project

Question 1

Can you give some examples of how CHIM studies can reduce time of clinical trials?

Question 2

Who are the current research leaders in this field? Have your explored the possibility of making a visit to their institution to see what you can learn?

Question 3

What’s the most important paper?

Question 4

Introduce a paper that you recently read

Question 5

What do you mean biased vaccine efficacy?

Question 6

Why there might be predictability issue in CHIM? How your project can address this problem?

Question 7

Higher doses may not necessarily increase the infection rate/attack rate, how will you address the methods are applicable in a variety of toxicity profiles?

Question 8

I understand that CHIMs are in a very controlled setting. But what if they suffer from invokable harm?

Question 9

You state that “results from CHIMs are not always robust as it’s not always possible to mimic natural infection”. How will you develop methods to ensure the results are robust given this fact? explain this and how this could limit your research

Question 10

What’s your favourite disease?

Question 11

Why UK is the leading country in CHIM research?

Question 12

Has anything in your application changed since submission?

Question 13

Why have you considered doing simulation studies and not more analytic techniques to develop your methods?

Question 14

You plan a series of simulation studies – what are the potential limitations of solely focusing on simulations?

Question 15

What are the potential challenges/limitations of delivering simulation studies?

Question 16

You use a 67.5% attack rate for CHANTS. How was this figure decided on?

Question 17

How does the model-based approach used in the CHANTS study improve the process of determining an optimal challenge dose compared to traditional rule-based designs in CHIM trials?

Question 18

Why you chose just the model-based method. What about other methods such as the model-assisted BOIN, and escalation with overdose control (EWOC)?

Question 19

CRM is the earliest method and there are a lot of other methods out there recently developed. Why are you still choosing it? Do you think the design would be very outdated?

Question 20

One may think that it is unethical to dose participants even if it is within 60-80. What is your view on this.

Question 21

What are the main ethical considerations in dose-finding studies for CHIM trials, especially when determining the optimal challenge dose?

Question 22

What if you get specific disease areas with very varying dose levels, how would you reconcile that?

Question 23

How will you overcome if one simulation study/work package fails?

Question 24

How might chat GPT or AI assist this proposed research?

Question 25

What’s your criteria of selecting the best model in your simulation

Question 26

Why this kind of dose-finding study can not be powered with predefined power in a traditional way? How are you going to consider the sample size?

Question 27

Please explain what is the ADEMP framework for your objective 1

Question 28

Please explain what is the ADEMP framework for your objective 2

Question 29

For project 1 why have you chosen 3 diseases, and why have you picked the 3 diseases you’ve chosen?

Question 30

What other diseases you will consider in the future?

Question 31

Will disease choices be generalisable to other diseases?

Question 32

How the selection of diseases is generalisable to future pandemic?

Question 33

You say model-based designs use fewer participants, and this is one of your motivations for using it. Are you sure they always use less people? In the CHANTS study, you look at 5 scenarios, 4 of them have the same median sample size and IQR, and scenario 3 has a bigger median for the CRM-based design

Question 34

The outcome in your dose-finding studies is binary. Have you considered using a continuous biomarker instead?

Question 35

Can the dose – attack rate, dose – vaccine efficacy, dose – sample size simulations be done independently, or do you need to do each one first?

Question 36

What if study level data points come from different population/study design, how this influence your simulation or deal with the potential bias if you put them together?

Question 37

How do you know exposure of people in epidemiological studies?

Question 38

Would the result of your current systematic review be helpful for your DF?

Question 39

What model you will use to simulate 3 curves in P1? Can you explain how they different from each other?

Question 40

Develop the dose -sample curve – alpha 0.05 and power 0.8 – why did you select 80% power rather than 90%?

Question 41

Will you test the validity of optimum challenge dose of 60% -80% as I was unclear. Are you planning any validation studies?

Question 42

You need to try different assumptions and models, do you think the timeframe is enough for you to run your simulation? What’s the plan if you could not find the answer within this time?

Question 43

Project 2 – you’re going to borrow information from external studies. Where do you plan to get those external data? Why you think it is important?

Question 44

How will advanced Bayesian models bring benefits to your results?

Question 45

Risk - What happens if Project 2 over runs (you do identify potential risks) as this will limit your time and compromise undertaking project 3 and hence developing the guidance could be at risk.

Question 46

How will you assess the accuracy of dose recommendations? /what do you mean by better dose recommendations?

Question 47

Dynamically borrowing information from previous studies to update your data, or using external information as prior distribution?

Question 48

Borrowing from animal studies and applying to humas – goal is usually quite different in terms of infection rate, and method of exposure. What if attack rate and/or immunological response are very different between animals and humans?

Question 48

Why you chose HCV and paratyphoid as your project 2?

Question 48

If I understand correctly, you are using previous data on dose attack and dose efficacy curves to find the optimal target attack rate for a vaccine efficacy trial. And this could be disease specific. Do you think this will be relevant for a future pandemic with a new unknown illness?

Question 49

You say that dose is continuous and isn’t constrained by tablet size for instance. So are you using fixed doses in your studies or are you able to recommend a dose based purely on the model?

Question 50

You say that dose is continuous and isn’t constrained by tablet size for instance. Are you looking at any measure of variance around that estimate?

Question 51

How will you ensure your developed methods will be robust?

Question 52

How would you make sure the prior distribution is unbiased and scientifically valid?

Question 53

What is the statistical techniques you will use to balance the bias when you use the information from previous studies, introduced by the differences between studies and different suggestions from clinicians?

Question 54

Regulators look at new approaches not willing to change current practice… what are your plans to contribute to dissemination of results to regulators to try to pave the way for future research? If they give pushback, what is your plan to try to push for your simulations in real-life trials?

Question 55

Given the difficulty in getting model-based designs used in the phase 1 oncology setting, what are you going to do differently to get them used quicker in CHIMS?

Question 56

How will you contact regulators (FDA, EMA, MHRA), who will you interview? How will you get in touch with them and how will you know the right people to speak to?

Question 57

Will you only talk to statisticians within your own team?

Question 58

Your stakeholder group is very heterogenous. Have you considered the potential conflict of interest might occur and how are you going to balance it?

Question 59

Why not include a course on NVivo to analyse the interviews?

Question 60

There are so many advanced and interesting method that never had the chance to the real life. How would you identify funders and regulators and contact them and convince them, as we know funders and regulators? What is your outcome and expectation?

Question 61

Have you considered to talk to regulators and funders before you run your analysis rather than after?

Question 62

What do you mean user friendly code ? Do you plan to develop a package or graphical interface?

Answer 62

adaptable and readable

Question 63

Can you tell me the implementation barriers you saw in SR and CHANTS?

Question 64

What is simulation you will do in the project 1 and 2, can you explain that more

Question 65

Can you explain what you mean by semi-structured interviews in groups? Are you planning 1:1 interview or focus groups? You describe interviews in some places but focus groups in others.

Question 66

Why are group interviews best placed to do this – rather than a consensus method such as Delphi?

Question 67

Can you explain what you mean by semi-structured interviews in groups? Are you planning 1:1 interview or focus groups? You describe interviews in some places but focus groups in others.

Question 68

Why are group interviews best placed to do this – rather than a consensus method such as Delphi?

Question 69

Justify why you have chosen a snowball sampling approach to identify participants given you want to propose guidance?

Question 70

What are the disadvantages of snowball sampling given the purpose of this phase of work to develop guidance?

Question 71

What are your sampling characteristics for each participant group being interviewed?

Question 72

How will you develop your interview guide?

Question 73

Study conduct- Explain how you will conduct these interviews? How can you share objective 2 findings and give them time to assimilate – will this be done prior to the interview?
Reason: Participants will need time to reflect

Question 74

How will you analyse your qualitative data?

Question 75

Justify your sample size of N=15

Question 76

What methods will you use to ensure bias is minimised?

Question 77

Describe what the output will be – what will it look like and how will you disseminate it

Question 78

You have allocated 5 months to recruit, interview and develop the guidance M25-30). Explain why this time is feasible (- no time allocated to analysis)

Question 79

What if the attack rates end up being 90% to 100% and the vaccine being investigated is not effective?

Question 80

Is it ethical to dose 80% attack rate to healthy volunteers? ->

Question 81

How do you foresee the analysis framework and practical guidance evolving or being applied to future CHIM trials beyond the scope of this fellowship

Question 82

Why haven’t you consider NIHR as a funder for a focus group?

Question 83

How will you ensure that the data extracted from such varied sources are comparable and reliable for your simulations?

Question 84

Given that your research involves data simulation, how do you plan to address any ethical concerns related to the applicability of your results to real-world clinical trials, especially in terms of participant safety in future CHIMs?

Question 85

What are the practical challenges in translating your findings from simulated data into real-world CHIM trial designs, and how will you ensure that your research has an impact on actual vaccine trials?

Question 86

Once you’ve developed the reasoning framework for researchers to use in dose-finding CHIMs, how will you disseminate this framework and encourage its adoption by other researchers and institutions?

Question 87

Do you recommend all infectious diseases/phase 2 trials to use CHIMs for vaccine evaluation?

Question 88

What is 100 day mission?

Question 89

How 60%-80% can be validated in your project 1?

Question 90

What’s your research impact in 3-5 years, and 5-10 years?