Empirical evaluation of the implementation of the EMA guideline on missing data in confirmatory clinical trials: Specification of mixed models for longitudinal data in study protocols

Häckl, Sebastian; Koch, Armin; Lasch, Florian

doi:10.1002/pst.1964

Cited by 8 publications

(10 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The means of changes from baseline values were calculated and analyzed using a restricted maximum likelihood (REML)-based repeated measures approach combined with the Newton Raphson algorithm. The analytical model included fixed categorical effects of treatment, visit, and treatment-by-visit interaction [ 26 ]. Akaike Information Criteria were used to choose the working correlation structures for modeling the within‐patient errors among an unstructured covariance structure, autoregressive (1) structure, and a compound symmetry structure.…”

Section: Methodsmentioning

confidence: 99%

Randomized phase 2 study of perampanel for sporadic amyotrophic lateral sclerosis

et al. 2021

View full text Add to dashboard Cite

Objective To evaluate the efficacy and safety of perampanel in patients with sporadic amyotrophic lateral sclerosis (SALS). Methods This randomized, double-blind, placebo-controlled, multicenter, phase 2 clinical study was conducted at 12 sites. Patients with probable or definite ALS as defined by revised El Escorial criteria were enrolled. Sixty-six patients were randomly assigned (1:1:1) to receive placebo, 4 mg perampanel, or 8 mg perampanel daily for 48 weeks. Adverse events (AEs) were recorded throughout the trial period. The primary efficacy outcome was the change in Amyotrophic Lateral Sclerosis Rating Scale-Revised (ALSFRS-R) score after 48 weeks of treatment. Results One patient withdrew before starting the treatment. Of 65 patients included, 18 of 22 patients randomized to placebo (82%), 14 of 22 patients randomized to 4 mg perampanel (64%), and 7 of 21 patients randomized to 8 mg perampanel (33%) completed the trial. There was a significant difference in the change of ALSFRS-R scores [− 8.4 (95% CI − 13.9 to − 2.9); p = 0.015] between the placebo and the perampanel 8 mg group, primarily due to worsening of the bulbar subscore in the perampanel 8 mg group. Serious AEs were more frequent in the perampanel 8 mg group than in the placebo group (p = 0.0483). Conclusions Perampanel was associated with a significant decline in ALSFRS-R score and was linked to worsening of the bulbar subscore in the 8 mg group.

show abstract

Section: Methodsmentioning

confidence: 99%

Randomized phase 2 study of perampanel for sporadic amyotrophic lateral sclerosis

et al. 2021

View full text Add to dashboard Cite

show abstract

“…The EMA multiplicity guideline 2 highlights this problem of (unnoticed) selection of a favorable analysis from a variety of analysis options all in line with imprecise model specifications. While the theoretical multiplicity problem is known and addressed in the guideline, preliminary research shows that the compliance with the guideline is poor for a specific type of analysis models 23 …”

Section: Introductionmentioning

confidence: 99%

“…If the specification of a single model item is altered, different results may be produced. Häckl et al 23 have conducted an empirical evaluation of the specifications of MMRMs in study protocols submitted to the Ethics committee at Hannover Medical School. It could be revealed that not a single study protocol specified all necessary model items (0/39).…”

Section: Introductionmentioning

confidence: 99%

“…While sponsors from the top‐21 pharmaceutical companies, as defined by global sales in 2017, show a slightly better compliance, still only 53% (9/17) protocols specified the main items, whereas sponsors not from the top‐21 pharmaceutical companies specified the main items in 14% (3/22) only. In summary, Häckl et al 23 showed that for MMRMs the pre‐specification in study protocols is poor with the consequence of a potential intended or unintended inflation of the type‐I‐error rate. The discovery of this problem is especially relevant, as 62% (24/39) of the trials are phase III trials and intended to provide pivotal evidence for drug licensing decisions.…”

Section: Introductionmentioning

confidence: 99%

“…Building on the empirical findings from Häckl et al, 23 this article investigates the magnitude of the type‐I‐error rate inflation in different clusters of analysis methods for different trial characteristics. Focusing on the same class of analysis methods – mixed models for longitudinal data – we investigate the inflation of the type‐I‐error rate in the common clinical setting of a randomized controlled parallel group trial depending on (a) the type of unspecified model items, (b) the number of unspecified model items, (c) the sample size and (d) the randomisation ratio.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Type‐I‐error rate inflation in mixed models for repeated measures caused by ambiguous or incomplete model specifications

Häckl

Koch

Lasch

2023

Pharmaceutical Statistics

Self Cite

View full text Add to dashboard Cite

Pre‐specification of the primary analysis model is a pre‐requisite to control the family‐wise type‐I‐error rate (T1E) at the intended level in confirmatory clinical trials. However, mixed models for repeated measures (MMRM) have been shown to be poorly specified in study protocols. The magnitude of a resulting T1E rate inflation is still unknown. This investigation aims to quantify the magnitude of the T1E rate inflation depending on the type and number of unspecified model items as well as different trial characteristics. We simulated a randomized, double‐blind, parallel group, phase III clinical trial under the assumption that there is no treatment effect at any time point. The simulated data was analysed using different clusters, each including several MMRMs that are compatible with the imprecise pre‐specification of the MMRM. T1E rates for each cluster were estimated. A significant T1E rate inflation could be shown for ambiguous model specifications with a maximum T1E rate of 7.6% [7.1%; 8.1%]. The results show that the magnitude of the T1E rate inflation depends on the type and number of unspecified model items as well as the sample size and allocation ratio. The imprecise specification of nuisance parameters may not lead to a significant T1E rate inflation. However, the results of this simulation study rather underestimate the true T1E rate inflation. In conclusion, imprecise MMRM specifications may lead to a substantial inflation of the T1E rate and can damage the ability to generate confirmatory evidence in pivotal clinical trials.

show abstract

Telmisartan for treatment of Covid-19 patients: An open multicenter randomized clinical trial

et al. 2021

View full text Add to dashboard Cite

Background: Angiotensin receptor blockers (ARBs), such as telmisartan, have been postulated to treat Covid-19-induced lung inflammation. Methods: This is a parallel-group, randomized, two-arm, open-label, adaptive, multicenter superiority trial with 1:1 allocation ratio. Participants included patients from 18 years of age hospitalized with Covid-19 with 4 or fewer days since symptom onset enrolled at a university and a community hospital in Buenos Aires, Argentina. Exclusion criteria included prior intensive care unit (ICU) admission and use of ARBs/angiotensin converting enzyme inhibitors at randomization. Control arm received standard care alone and treatment arm telmisartan 80 mg twice daily for 14 days. Primary outcomes were C-reactive protein (CRP) plasma levels at day 5 and 8 after randomization. Secondary outcomes included time to discharge within 15 days, admission to ICU and death at 15-and 30-days. NCT04355936 (Completed). Findings: A pragmatic decision to end the study before the third interim analysis was made on Oct. 30th due to sharp reduction in recruitment. A total of 162 patients were randomized. 158 patients enrolled between May 14 and October 30 2020, were included in the analysis, 80 in the standard care and 78 in the telmisartan added to standard care group. Baseline absolute CRP serum levels were 5.53 § 6.19 mg/dL (95% CI 6.91 to 4.15, n = 80) and 9.04 § 7.69 (95% CI 9.04 to 10.82, n = 74) in the standard care and telmisartan added to standard care groups, respectively. Day 5 control-group CRP levels were 6.06 § 6.95 mg/dL (95% CI 7.79À4.35, n = 66) while telmisartan group were 3.83 § 5.08 mg/dL (95% CI 5.08À2.59, n = 66, p = 0.038). Day 8 CRP levels were 6.30 § 8.19 mg/dL (95% CI 8.79À3.81, n = 44) and 2.37 § 3.47 mg/dL (95% CI 3.44À1.30, n = 43, p = 0.0098) in the control and telmisartan groups, respectively (all values expressed as mean § SD). Kaplan-Meier analysis showed that telmisartan-treated patients had a lower median time-to-discharge (control=15 days; telmisartan=9 days). Death by day 30 was reduced in the telmisartan-treated group (control 22.54%,

show abstract

Empirical evaluation of the implementation of the EMA guideline on missing data in confirmatory clinical trials: Specification of mixed models for longitudinal data in study protocols

Cited by 8 publications

References 20 publications

Randomized phase 2 study of perampanel for sporadic amyotrophic lateral sclerosis

Randomized phase 2 study of perampanel for sporadic amyotrophic lateral sclerosis

Type‐I‐error rate inflation in mixed models for repeated measures caused by ambiguous or incomplete model specifications

Telmisartan for treatment of Covid-19 patients: An open multicenter randomized clinical trial

Contact Info

Product

Resources

About