Beyond "Intent-to-treat" and "Per protocol": Improving assessment of treatment effects in clinical trials through the specification of an estimand.
Ontology highlight
ABSTRACT:
SUBMITTER: Petavy F
PROVIDER: S-EPMC7319001 | biostudies-literature | 2020 Jul
REPOSITORIES: biostudies-literature
Publications
Beyond "Intent-to-treat" and "Per protocol": Improving assessment of treatment effects in clinical trials through the specification of an estimand.
Pétavy Frank F Guizzaro Lorenzo L Antunes Dos Reis Inês I Teerenstra Steven S Roes Kit C B KCB
British journal of clinical pharmacology 20200327 7
There is a key problem in randomised clinical trials as outcomes can be distorted due to informative post-randomisation events. This is inadequately addressed by the use of traditional intention-to-treat or per protocol analysis sets and often either ignored or wrongly labelled as missing data. As a consequence, the treatment effects of interest in a clinical trial are not well defined and their estimates might be misinterpreted. The estimand framework should help all those planning, conducting ...[more]
Similar Datasets
Project description:BackgroundCluster randomized trials are designed to evaluate interventions at the cluster or group level. When clusters are randomized but some clusters report no or non-analyzable data, intent-to-treat analysis, the gold standard for the analysis of randomized controlled trials, can be compromised. This article presents a very flexible statistical methodology for cluster randomized trials whose outcome is a cluster-level proportion (e.g. proportion from a cluster reporting an event) in the setting where clusters report non-analyzable data (which in general could be due to nonadherence, dropout, missingness, etc.). The approach is motivated by a previously published stratified randomized controlled trial called, "The Randomized Recruitment Intervention Trial (RECRUIT)," designed to examine the effectiveness of a trust-based continuous quality improvement intervention on increasing minority recruitment into clinical trials (ClinicalTrials.gov Identifier: NCT01911208).MethodsThe novel approach exploits the use of generalized estimating equations for cluster-level reports, such that all clusters randomized at baseline are able to be analyzed, and intervention effects are presented as risk ratios. Simulation studies under different outcome missingness scenarios and a variety of intra-cluster correlations are conducted. A comparative analysis of the method with imputation and per protocol approaches for RECRUIT is presented.ResultsSimulation results show the novel approach produces unbiased and efficient estimates of the intervention effect that maintain the nominal type I error rate. Application to RECRUIT shows similar effect sizes when compared to the imputation and per protocol approach.ConclusionThe article demonstrates that an innovative bivariate generalized estimating equations framework allows one to implement an intent-to-treat analysis to obtain risk ratios or odds ratios, for a variety of cluster randomized designs.
| S-EPMC9497422 | biostudies-literature
Project description:BackgroundIn non-inferiority trials, there is a concern that intention-to-treat (ITT) analysis, by including participants who did not receive the planned interventions, may bias towards making the treatment and control arms look similar and lead to mistaken claims of non-inferiority. In contrast, per protocol (PP) analysis is viewed as less likely to make this mistake and therefore preferable in non-inferiority trials. In a systematic review of antibiotic non-inferiority trials, we compared ITT and PP analyses to determine which analysis was more conservative.MethodsIn a secondary analysis of a systematic review, we included non-inferiority trials that compared different antibiotic regimens, used absolute risk reduction (ARR) as the main outcome and reported both ITT and PP analyses. All estimates and confidence intervals (CIs) were oriented so that a negative ARR favored the control arm, and a positive ARR favored the treatment arm. We compared ITT to PP analyses results. The more conservative analysis between ITT and PP analyses was defined as the one having a more negative lower CI limit.ResultsThe analysis included 164 comparisons from 154 studies. In terms of the ARR, ITT analysis yielded the more conservative point estimate and lower CI limit in 83 (50.6%) and 92 (56.1%) comparisons respectively. The lower CI limits in ITT analysis favored the control arm more than in PP analysis (median of - 7.5% vs. -6.9%, p = 0.0402). CIs were slightly wider in ITT analyses than in PP analyses (median of 13.3% vs. 12.4%, p < 0.0001). The median success rate was 89% (interquartile range IQR 82 to 93%) in the PP population and 44% (IQR 23 to 60%) in the patients who were included in the ITT population but excluded from the PP population (p < 0.0001).ConclusionsContrary to common belief, ITT analysis was more conservative than PP analysis in the majority of antibiotic non-inferiority trials. The lower treatment success rate in the ITT analysis led to a larger variance and wider CI, resulting in a more conservative lower CI limit. ITT analysis should be mandatory and considered as either the primary or co-primary analysis for non-inferiority trials.Trial registrationPROSPERO registration number CRD42020165040 .
| S-EPMC8054385 | biostudies-literature
Project description:Under exceptional circumstances, including high rates of protocol non-compliance, per-protocol (PP) analysis can better indicate the real-world benefits of a medical intervention than intention-to-treat (ITT) analysis. Exemplifying this, the first randomized clinical trial (RCT) considered found that colonoscopy screenings were marginally beneficial, based upon ITT analysis, with only 42% of the intervention group actually undergoing the procedure. However, the study authors themselves concluded that the medical efficacy of that screening was a 50% reduction in colorectal cancer deaths among that 42% PP group. The second RCT found a ten-fold reduction in mortality for a COVID-19 treatment drug vs. placebo by PP analysis, but only a minor benefit by ITT analysis. The third RCT, conducted as an arm of the same platform trial as the second RCT, tested another COVID-19 treatment drug and reported no significant benefit by ITT analysis. Inconsistencies and irregularities in the reporting of protocol compliance for this study required consideration of PP outcomes for deaths and hospitalizations, yet the study coauthors refused to disclose them, instead directing inquiring scientists to a data repository which never held the study's data. These three RCTs illustrate conditions under which PP outcomes may differ significantly from ITT outcomes and the need for data transparency when these reported or indicated discrepancies arise.
| S-EPMC10253430 | biostudies-literature
Project description:Promoting child wellbeing necessarily goes beyond the clinic as risks to child health and development are embedded in the social and physical environmental conditions in which children live. Pediatricians play a vital role in promoting the health of children in the communities they serve and can maximize their impact by advocating for and supporting efficacious, evidence-based strategies in their communities.To provide a succinct guide for community pediatric efforts to advance the wellbeing of all children and particularly disadvantaged children in a community, we conducted a theory-driven and structured narrative review to synthesize published systematic and meta-analytic reviews of policy-relevant, local-level strategies addressing potent and malleable influences on child health and development. An exhaustive list of policy-relevant, local-level strategies for improving child health was used to conduct a comprehensive search of recent (1990-2012), English language peer-reviewed published meta-analyses and systematic reviews in the 10 core databases of scientific literature. Our review of the literature encompassed six key conceptual domains of intervention foci, including distal influences of child health (i.e., income and resources, social cohesion, and physical environment) and proximal influences (i.e., family, school and peer). We examined intervention effects on four key domains of child health and development: cognitive development, social and emotional competence, psychological and behavioral wellbeing, and physical health.Published reviews were identified for 98 distinct policy-relevant community interventions, evaluated across 288 outcomes. We classified 46 strategies as meeting scientific criteria for efficacy by having consistent, positive outcomes from high-quality trials (e.g., tenant-based rental assistance, neighborhood watch programs, urban design and land use policies, access to quality childcare services, class size reductions, after-school programs that promote personal/social skills). Another 21 strategies were classified as having consistent evidence of positive outcomes from high-quality observational studies only, while 28 strategies had insufficient evidence available to assess their effectiveness based on published reviews. We did not limit the review to studies conducted in the United States, but the vast majority of them were U.S.-based, and the results therefore are most applicable to the U.S. context.Based on our synthesis of published literature on community development strategies, we provide an illustration combining a comprehensive set of evidence-based strategies to promote child health and development across a wide-range of child health outcomes.
| S-EPMC4016148 | biostudies-other
Project description:Background In event-driven clinical trials, study termination is based on accrual of a target number of primary efficacy events. For noninferiority trials in which superiority is conditionally examined, the ideal cohort in which to track event accrual is unclear. We used data from the ROCKET AF (Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared With Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation) trial to determine the effect of primary efficacy-event tracking in the per-protocol cohort during the on-treatment period versus the intention-to-treat (ITT) cohort during the ITT period. Methods and Results ROCKET AF was terminated after accruing 429 primary efficacy events (stroke or systemic embolism) in the per-protocol cohort during the on-treatment period for noninferiority. We identified the date on which 429 events occurred in the ITT cohort during the ITT period. We performed noninferiority and superiority analyses based on hypothetical study termination on this date. ROCKET AF would have terminated 226 days earlier if events were tracked during the ITT period. Similar to the main trial findings, rivaroxaban would have met noninferiority versus warfarin for the primary efficacy end point (hazard ratio [HR], 0.77; 95% CI, 0.62-0.96; P<0.001). In contrast to the main trial findings, rivaroxaban would have met superiority for the primary efficacy end point (HR, 0.82; 95% CI, 0.68-0.99; P=0.038). In both termination scenarios, rivaroxaban was associated with a lower risk of intracranial hemorrhage and similar risk of other safety end points. Conclusions Clinical trial termination based on event accrual in the ITT cohort versus the per-protocol cohort may have important implications on trial results depending on rates of study drug discontinuation and event rates off treatment.
| S-EPMC8649140 | biostudies-literature
Project description:BACKGROUND:The per-protocol effect is the effect that would have been observed in a randomized trial had everybody followed the protocol. Though obtaining a valid point estimate for the per-protocol effect requires assumptions that are unverifiable and often implausible, lower and upper bounds for the per-protocol effect may be estimated under more plausible assumptions. Strategies for obtaining bounds, known as "partial identification" methods, are especially promising in randomized trials. RESULTS:We estimated bounds for the per-protocol effect of colorectal cancer screening in the Norwegian Colorectal Cancer Prevention trial, a randomized trial of one-time sigmoidoscopy screening in 98,792 men and women aged 50-64 years. The screening was not available to the control arm, while approximately two thirds of individuals in the treatment arm attended the screening. Study outcomes included colorectal cancer incidence and mortality over 10 years of follow-up. Without any assumptions, the data alone provide little information about the size of the effect. Under the assumption that randomization had no effect on the outcome except through screening, a point estimate for the risk under no screening and bounds for the risk under screening are achievable. Thus, the 10-year risk difference for colorectal cancer was estimated to be at least -0.6 % but less than 37.0 %. Bounds for the risk difference for colorectal cancer mortality (-0.2 to 37.4 %) and all-cause mortality (-5.1 to 32.6 %) had similar widths. These bounds appear helpful in quantifying the maximum possible effectiveness, but cannot rule out harm. By making further assumptions about the effect in the subpopulation who would not attend screening regardless of their randomization arm, narrower bounds can be achieved. CONCLUSIONS:Bounding the per-protocol effect under several sets of assumptions illuminates our reliance on unverifiable assumptions, highlights the range of effect sizes we are most confident in, and can sometimes demonstrate whether to expect certain subpopulations to receive more benefit or harm than others. TRIAL REGISTRATION:Clinicaltrials.gov identifier NCT00119912 (registered 6 July 2005).
| S-EPMC4666083 | biostudies-literature
Project description:Assessing per-protocol treatment effcacy on a time-to-event endpoint is a common objective of randomized clinical trials. The typical analysis uses the same method employed for the intention-to-treat analysis (e.g., standard survival analysis) applied to the subgroup meeting protocol adherence criteria. However, due to potential post-randomization selection bias, this analysis may mislead about treatment efficacy. Moreover, while there is extensive literature on methods for assessing causal treatment effects in compliers, these methods do not apply to a common class of trials where a) the primary objective compares survival curves, b) it is inconceivable to assign participants to be adherent and event-free before adherence is measured, and c) the exclusion restriction assumption fails to hold. HIV vaccine efficacy trials including the recent RV144 trial exemplify this class, because many primary endpoints (e.g., HIV infections) occur before adherence is measured, and nonadherent subjects who receive some of the planned immunizations may be partially protected. Therefore, we develop methods for assessing per-protocol treatment efficacy for this problem class, considering three causal estimands of interest. Because these estimands are not identifiable from the observable data, we develop nonparametric bounds and semiparametric sensitivity analysis methods that yield estimated ignorance and uncertainty intervals. The methods are applied to RV144.
| S-EPMC3811958 | biostudies-literature
Project description:ObjectivesFollowing well-established practices in demography, this article discusses several measures based on the number of COVID-19 deaths to facilitate comparisons over time and across populations.SettingsNational populations in 186 United Nations countries and territories and populations in first-level subnational administrative entities in Brazil, China, Italy, Mexico, Peru, Spain and the USA.ParticipantsNone (death statistics only).Primary and secondary outcome measuresAn unstandardised occurrence/exposure rate comparable to the Crude Death Rate; an indirectly age-and-sex standardised rate that can be derived even when the breakdown of COVID-19 deaths by age and sex required for direct standardisation is unavailable; the reduction in life expectancy at birth corresponding to the 2020 number of COVID-19 deaths.ResultsTo date, the highest unstandardised rate has been in New York, at its peak exceeding the state 2017 crude death rate. Populations compare differently after standardisation: while parts of Italy, Spain and the USA have the highest unstandardised rates, parts of Mexico and Peru have the highest standardised rates. For several populations with the necessary data by age and sex for direct standardisation, we show that direct and indirect standardisation yield similar results. US life expectancy is estimated to have declined this year by more than a year (-1.26 years), far more than during the worst year of the HIV epidemic, or the worst 3?years of the opioid crisis, and to reach its lowest level since 2008. Substantially larger reductions, exceeding 2?years, are estimated for Panama, Peru, and parts of Italy, Spain, the USA and especially, Mexico.ConclusionsWith lesser demand on data than direct standardisation, indirect standardisation is a valid alternative to adjust international comparisons for differences in population distribution by sex and age-groups. A number of populations have experienced reductions in 2020 life expectancies that are substantial by recent historical standards.
| S-EPMC7948156 | biostudies-literature
Project description:Following well-established practices in demography, this article discusses several measures based on the number of CoViD-19 deaths to facilitate comparisons over time and across populations. National populations in 186 UN countries and territories and populations in first-level sub-national administrative entities in Brazil, China, Italy, Mexico, Peru, Spain, and the USA. None (death statistics only). An unstandardized occurrence/exposure rate comparable to the Crude Death Rate; an indirectly age-and-sex standardized rate that can be derived even when the breakdown of CoViD-19 deaths by age and sex required for direct standardization is unavailable; the reduction in life expectancy at birth corresponding to the 2020 number of CoViD-19 deaths. To date, the highest unstandardized rate has been in New York, at its peak exceeding the state 2017 Crude Death Rate. Populations compare differently after standardization: while parts of Italy, Spain and the USA have the highest unstandardized rates, parts of Mexico and Peru have the highest standardized rates. For several populations with the necessary data by age and sex for direct standardization, we show that direct and indirect standardization yield similar results. US life expectancy is estimated to have declined this year by more than a year (-1.26 years), far more than during the worst year of the HIV epidemic, or the worst three years of the opioid crisis, and to reach its lowest level since 2008. Substantially larger reductions, exceeding two years, are estimated for Panama, Peru, and parts of Italy, Spain, the USA, and especially, Mexico. With lesser demand on data than direct standardization, indirect standardization is a valid alternative to adjust international comparisons for differences in population distribution by sex and age-groups. A number of populations have experienced reductions in 2020 life expectancies that are substantial by recent historical standards.
| S-EPMC7273293 | biostudies-literature
Project description:This SuperSeries is composed of the SubSeries listed below.
2016-08-24 | GSE79565 | GEO