Marion Campbell
banner
marionkcampbell.bsky.social
Marion Campbell
@marionkcampbell.bsky.social
Trialist; Methodologist; Statistician; Professor of Health Services Research; University of Aberdeen. Views my own. Hebridean.
https://www.abdn.ac.uk/people/m.k.campbell
The work provides accessible definitions of the core elements of an estimand which will help aid understanding and should ease the use of the estimand framework across the range of stakeholders 8/8
November 3, 2025 at 7:44 AM
A new paper by @suziecro.bsky.social and colleagues will help in this regard. Working with public partners, they developed a practical tool to explain the attributes of an estimand in an accessible manner 7/8
trialsjournal.biomedcentral.com/articles/10....
November 3, 2025 at 7:44 AM
However, the technical language around estimands makes it hard to discuss the concept with patients and public (and indeed with many researchers and clinicians). 6/8
November 3, 2025 at 7:44 AM
Kahan et al provide a useful introduction to the estimand approach and to the different terms/approaches used: 5/8
www.bmj.com/content/384/...
November 3, 2025 at 7:44 AM
We may want to know how effective the treatment was in all patients allocated to it (this is called a “treatment policy” estimand strategy); or only in those patients who received it as planned (called a “principal stratum” approach) or some other variation. 4/8
November 3, 2025 at 7:44 AM
“Intercurrent events” are post-randomisation events (eg discontinuation of treatment or treatment switching) and the estimand describes precisely how these events will be handled in the analysis. 3/8
November 3, 2025 at 7:44 AM
An #estimand is effectively a more precise specification of the research question(s) and how the treatment effect is to be estimated (hence estimand). It also specifically describes how so called “intercurrent events” will be handled in the analysis. 2/8
November 3, 2025 at 7:44 AM
Reposted by Marion Campbell
The world’s largest and longest clinical trial in critical care, SuDDICU found a significant reduction in hospital-acquired infection in critically ill patients on mechanical ventilation, when given antibiotics preventatively before a new infection can develop.
October 31, 2025 at 3:31 PM
Reposted by Marion Campbell
Recruitment has begun for a new National Statistician (£200k). Important role as the strategic head of the UK statistical system. There is less operational work as there is now a Perm Sec for ONS. So they are seeking an ‘internationally recognised leader in statistics’
www.saxbam.com/appointment/...
UK Statistics Authority
Official statistics play a vital role in supporting society and the economy - informing government policymaking and evaluation, guiding economic and
www.saxbam.com
October 25, 2025 at 6:56 AM
Bigirumurame et al also published a useful review of SMART trials and highlighted the need for accurate reporting 7/7
www.sciencedirect.com/science/arti...
October 20, 2025 at 6:39 AM
Kidwell & Almirall wrote a short primer paper on the design 6/7
jamanetwork.com/journals/jam...
October 20, 2025 at 6:39 AM
They have also be used in scenarios where one might begin with testing less costly or less intensive treatment strategies in the first phase and stepping up to testing more intensive strategies only if there is lack of response 5/7
October 20, 2025 at 6:39 AM
The SMART design is typically more used in chronic conditions where response can be measured over time and different interventions/doses/sequences are available should individuals not respond to their first line allocated intervention 4/7
October 20, 2025 at 6:39 AM
These designs are often seen as attractive to patients as the intervention they receive can be changed if they see no improvement 3/7
October 20, 2025 at 6:39 AM
In a SMART trial each individual can undergo multiple randomisations. All get initially randomised in stage 1, but non-responders can then get randomised in stage 2 to further treatment options etc 2/7
October 20, 2025 at 6:39 AM