A confidence interval (CI) describes the range of values within which an unknown parameter of the population \u2013 for example, a mean reduction in blood pressure or a hazard ratio \u2013 can lie with a predefined level of confidence. In clinical trials, the CI is typically reported together with a point estimate because it visualizes both the direction and the precision of an effect estimate. For sponsors, CROs and assessment bodies, the confidence interval is thus often more meaningful than an isolated p-value.<\/p>\n
A 95% CI is constructed such that if a great many studies with an identical design were theoretically repeated, the calculated interval would cover the true parameter value in approximately 95% of these studies. Importantly, this does not mean that the true value lies in exactly this one interval “with 95% probability”; in classical frequentist statistics, the parameter is considered fixed, while the interval is random. In practice, however, the CI is used as a measure of uncertainty and serves the interpretation of clinical relevance.<\/p>\n
For relative measures such as odds ratio or hazard ratio, the CI is frequently calculated on the log scale and subsequently back-transformed. As a result, the intervals are usually asymmetrical. For continuous endpoints (e.g., change in a score), CIs are often approximately symmetrical if assumptions of normality are plausible or large samples are available.<\/p>\n
For many standard tests, the following applies: if a two-sided 95% confidence interval does not contain the null value (e.g., difference = 0 or ratio = 1), this corresponds to a p-value of less than 0.05. However, the CI additionally provides how large the effect can plausibly be: a narrow CI indicates a precise estimate; a wide CI indicates high uncertainty, for instance due to a small sample, high variability or rare events. Especially for safety endpoints, wide CIs are common and must be transparently addressed in the benefit-risk assessment.<\/p>\n
In regulatory dossiers and clinical study reports, therefore, frequently both the p-value and the CI are tabulated. For internal decision-making processes (go\/no-go, dose selection, trial continuation), CIs are particularly helpful because they depict the range of potential effects.<\/p>\n
In superiority trials, the CI is used to evaluate whether the data are compatible with a clinically relevant benefit. In non-inferiority trials, the comparison with the non-inferiority margin is the focus: it is crucial that the “worst” plausible effect (depending on the effect measure, the lower or upper CI limit) does not exceed the margin. Equivalence trials typically require that the entire CI lies within a predefined equivalence margin.<\/p>\n
This logic is closely linked to the statistical protocol (determination of alpha, one- or two-sided consideration, hierarchies). Changes to analysis windows, populations (ITT\/FAS vs. PP) or model assumptions can alter CIs and should be consistently documented in the statistical analysis plan and in the CSR.<\/p>\n
A frequent misinterpretation is the equation of “not significant” with “no effect”. A CI that includes the null value can still not exclude a clinically relevant effect if it is wide. Conversely, a very narrow CI can indeed be statistically significant, but only allow for a small, clinically scarcely relevant difference. Therefore, CIs should always be discussed in the context of minimal clinically important difference, endpoint definition and measurement precision.<\/p>\n
Further pitfalls are multiple comparisons and data-driven subgroups. Without appropriate adjustment, reported CIs can convey a too optimistic picture of precision. In practice, sensitivity analyses and robust variance estimators are therefore used to check the stability of the CIs.<\/p>\n
For project teams, it is also relevant how confidence intervals are interpreted in interplay with protocol deviations and missing data. If, for example, a model makes an assumption about missing follow-up, the CI can become significantly narrower or wider without the point estimate changing strongly. Therefore, sponsors should specify in the SAP which imputation or modeling approaches are primary and which serve as a sensitivity analysis.<\/p>\n
A practical tip for medical writing: do not merely describe whether the CI crosses the null value, but explain which effect sizes are plausibly supported or excluded by the data. This facilitates the discussion of clinical relevance, particularly if the trial was designed for a specific minimal difference.<\/p>\n
In adaptive designs or interim analyses, confidence intervals are frequently adjusted to the alpha-spending strategy. Then 95% CIs can be replaced by other confidence levels to control the overall error probability. These details should be consistently presented in the protocol and in the CSR so that traceability for authorities is maintained.<\/p>\n
Authorities and notified bodies expect a traceable presentation of effect sizes and uncertainties. This applies both to medicinal product trials under Regulation (EU) No 536\/2014 and to clinical investigations of medical devices under MDR 2017\/745. In the context of ICH E9 (statistical principles) and ICH E6(R3) (GCP modernization), transparent analysis and reporting chains are central, including the question of which confidence levels are used and how sensitivity analyses address the uncertainty.<\/p>\n
FAQ<\/strong><\/p>\n Why is a 95% confidence interval mostly used?<\/strong><\/p>\n 95% is a historically established convention level that fits well with a two-sided alpha of 0.05. Depending on the context (e.g., interim analyses or multiple endpoints), other levels can be reasonable.<\/p>\n Can a confidence interval be “wrong”?<\/strong><\/p>\n The interval follows from model assumptions. If these are violated (e.g., strong deviation from distributional assumptions, informative censoring), the CI can underestimate the uncertainty. Then alternative models or robust procedures are indicated.<\/p>\n How are CIs used for clinical relevance?<\/strong><\/p>\n The CI limits are compared with predefined thresholds (e.g., non-inferiority margin or clinically relevant difference). This makes it visible which effect sizes are compatible with the data.<\/p>\n Regulatory references:<\/strong> ICH E9 (Statistical Principles for Clinical Trials), ICH E6(R3) Guideline for Good Clinical Practice, Regulation (EU) No 536\/2014 (Clinical Trials Regulation), Regulation (EU) 2017\/745 (MDR).<\/p>\n","protected":false},"excerpt":{"rendered":" A confidence interval (CI) describes the range of values within which an unknown parameter of the population \u2013 for example, a mean reduction in blood pressure or a hazard ratio \u2013 can lie with a predefined level of confidence. In clinical trials, the CI is typically reported together with a point estimate because it visualizes […]<\/p>\n","protected":false},"author":10,"featured_media":0,"parent":0,"template":"","meta":{"_acf_changed":false,"site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"default","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","ast-disable-related-posts":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"default","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"footnotes":""},"glossary-cat":[],"class_list":["post-6711","glossary","type-glossary","status-publish","hentry"],"acf":[],"related_terms":"","external_url":"","internal_reference_id":"","_links":{"self":[{"href":"https:\/\/mediconomics.com\/en\/wp-json\/wp\/v2\/glossary\/6711","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/mediconomics.com\/en\/wp-json\/wp\/v2\/glossary"}],"about":[{"href":"https:\/\/mediconomics.com\/en\/wp-json\/wp\/v2\/types\/glossary"}],"author":[{"embeddable":true,"href":"https:\/\/mediconomics.com\/en\/wp-json\/wp\/v2\/users\/10"}],"version-history":[{"count":1,"href":"https:\/\/mediconomics.com\/en\/wp-json\/wp\/v2\/glossary\/6711\/revisions"}],"predecessor-version":[{"id":7498,"href":"https:\/\/mediconomics.com\/en\/wp-json\/wp\/v2\/glossary\/6711\/revisions\/7498"}],"wp:attachment":[{"href":"https:\/\/mediconomics.com\/en\/wp-json\/wp\/v2\/media?parent=6711"}],"wp:term":[{"taxonomy":"glossary-cat","embeddable":true,"href":"https:\/\/mediconomics.com\/en\/wp-json\/wp\/v2\/glossary-cat?post=6711"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}