Dominic Magirr

Statistician @Novartis. Survival analysis, multiplicity, adaptive design.

[Jonathan Bartlett @TheStatsGeek]

PSI online event 1st Dec on covariate adjustment in trials, featuring among others
@tmorris_mrc @statnav https://t.co/bvSFuUQ78A

[Dominic Magirr @DominicMagirr]

https://t.co/opjizeFew5

[Pavel Mozgunov @pmozgunov]

Still time to register for our dose-escalation courses which are built around innovative trials that we have worked on in (oncology, infectious diseases, and opiate detox). We will talk about how we communicate complex designs to clinical teams, funders, and health authorities.

[Kert Viele @KertViele]

Recent threads on non-concurrent controls...

Tweetorial
https://t.co/EXM3rBIJBN

Roig et al
https://t.co/xdzZXdBQq0

Saville et al
https://t.co/JlVnABLjkB

Marschner and Schou
https://t.co/SWaTQKIyvU

Wang et al
https://t.co/Q60pJ52AZk

Current picture
https://t.co/vG22RYnTLb

[Dominic Magirr @DominicMagirr]

Great opportunity here to join the @Novartis Analytics PhD Graduate Programme.

https://t.co/1e0TEQKPxg

[Haiyan Zheng @haiyan_zheng]

📚Check out our NEW PAPER in #SampleSize determination for Bayesian #BasketDesigns that permit borrowing of information across subtrials. @MRC_BSU @CRUKresearch

Joint work with Michael Grayling, @pmozgunov @thomas_jaki @JMSWason

https://t.co/4FJrHXzez3

1/8

[Dominic Magirr @DominicMagirr]

Huge thanks to @IsobelBarrott for her fantastic work on the implementation.

[Dominic Magirr @DominicMagirr]

{nphRCT} on CRAN. A new #rstats package for weighted log-rank tests.
https://t.co/wgoS2B9iIt

[Dominic Magirr @DominicMagirr]

What I think is unique, and I really like, is that we can view a weighted log-rank test as a difference in mean "score" on the two treatment arms.

[Kert Viele @KertViele]

(1/n) Next up on non-concurrent controls in platforms…Marschner and Schou
https://t.co/JQnqu9MDIC

[Kert Viele @KertViele]

(1/n) More on non-concurrent controls in platform trials
Saville et al Clinical Trials
This paper has a different scope than Roig et al, describing the time trend modeling used in ISPY2, GBM-AGILE, Precision Promise, PRINCIPLE, and others.
https://t.co/AhYuxcwEA9

[Dominic Magirr @DominicMagirr]

@KertViele @RWJE_BA @LindsayBerry_ @JasonConnorPhD @martabofillr @GoldLabCharite @KH_Stats @FraKoenig @posch_m @imi_eupearl Yes, we need to keep going. Look forward to it!

[Dominic Magirr @DominicMagirr]

@KertViele @RWJE_BA @LindsayBerry_ @JasonConnorPhD @martabofillr @GoldLabCharite @KH_Stats @FraKoenig @posch_m @imi_eupearl Agree, nicely put. But isn't it a bigger issue with nonconcurrent control because time trends on trt 1 effect can drag up trt 2 effect (going back to the example) when trt 2 does absolutely nothing? Or is this too heavily focussed on the null hypothesis?

[Munya Dimairo @mdimairo]

The School of Health and Related Research (ScHARR) at the University of Sheffield are running a free online course on 8th September for statisticians on the design and analysis of n-of-1 trials.
To register, and for more information, see here: https://t.co/tvfixzFJ6s

[Marta Bofill Roig @martabofillr]

Our paper on non-concurrent controls in platform trials is now online at @BMC_series Medical Research Methodology!

https://t.co/iFBrKeOh7t

@pavlakrotka @GoldLabCharite @KH_Stats @BerryConsultant @FraKoenig @DominicMagirr @Novartis @KertViele @posch_m @imi_eupearl @MedUni_Wien

[Kert Viele @KertViele]

(1/n) More non-concurrent controls in platform trials, Roig et al from the EU-PEARL group (great group to work with!!)
@martabofillr @GoldLabCharite @KH_Stats @FraKoenig @DominicMagirr @KertViele @posch_m @imi_eupearl
(tagging so they can correct me!)
https://t.co/j9UnqgTvQy

[Kert Viele @KertViele]

(1/13) At least three papers came out recently on non-concurrent controls in platform trials This is a rapidly evolving area both for academics and for regulators. I’m going to try to summarize them, as they all have overlapping content. 4 threads…unless another paper appears!

[Colin Begg @ColinBBegg]

Ben Saville and colleagues propose modeling of temporal drift to account for time effects when analyzing platform trials. @Stats_Ninja_ @SCTorg @BerryConsultant @statberry @KertViele @berryni
https://t.co/5NPMJPG7X8

[Frank Bretz @Frank_Bretz]

CEN2023 Conference: Call for session proposals now open!

The Central European Network (CEN) conference from 3-7 September, 2023, in Basel, Switzerland, is organized by the Austro-Swiss, German & Polish IBS Regions

Visit https://t.co/aDZYDujHEG for details

@IBSstats #CEN2023

[Dominic Magirr @DominicMagirr]

@noah_greifer @TheStatsGeek I think it's: what is the population of my estimand? Is it the trial population at time 0, or is it the selected population at time > 0? I would always think the former unless explicitly told otherwise.