News

Annie Wright, a statistician and NIHR BRC PhD student, takes us behind the scenes at ICTMC 2024 in Edinburgh through two short videos. It's fantastic to see the diverse activities our team members engage in, and this is a creative, engaging way to showcase that. Stay tuned for a blog post highlighting our key moments from ICTMC!

The 7th International Clinical Trials Methodology Conference will be held in Edinburgh 30th September to 3rd October. Make sure not to miss ICTU’s statistical team who will be presenting at the conference at the following times:

TUESDAY 1st October

>13.25 - 13.45, Suzie Cro, Lightning Talk (Poster session PS.1B), Introducing retrieved dropout reference-based centred multiple imputation for estimation of treatment policy strategies with missing data

>14:45-15:05, Zhangyi He, Poster, How to improve the efficiency of group sequential designs by optimising the timing of interim analyses

>14:45-15:05, Ed Waddingham, Poster, SepTIC: A multi-factorial trial   with an information-rich outcome to efficiently evaluate three interventions in sepsis clinical care

>14:45-15:05, Emanuela Falaschetti, Poster, Importance of control group definition to address the research question. Lesson learned from Endobarrier trial

>15:10 - 16:30, Rachel Phillips, Talk, Public  perspective on potential treatment intervention harm in clinical trials –   terminology and communication

WEDNESDAY 2nd October

>8:30 AM -   9:50 AM, Victoria Cornelius, Talk, Design,   analysis and inference for follow-up trials where sample size is fixed and has inadequate power for hypothesis testing

>13.25 – 13.45, Zohra Zenasni, Poster, Methods to investigate contemporary trial practices in research involving preterm and full-term neonates, a systematic review

>13.25 – 13.45, Jack Elkes, Poster, Analysing the impact of engagement on efficacy for Digital Mental Health Interventions: An optimal strategy to identify user groups for analysis using unsupervised clustering methods

>13.25 – 13.45, Consuelo Nohpal de la Rosa, Poster, Causal inference methods to account for compliance in trials of device interventions

>13:45 - 14:05, Suzie Cro, Poster, Improving the Analysis and Interpretation of Studies Within A Trial using Bayesian Methods and ACCEPT Analyses

Earlier this year Dr Brennan Kahan from MRC CTU @ UCL, along with Dr Suzie Cro, Dr Richard Emsley (KCL) and Dr Ian White (MRC CTU @ UCL) taught a MRC-NIHR TMRP funded workshop covering the statistical methods that can be used to estimate different estimands. The taught materials from the workshop are now freely available to all online following the link below and includes: (i) slides for each lecture; (ii) computer practical’s, which include statistical code in Stata and R to implement different estimators, along with datasets and instructions for their use. These materials can be used to learn different methods that can be used to estimate each intercurrent event strategy including treatment policy, composite, hypothetical, principal stratum and while-on-treatment and the statistical code necessary to implement these estimators.

We have recently advertising our survey to statisticians, clinical investigators and trial operational staff who have been involved in clinical trials for digital interventions. We are looking to answer the question: Are efficient methodological designs being implemented in RCTs of DHIs and what are the barriers and facilitators to their use?

If you think you are elgible please take our survey using the link below. If you have any questions please email Jack Elkes (j.elkes@imperial.ac.uk).

Congratulations to Jack Elkes for winning the judges award for best poster at the Imperial Mental Health Research Network 

summer event. He presented his latest research using unsupervised cluster methods to identify groups of users from engagement data in a self-guided digital mental health intervention.

In June 2024, Laetitia de Abreu Nunes attended the Statisticians in the Pharmaceutical Industry (PSI) conference in Amsterdam, Netherlands. Laetitia presented a poster on her methodological review of statistical methods leveraging the natural structure of adverse events for signal detection in randomised controlled trials, which she had the chance to introduce during the lightning poster session. This review was conducted as part of her PhD at Queen Mary University of London. The PSI conference was a great opportunity to discuss current approaches to safety signal detection and how to move towards better practices with other researchers and industry members. Laetitia also won one of the best poster prizes for her poster.

On the 25th April Imperial College Clinical Trials Unit and the NIHR RSS Imperial & Partners Hub hosted the 1st Contemporary Clinical Trials Methodology Meeting.

The day focussed on the theme of adaptive designs. Dr Leila Janani, Professor Stuart Pocock and Professor Laurent Billot covered Group Sequential Designs and Interim Analysis with some examples in action!

Mr Lukas Pin, Dr Sofia Villar and Professor John Norrie talked about Response Adaptive Randomisation including a demonstration of an Rshiny application.

The day rounded off with Sample Size Re-estimation from Professor Alan Montgomery, Dr Dominique-Laurent Couturier and Ruth Goodall. A massive thank you to all the speakers and the organising comittee Suzie Cro (lead), Victoria Cornelius, Rachel Phillips, Rani Rayat, Matenyeh Kaba, Anneliese Levy, Sara Gazi, Gareth O’Brien, Lauren Noto.

We are all looking forward the next one!

The programme for the first ICTU & NIHR RSS Imperial and partners hubs' contemporary clinical trials meeting is now available by clicking the below link. Inside find the abstracts and biographies for the speakers who will be talking about group sequential trial designs and interim analysis, response adaptive randomisation, or sample size re-estimation. Registration for the meeting is now closed. We look forward to seeing all those who have registered to attend shortly on 25th April.

We are delighted to announce registration is now open to attend Imperial Clinical Trials Unit (ICTU) and the NIHR RSS Imperial and partners Hub's 1st Contemporary Clinical Trials Methodology Meeting, 25th April 2024.

If you are an applied trial statistician or a statistically minded trialist/clinician and want to know more about adaptive clinical trials this trial methodology meeting is for you.

Learn about Group Sequential Trial Designs, Response-Adaptive Randomisation and Sample Size Re-estimation across three main sessions. Including when and how to implement, examples of their use in practise and time for discussion of methodological challenges.

Speakers include:

Group Sequential Designs and interim analysis - Stuart Pocock, Laurent Billiot, Leila Janani (Chair: Victoria Cornelius) Response-Adaptive Randomisation: Sofia Villar, Lukas Pin, John Norrie (Chair: Suzie Cro) Sample size re-estimation: Domique-Laurent Couturier, Alan Montgomery, Ruth Goodall (Chair: Rachel Phillips)

Register to attend at the below link!

Annie Wright, formerly a clinical trials statistician at the unit, has embarked on her NIHR Biomedical Research Centre Digital Health PhD. Under the supervision of Professor Victoria Cornelius and in collaboration with Professor David Sharp of the UK DRI Care and Technology Centre at Imperial College London, she brings her expertise and enthusiasm for clinical trials and routine data to her new role.

Her research will delve into the practical implementation of decentralised clinical trials (DCT), aiming to enhance their trial design for seamless integration of routine and remote data collection. DCTs, which leverage digital technology with a patient-centric approach, depart from the traditional centralised site model by enabling trial activities to be conducted partially or entirely from patients' homes. DCTs have numerous perceived benefits and offer a unique way of thinking about running clinical trials.

Annie's project will analyse the current landscape of DCTs, optimise the integration of longitudinal data, and engage with the public to gauge their perspectives on this approach to clinical trials.

On February 26th an NIHR Public and Patient Involvement and Engagement (PPIE) for Statistical Methodology workshop, was held at the University of Leicester. Dr Suzie Cro presented on her surprises and learnings from co-chairing the HEALTHY STATS public involvement group with her public partner co-hair Ania Henley. Other researchers and public partners presented their experiences of conducting PPIE for Statistical Methodology research and Nifty Fox presented a workshop on engaging story telling. Click the below link to see all the presentation slides from the day.

Dr Suzie Cro explores public partners perspectives on the importance of discussing estimands when designing a trial in a Health Research Board Trial Methodology Research Network (HRB TMRN) webinar, 18th January. She also presents a co-developed practical tool that helps explains what an estimand is and what impact it may have in trial results. The recording is available at the following link.

Registrations are open to attend a free webinar on a guide to pre-doctoral and doctoral NIHR fellowships for aspiring and early career statisticians. Hear from Statci's Ellie Van Vogt who will be talking about their pre-doctoral experiences and Victoria Cornelius who will be giving tips about applying to the doctoral fellowship.  The webinar will be held on 23rd January, 12.45-3 and registration is at the below link.

Registrations are now open to attend a free workshop to be held December 1st 2023 (Bristol),  funded by the NIHR, to learn practical methods for conducting missing data sensitivity analyses in randomised controlled trials using Stata or R.

In this workshop, facilitated by Daisy Gaunt (UoBristol, previous NIHR Doctoral Fellow), Baptise Leurent (University College, previous NIHR Doctoral Fellow), and Suzie Cro (Imperial College London, current NIHR advanced fellowship) available accessible methods to conduct such sensitivity analysis will be de-mystified and practically demonstrated. Register following the link below.

This year, at the Australian Trials Methodology Conference, Ellie Van Vogt was invited to present their research virtually on causal machine learning for the detection of heterogeneous treatment effects.  They presented work that was part of an ongoing project with the Alan Turing Institute and the Medical Research Council Clinical Trials Unit at University College London. Ellie presented in the Trial Analysis session, chaired by Rory Wolfe. Other presenters in the session were Jonathan Bartlett (London School of Hygiene and Tropical Medicine) and Thomas Jaki (Cambridge University). Ellie’s presentation highlighted the benefits of incorporating a machine learning approach into the subgroup analysis of trials, to identify previously not-considered subgroups and investigate high-level interactions between baseline covariates. Recordings of the presentation are available to attendees and a publication of this work will be available soon.

In September, Dr Rachel Phillips presented at the 6th Conference of the Central European Network (CEN), Basel, Switzerland. This was a joint conference of: the Austro-Swiss Region (ROeS), the German Region (DR), and the Polish Region of the International Biometric Society (IBS) and supported by the Basel Biometric society (BBS). Rachel was invited to join the session on best practices for Data Monitoring Committees and how to get there. In this session she presented her collaborative work on recommendations for visualising harm outcomes that aim to improve reporting practice in journal articles and extended these ideas to applications for Data Monitoring Committee (DMC) reports.

In August this year, Ellie Van Vogt attended the International Society for Biostatistics Conference (ISCB) in Milan, Italy. Ellie presented their poster from work with the Alan Turing institute and the MRC clinical trials unit on the use of causal machine learning for the detection of heterogeneous treatment effects in clinical trials. This conference was a wonderful opportunity to network with other researchers across different medical disciplines and discuss the implementation of causal machine learning methods into clinical trials. Follow the link below to view their poster!

Congratulations to Bao Khanh Nguyen who has been awarded a pre-doctoral fellowship by the NIHR. Bao has an MSc in Mathematics and will be joining us at ICTU to gain experience in medical statistics and clinical trials. She will be undertaking research to evaluate novel multiple imputation methods for handling missing data in clinical trials, working with Dr Suzie Cro and Dr Matteo Quartagno (MRC CTU at UCL).  

NIHR Pre-doctoral Fellowships are designed to support people who are looking to start or advance their career within a health and social care research methodology and we welcome you to get in touch with us if this is an opportuntiy you'd be interested in exploring with us. Follow the link below to find out more about the NIHR pre-doctoral fellowship scheme.

The HEALTHY STATS public involvement group provides advice on an NIHR funded project to develop accessible statistical methods to determine treatments effects that matter to patients and prescribing physicians in randomised controlled trials (Suzie Cro’s Advanced fellowship NIHR300593). The group have developed a practical tool for public partners involved in trial design that describes what an estimand is and what impact it may have in trial results interpretation. The aim of the tool is to break down the barriers of scientific jargon and explain what an estimand is to public partners, to facilitate involvement of public partners in estimand discussions. The tool and more information about this work is now published in Trials, which can be accessed via the below link.

At this years PSI conference Dr Suzie Cro gave a presentation of her work on ‘Reference-based Multiple Imputation for longitudinal binary outcomes’. This included presentation of a simulation study exploring two algorithms for implementing reference-based Multiple Imputationin this setting: (i) joint modelling with the multivariate normal distribution and an adaptive rounding algorithm and (ii) joint modelling with a latent multivariate normal model. A recording of the talk will soon be available via PSI on demand for PSI members.

The CONSORT group includes many initiatives to improve reporting of results from randomised controlled trials (RCTs). The latest output from the group demonstrate how items related to harm outcomes could be integrated into the main CONSORT checklist. Thirteen items from the main CONSORT were modified to improve harms reporting and three new items were added. The CONSORT group recommend until future work produces an updated checklist, trial authors, journal reviewers, and editors should use the integrated CONSORT Harms 2022 checklist presented in this paper. This publication is the culmination of a large international research project that began in 2017 and which our own Dr Rachel Phillips became involved in 2019. You can read more about this in this 2019 blog post.  Alongside this latest CONSORT update, Rachel and Professor Victoria Cornelius have published an opinion piece on the future direction of research into harms in RCTs. They call for clear advice on best analysis practice for harm outcomes that is endorsed by key stakeholders and can be easily implemented. Arguing that better reporting encouraged by the new CONSORT harm extension along with guidance for analysis would enable us to better assess the true benefit-risk of interventions. You can read more on their thoughts about the future direction of harms reseach here and how their ongoing DoRAH project hopes to address this.

On June 22nd we will be at the NIHR statistics conference. This is a fantastic event including talks and presentations focused on, Achieving Best Statistical Practice in Health and Care Research. Along with Dr Pavel Mozgunov (MRC Biostatistics unit), StatsCI's Dr Rachel Philips, Dr Suzie Cro and Ms Ellie Van Vogt will be holding a workshop on “Lessons learned from implementing an adaptive design in an early phase dose-finding clinical trial“ . Additionally Mr Jack Elkes has a poster on the results of his systematic review of reporting and analysis of user engagement in trials of digital mental health interventions.  For more details including other sessions and how to register for the event follow the below link.

We had another fantastic research away day with a focus on Adaptive Designs on the 27th April organised by Suzie Cro and Leila Janani. There was a great variety of talks from both internal and external speakers in the morning. Professor Yin’s talk about ‘Control of Type 1 Error Rates in Bayesian Sequential Designs’ sparked a lot of interest in the group. We also enjoyed Professor Billot’s talk on ‘Group Sequential Designs with Unblinded Sample Size Reassessment’. In the afternoon we used these talks as well as thinking about our own problems we face in designing trials in an idea generating exercise to identify issues and research areas we wanted to focus on both personally and as a unit. This exercise was a great success with Zhangyi He conducting research on optimising the timing of interim analyses off the back of discussions; we all look forward to seeing his results!

We are very pleased to announce that Mrs Emanuela Falaschetti is our new Deputy Head of Statistics, and Dr Leila Janani is our new Deputy Head of Trial Methodology.

They will closely work with Dr Rachel Phillips and Dr Suzie Cro to improve the design, analysis and delivering trials at ICTU and to build on trial methodology work.

Congratulations to you both!

We are delighted to welcome back Dr Rachel Phillips to the ICTU statistics team. Dr Rachel Phillips and Dr Suzie Cro now head up the statistics team for ICTU. They bring with them extensive experience in clinical trials, CTUs, and a passion to improve the way we design and deliver trials through research.

Our previous pre-doctoral fellow, Ellie Van Vogt, has successfully been awarded a 3-year doctoral fellowship by the NIHR to use causal machine learning to improve the detection of heterogeneous treatment effects in complex RCT settings.

They will be undertaking research to develop causal machine learning techniques for the estimation of heterogeneous treatment effects in complex settings with competing risks, with a particular focus on critical care settings. Ellie aims to develop sound guidance on the use of causal machine learning methods for determining subgroups of heterogeneous treatment effects, how to communicate results effectively and investigate methods for the incorporation of subgroups into ongoing trials.

On 7th December PSI held their annual medical statistics careers event which we attended. The event included a live presentation from our Ellie Van Vogt on her experiences as a pre-doctoral NIHR research fellow in medical statistics with us. Ellie summarised how her fellowship enabled her to conduct an MSc in health data analytics and machine learning at Imperial, followed by a year conducting a research project exploring causal machine learning methods to detect differences in treatment effects by patient characteristics in trials (known as treatment heterogeneity). The event also included a panel discussion with statisticians from academia and industry, followed by an exhibition by various companies (including us) giving individuals the opportunity to find out more about job opportunities and life as a medical statistician. We invited expressions of interest from individuals who would like to apply for a NIHR Pre-Doctoral or Doctoral Fellowship in medical statistics or clinical trial design with us.  Get in touch with us if you are interested in either of these schemes!

Ellie Van Vogt and Jack Elkes represented ICTU at the School of Public Health showcase of research activities to the Imperial College President - Hugh Brady. Both Jack and Ellie spoke with the president about their research, talking him through each of their posters. The event was a great opportunity to share more of our methods research and a chance to meet the president in person for the first time.

StatsCI had our second research away day of the year on the South Kensington campus of Imperial College at the end of October 2022. We started the day off with some games and looked back at the past few months to celebrate what the team has achieved. Ellie, Suzie and Victoria gave us some insight into their research networks and collaborations that they have made and nurtured through their careers as statisticians.

Our theme for this research away day was ‘Adaptive Trial Designs’. We were lucky enough to have Dr Munya Dimairo from the University of Sheffield with us for the day. Munya is the lead on a collaborative initiative that has set up PANDA; a platform that brings together a wealth of information and resources about adaptive designs and presents them in an accessible manner to anyone involved in the design and implementation of trials using adaptive designs. We were very excited to hear him talk about the platform before exploring it ourselves and then utilising PANDA to answer a mock trial proposal based on real world examples of requests from investigators. This was such a valuable experience, and we are grateful for Munya travelling to us to help deliver such a great away day. We all highly recommend using the PANDA platform to learn more about adaptive designs.

We finished off the day with a scavenger hunt in the Science Museum and some drinks together as a team. Follow the link to our twitter page where you can find a video round up of the day

The EMBIO study, led by Mr. Ahmed Ahmed (Imperial College Hospital NHS Trust), is a multi-centre, double-blinded randomised controlled trial that will assess whether a non-invasive “pinhole surgery” procedure (known as left gastric artery embolization, or LGAE) is effective in helping participants with obesity lose weight over 12 months.

The study, run by Imperial College London, Imperial College Healthcare NHS Trust and University College London Hospital NHS Trust, will compare the weight change of those undergoing LGAE to the weight change in a group of participants undergoing a placebo procedure.

Imperial Clinical Trials Unit, including statisticians Elli Bourmpaki, Nicholas Johnson, and Graham Wheeler, are helping to conduct and analyse the study, which opened to recruitment earlier this year. ITV News reported on the experience of Tom Keech, one of the participants recruited to the trial. For more information on the study and to read about Tom’s experience, click on the link below.

Congratulations to Jack Elkes from our team who was runner up in the Student Poster Presentation award at the International Clinical Trial Methodology Conference earlier this month. The poster was his first research output from his PhD, the poster presents results from his systematic review on current practise to describe user engagement in digital mental health interventions and how engagement is considered in the efficacy analysis. A very well deserved prize for his presentation, if you missed it you can see a recording from the link below! 

October brought along the 2022 International Clinical Trials Methodology Conference in Harrogate, and it was quite a busy one for ICTU’s statistical team with several members of the team presenting posters during the conference:

Eleanor Van Vogt presented her work on exploring Treatment Effect Heterogeneity and novel methods to obtain data driven subgroups.

Victoria Cornelius discussed the power of visualising harm in randomised controlled trials.

Rachel Phillips presented recommendations from a national consensus for visualising harms in randomised controlled trial publications.

Suzie Cro brought to light the importance of estimands for patients and the public through presenting a practical tool that can be used to start a conversation with public partners about estimands. She also presented a poster on a systematic review of estimands

There were two posters from Leila Janani on incorporating clinicals elicited informative prior distributions into a planned Bayesian analysis and the use of adaptive analysis for late phase randomised controlled trials in intensive care with mortality outcomes.

Jack Elkes presented a poster on a systemic review of user engagement in randomised controlled trials for digital mental health interventions.

Elli Bourmpaki and Graham Wheeler presented a poster on Percentage change as a primary outcome in randomised controlled trials of obesity/overweight: a systematic review of design and analysis approaches.

There was also a poster from Nicholas Johnson looking at how inclusive UK-based randomised controlled trials of COVID-19 vaccines were.

Our team (Victoria, Elli and Jack) are speaking at the Imperial North West London Research Symposium on September 22nd. The event is arranged by the Clinical Academic Training Office and took place at South Kensington campus, and was aimed to provide guidance to those interested in pursuing a research career. Our lovely team presented an overview of our Contemporary Clinical Trials Course run by Imperial CTU (ICTU) and what support is available through ICTU Collaborations & NIHR Research and Design Services London. Special thanks to Victoria Cornelius, Jack Elkes, Elli Bourmpaki for presenting the workshops and Ana Sousa Marcelino Boshoff, Claire Smith and Ranjit K Rayat for all their help and support!

If you are heading to the 43rd Annual Conference of the International Society for Clinical Biostatistics (ISCB) look out for poster 43 by Eleanor Van Vogt on Causal Forests for uncovering treatment effect heterogeneity and data driven subgroups in clinical trials. Eleanor will be giving an introduction to her research in the 11-12.30 lightning poster session. Also don’t missing parallel session 26: Missing data in studies chaired by Suzie Cro. This session includes 5 fantastic speakers including Mike Daniels (US), Mia Tackney (UK), Camila Olarte Para (UK), Melissa Middleton (AU) and Audinga-Dea Hazewinkel (UK).

The new Stata command cemimix is now available to conduct reference-based (MNAR) multiple imputation of cost-effectiveness data. Type: ssc install cemimix within Stata to install or downloadable from the below link. It follows the method described in Leurent et al. 2020, based on the work by Carpenter, Roger, and Kenward (Journal of Biopharmaceutical Statistics, 2013). It allows for various missing data assumptions, including 'Missing at random' (MAR), 'Jump to reference' (J2R), 'Copy increments in reference' (CIR), 'Last mean carried forward (LMCF), and 'Baseline mean carried forward' (BMCF). It was primarily developed for the imputation of cost-effectiveness data in randomised controlled trials but can be used more generally for the imputation of up to two continuous outcomes, measured at a single follow-up point or longitudinally (repeated measures).

We are delighted to announce new appointments to two key leadership positions at Imperial Clinical Trials Unit (ICTU). Congratulations to Dr Victoria Cornelius who has been appointed as the Director of ICTU and Dr Suzie Cro who has been appointed as Head of Trial Methodology. To read more about their appointments head to the below link.

On 8th June at a NIHR benefit-risk working group webinar, Dr Ed Waddingham  presented his work from his recent PhD on Bayesian estimation of preference weights for quantitative benefit-risk assessments or other multi-criteria decision problems.

Topics covered included:

· The basics of quantitative benefit-risk assessment, based on the methodology known as multi-criteria decision analysis (MCDA);

· Some commonly used deterministic methods for eliciting preferences;

· Probabilistic models that may be appropriate for preference data;

· Aggregation of preference data from multiple sources;

· The parallel application of these modelling approaches in a benefit-risk assessment.

The talk was illustrated with examples using the Bayesian MCMC approach in BUGS software.

On 10th May we had our Inaugural meeting for a new group led by Rachel Phillips and Victoria Cornelius. The aim of the group is to bring together researchers interested in improving the collection, analysis, selection, and reporting of adverse events in RCTs for pharmacological and complex interventions. The first meeting had an international flavour attracted over 40 delegates from seven countries.

The meeting started with an introduction and orientation to progressing research in harms by Victoria Cornelius and Rachel Phillips. Diana Papaioannou from the University of Sheffield gave a talk on “Working towards efficient and appropriate recording of adverse events in behavioural change trials the RHABIT project”, followed by Kaspar Rufibach from Roche in Basel who gave a talk on “Stop the abuse: A plea for a more principled approach to the analysis of adverse events. The SAVVY project”.

The final session was hosted by Dr Tianjing Li from University of Colorado on determining the Groups first activity- to write a series of educations articles to promote better practice. 

On 28th and 29th April Dr Suzie Cro lead a workshop for trialists introducing estimands, which was supported by the MCR-NIHR TMRP. Speakers included clinicians and statisticians from academia and industry who covered, what an estimand is, why we need them and how to implement the estimand framework. All the sessions were recorded and are freely available to watch via the link below.

This year we will be at the Great Exhibition Road Festival! This is a fantastic event where research teams share insights of their work with the general public. We will be hosting several activities that aim to teach you more about how a clinical trial works and some of the key concepts we use to make sure we can be confident about our answers. 

We will be in the Medicine Zone, you can find out more below.