
Negative Study
A negative study is one whose main result shows no statistically significant or clinically important effect, which is not proof that no effect exists.
Network Meta-Analysis
Network meta-analysis (NMA) compares three or more treatments in one analysis, pooling direct and indirect trial evidence to inform HTA decisions.
Network Meta-Analysis HTA
The application of network meta-analysis specifically within health technology assessment, generating comparative effectiveness estimates when treatments were never studied together.
Node-Splitting
Node-splitting is a network meta-analysis inconsistency check that estimates one comparison separately from direct and indirect evidence and compares them.
Non-Inferiority Trial
A non-inferiority trial tests whether a new treatment is worse than an active comparator by less than a preset margin, often because it costs less.
Non-Randomised Controlled Trial
A non-randomised controlled trial is a trial where investigators assign people to treatment or control without chance, so its estimates risk confounding.