The Effective Statistician - in association with PSI

Estimands for Repeated Continuous Outcomes

Dec 18, 2023

Oliver Keene, a statistician, discusses the evolving landscape of estimands. They cover topics such as estimates for repeated continuous outcomes, treatment policy, intent-to-treat analysis, communication challenges, and the validity of last observation analysis.

Ask episode

Chapters

Transcript

Episode notes

Introduction

00:00 • 2min

Estimating Continuous Endpoints Collected Over Time

01:44 • 4min

Collaboration on Estimates Project

05:43 • 4min

Challenges of Estimating Treatment Policy with Missing Data

09:25 • 12min

Defining Treatment Effects and Validity of Last Observation Analysis

21:14 • 6min

Estimation Methods and Effective Communication

27:28 • 5min

Interview with Oliver Keene

Discussion on Oliver's Career
Estimates for Repeated Continuous Outcomes
Response to Ting's Article
Defining Estimates
Discussion on Treatment Policy
Treatment Policy and Intent-to-Treat Analysis
Hypothetical Strategy
Communication Challenges:
One Patient, One Vote Concept
Last Observation Analysis
Communication of Statistical Approaches

We invite statisticians to reflect on the evolving landscape of estimands, encouraging thoughtful consideration of estimation techniques and a deeper exploration of causal inference in clinical trials. The journey through this nuanced statistical terrain unfolds, offering valuable insights for both seasoned professionals and newcomers to the field.

Reference: Wright et al 2023 Response to Ting

Role play reference:

Keene ON, Ruberg S, Schacht A, Akacha M, Lawrance R, Berglind A, Wright D. What matters most? Different stakeholder perspectives on estimands for an invented case study in COPD. Pharmaceutical Statistics. 2020 Jul;19(4):370-87. https://pubmed.ncbi.nlm.nih.gov/31919979/

Other references:

Naitee Ting (2023) Emerging insights and commentaries – MMRM vs LOCF, Journal of Biopharmaceutical Statistics, 33:2, 253-255, DOI: 10.1080/10543406.2023.2184828
https://www.tandfonline.com/doi/abs/10.1080/10543406.2023.2184828
Wright D, Bratton DJ, Drury T, Keene ON, Rehal S, White IR. Response to Comment on” Emerging insights and commentaries–MMRM vs LOCF by Naitee Ting”. Journal of Biopharmaceutical Statistics. 2023 Sep 23:1-3.

Keene ON. Adherence, per-protocol effects, and the estimands framework. Pharmaceutical Statistics. 2023;1‐4. doi:10.1002/pst.232
Keene ON. Intent-to-treat analysis in the presence of off-treatment or missing data. Pharmaceutical Statistics 2011, 10:191–195, doi: 10.1002/pst.421.
Keene ON, Wright D, Phillips A, Wright M. Why ITT analysis is not always the answer for estimating treatment effects in clinical trials. Contemporary Clinical Trials. 2021 Sep 1;108:106494.
Keene ON, Lynggaard H, Englert S, Lanius V, Wright D. Why estimands are needed to define treatment effects in clinical trials. BMC Medicine. 2023 Jul 27;21(1):276.