Sensitivity analysis is a statistical procedure used to verify the robustness of the results of a clinical trial or analytical model against changes in assumptions, alternative methods, or the handling of missing data. It answers the question of whether the conclusions of the primary analysis remain stable when key methodological decisions are varied. Sensitivity analyses are a mandatory component of the Statistical Analysis Plan (SAP) and are required by regulatory authorities. Their importance has increased significantly with the introduction of ICH E9(R1), as the updated guideline explicitly focuses on the handling of missing data and the robustness of the primary analysis. Today, no marketing authorization dossier for a confirmatory tested endpoint is acceptable without appropriate sensitivity analyses. <\/p>\n
The primary analysis of a clinical trial is based on pre-defined assumptions regarding the analysis population (e.g., Intent-to-treat vs. Per-protocol), the handling of missing data (e.g., Multiple Imputation vs. Last Observation Carried Forward), model specification, and the treatment of protocol deviations. Sensitivity analyses systematically vary individual assumptions to check whether the results remain consistent. <\/p>\n
Unlike the primary analysis and pre-specified subgroup analyses, sensitivity analyses do not serve to provide confirmatory proof of an effect, but rather to evaluate methodological robustness. They do not provide new hypotheses but instead strengthen or weaken confidence in the primary evaluation. For this reason, their results are classified as contextual evidence within the framework of EMA assessments and HTA procedures. A decisive quality feature of sensitivity analyses is their complete pre-specification in the Statistical Analysis Plan (SAP)\u2014at the latest before database lock and unblinding. Only pre-specified analyses can be considered part of the confirmatory analysis strategy; all others are classified as post-hoc and exploratory. <\/p>\n
Sensitivity analyses are used in clinical trials in various contexts. The most common use case is the handling of missing data: according to ICH E9(R1), the SAP for the primary endpoint must contain at least one sensitivity analysis that tests alternative assumptions regarding the missing data mechanism (MCAR, MAR, MNAR). Typical methods include pattern-mixture models, tipping-point analyses, and worst-case imputations. <\/p>\n
Other types of sensitivity analyses include: variation of the analysis population (e.g., comparison of ITT vs. PP population), alternative statistical models (e.g., parametric vs. non-parametric tests), exclusion of centers with high protocol deviation rates, and checking the influence of outliers. In meta-analyses and systematic reviews, sensitivity analyses examine how results react when individual studies or subgroups are excluded. Furthermore, sensitivity analyses are used in pharmacoeconomic models to verify the robustness of cost-benefit estimates against uncertain model parameters\u2014an area of application that is becoming increasingly important in the context of HTA procedures. <\/p>\n
ICH E9(R1) contains specific requirements for sensitivity analyses for the primary endpoint in the case of missing data. The EMA expects the SAP to contain a clear description of all planned sensitivity analyses, including the rationale for the choice of alternative assumptions. Sensitivity analyses must be defined in advance in the SAP; sensitivity analyses performed post-hoc are critically evaluated by authorities and may be viewed as an indication of a results-oriented approach. <\/p>\n
In the Clinical Study Report (CSR) according to ICH E3, the results of all pre-specified sensitivity analyses must be fully presented and interpreted. GCP inspectors check during the review of the SAP and CSR whether sensitivity analyses were performed consistently and reported transparently. Deviations between the planned and actually performed analysis scope are considered protocol deviations. <\/p>\n
Well-planned sensitivity analyses strengthen the credibility of clinical trials with authorities, HTA bodies, and the scientific community. If the primary analysis and all sensitivity analyses provide consistent results, the robustness of the treatment effect is proven. Conversely, diverging results indicate methodological uncertainties that must be discussed transparently in the evaluation process. Studies in which sensitivity analyses lead to significantly different conclusions may be subject to conditions or requests for additional information from authorities before a positive marketing authorization decision is granted. <\/p>\n
For full-service CROs like mediconomics, the planning of sensitivity analyses involves close cooperation between biostatisticians, clinical scientists, and regulatory experts as early as the study design phase. Sensitivity analyses planned early and anchored in the SAP avoid discussions with authorities regarding methodological weaknesses and contribute to the quality of the entire clinical dossier. <\/p>\n
How many sensitivity analyses are required in a clinical trial?<\/strong><\/p>\n There is no fixed number. At least one sensitivity analysis for the primary endpoint regarding the handling of missing data is required by regulation (ICH E9(R1)). Beyond that, the number of sensitivity analyses should correspond to the risk profile of the study: trials with high dropout rates, heterogeneous populations, or complex designs require more sensitivity analyses than simple trials with complete data. All planned analyses must be documented in the SAP. <\/p>\n What is the difference between sensitivity analysis and subgroup analysis?<\/strong><\/p>\n A sensitivity analysis varies methodological assumptions of the overall analysis (e.g., imputation method, analysis population) to test the robustness of the overall result. A subgroup analysis divides the total population into subgroups (e.g., by age, gender, biomarker status) to identify effect-modifying factors. Both are pre-specified but serve different purposes and are interpreted differently from a regulatory perspective. <\/p>\n What happens if sensitivity analyses yield different results than the primary analysis?<\/strong><\/p>\n Divergent results in sensitivity analyses must be discussed and interpreted transparently in the CSR. They may indicate uncertainties in the data, methodological dependencies, or genuine heterogeneity. Authorities such as the EMA expect a substantial discussion of these discrepancies in the marketing authorization dossier. Depending on the extent of the deviation, the regulatory assessment of the primary endpoint result may be restricted or linked to conditions. <\/p>\n","protected":false},"excerpt":{"rendered":" Sensitivity analysis is a statistical procedure used to verify the robustness of the results of a clinical trial or analytical model against changes in assumptions, alternative methods, or the handling of missing data. It answers the question of whether the conclusions of the primary analysis remain stable when key methodological decisions are varied. Sensitivity analyses […]<\/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-6749","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\/6749","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":0,"href":"https:\/\/mediconomics.com\/en\/wp-json\/wp\/v2\/glossary\/6749\/revisions"}],"wp:attachment":[{"href":"https:\/\/mediconomics.com\/en\/wp-json\/wp\/v2\/media?parent=6749"}],"wp:term":[{"taxonomy":"glossary-cat","embeddable":true,"href":"https:\/\/mediconomics.com\/en\/wp-json\/wp\/v2\/glossary-cat?post=6749"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}