Skip to main content
Search form
  • 00:00

    [MUSIC PLAYING][An Introduction to Studies Within a Trial (SWAT)]

  • 00:14

    SHAUN TREWEEK: I'm Shaun Treweek.I'm from the Health Services Research unitat the University of Aberdeen.Trial methods are what really light my fire.And what I mean by that is, if we are tryingto run a randomized trial, what I'm really interested inis how we do that more efficiently,so how we improve processes like recruitment and retention, data

  • 00:38

    SHAUN TREWEEK [continued]: collection, and dissemination of results.How we improve those processes isthe main focus of my research.So what I'm going to talk about in this tutorialis something called a study within a trial, a SWAT,a way of studying what happens in trial processes.[How did you become interested in this type of research?]

  • 01:03

    SHAUN TREWEEK [continued]: I became interested because as anybody who works in trialswill know, we do a lot of stuff because that'sthe way we do it.There's no great evidence that we can lean onthat the way we are doing things is the best, most effective wayof doing things.There is inefficiency.

  • 01:23

    SHAUN TREWEEK [continued]: That is plain for all to see.So the reason I got interested in SWATs,Studies Within A Trial, is because itis a great way of providing evidenceto help us make better decisions about how we do processeslike recruitment, like retention, like datacollection, like dissemination within trials.

  • 01:46

    SHAUN TREWEEK [continued]: [What is a Study Within A Trial (SWAT)?]A study within a trial is conceptually really simple.As the name suggests, it's a study of one or more trialprocess alternatives.As concrete examples, you might imagine that to help people

  • 02:09

    SHAUN TREWEEK [continued]: think about taking part in a trial,you might present the information in one of two ways.What a study within a trial would do is ask the question,which of these two ways is the best way?So you would compare one against the other,very often randomized.So half of the trial participants

  • 02:30

    SHAUN TREWEEK [continued]: would receive one way of seeing the information.The other half would receive the other wayof receiving the information or having the informationpresented to them.And you'd compare at the end of your SWATwhether one or other method of presentingthat information to participants improved recruitment.

  • 02:51

    SHAUN TREWEEK [continued]: The great thing about a SWAT is it'sresolving, or helping to resolve, uncertaintywith regard to the choice of trial process alternatives.And we're using exactly the same approachthat we would use in our clinical trialsto resolve clinical uncertainty.It is a study within a study.

  • 03:12

    SHAUN TREWEEK [continued]: That's what it's all about.Nothing rocket science or anything like that.We're just looking to resolve uncertainty in trial processdecision-making, in the same way that wewould want to resolve uncertaintyin clinical decision-making.[What type of trials would this be suitable for?]

  • 03:35

    SHAUN TREWEEK [continued]: As a great fan of SWATs, what I'm going to say is, anything.You have any trial process, you could use a SWAT.So if we think about a trial pathway, all the processes thatare there in a trial, right at the start,you would have choosing the research question, making surethat question is important.

  • 03:55

    SHAUN TREWEEK [continued]: At the other end, you'd have disseminating the results.How do we make sure these results arrivein the hands of those people we want to use our results,and indeed the participants who took part?And we have everything in between--approvals, retention, recruitment, data collection.All of that is in between.All of those trial processes you could think of a SWAT,

  • 04:18

    SHAUN TREWEEK [continued]: so a look at how you might improve it.Comparing different data collection techniques,comparing different ways of deciding what to measure,comparing different trial recruitment methods,comparing different ways of keeping people engagedwith the trial so that your retention isexcellent at the end, comparing different waysof providing the results of your trial

  • 04:40

    SHAUN TREWEEK [continued]: to the people who helped you with that trial.All of these things are things you can do SWATs on.So to come straight back to your questionin a summary, any trial process, Ithink you could sit down and think,yeah, we could do a SWAT here.We could evaluate what it is we are thinking of doing here,because SWATs are just the same as we do with randomized trials

  • 05:04

    SHAUN TREWEEK [continued]: for resolving uncertainty clinically.It's the same site type of approach.[What data can be collected and analyzed using this method?]To date, most of what we've been doing with regards to SWATshas collected one form or other of quantitative numerical data.

  • 05:24

    SHAUN TREWEEK [continued]: We'd be counting things.How many people were recruited usingmethod 1 compared to method 2?How many people returned a questionnaireusing method 1 compared to question 2?So it's mostly been numerical.But there are some SWATS which are juststarting to come onstream where we're asking more qualitative

  • 05:46

    SHAUN TREWEEK [continued]: questions.We're asking for people's opinions.Or we're asking people, why would this make it hardfor you to take part?What would make it easier for you to take part so?That data is rich, textual, opinion-based data,which we would then have to collate and synthesizeto draw conclusions about whether we wouldtake one approach or another.

  • 06:07

    SHAUN TREWEEK [continued]: So essentially, if it's quantitative counting,we're all pretty familiar with that.Does one approach lead to an improvement in recruitment?Does it increase completeness of data counting stuff?But this qualitative work-- much newer.And I think we're still working through howbest to use those sorts of data within SWATs, because one

  • 06:29

    SHAUN TREWEEK [continued]: of the key features of a SWAT is,you would want the information you collectto have at least some possibility of beinggeneralized to other situations.And for the numerical work, that's--not straightforward.It's more straightforward.For the qualitative work I think that's trickier,because often, it's very, very focusedon that particular trial and the issues

  • 06:50

    SHAUN TREWEEK [continued]: of those particular types of participants.But the aim of the SWAT would be that it would be generalizable,at least in principle.And I think we're working through thatfor qualitative SWATs.I'll give you one particular example

  • 07:11

    SHAUN TREWEEK [continued]: of what we think of as a bit of a poster child for SWATs.And we call it SWAT 24.There is a repository of SWATs, and this is number 24.And it's a theory-based cover letterthat you would send with a questionnaire.So many, many trials collect databy sending out questionnaires to participants.

  • 07:32

    SHAUN TREWEEK [continued]: So you've been in the trial for a year.We want to know how you are.Send out a questionnaire.A lot of those questionnaires never come back.So what we wanted to do with this particular SWATintervention was, OK, we're going to send a cover letter.We're going to say, dear Shaun, thank youfor taking part in a trial.Please complete this questionnaire.But what we want me to do as a participant

  • 07:53

    SHAUN TREWEEK [continued]: is change my behavior.My behavior is returning the questionnaire.We want it to be more likely that Iwill return that questionnaire.So this cover letter was designed as a behavioral changeintervention.And the way it was written is tapping into behavioral theory.What will make it more likely that I will actually

  • 08:13

    SHAUN TREWEEK [continued]: return that questionnaire?It's actually a very small interventionit's a one-page document.You're probably going to say, dear Shaun, thank youfor taking part at the trial anyway.But you're going to send a one-page document anyway.So you replace one document with another documentthat you are going to send out anyway.That intervention-- I think the great thing

  • 08:34

    SHAUN TREWEEK [continued]: that we've managed to do with that one-- it's cheap as chips.It's a letter.We have a template for how it's put together.We learn to also think about the value of having a template.We also learn a lot about the value of having textthat we could give to different groups whoare trying to act in their trials, textthat they could send to their ethics committees

  • 08:54

    SHAUN TREWEEK [continued]: so they don't have to invent it themselves.We helped groups.So groups got active support from usin implementing this SWAT intervention.They got support from us in workingwith their ethical committees and their trial teamsto explain why this was worth doing, why it was important.And the great joy of the entire exercise--

  • 09:15

    SHAUN TREWEEK [continued]: we've run it in at least four trials,and the overall effect is lookinglike a small benefit for something that is really cheap,is really easy to do.And in the trials world, anything that increasesstuff you need to be increased-- fabulous.And if it's cheap?Even better.What we've also learned from thatis, I think we need to be a little a little bit

  • 09:37

    SHAUN TREWEEK [continued]: smarter about the way we decide whatgoes into letters like that.So one thing we didn't do, which we perhapsshould have done to a much greater extent,is involve members of the public and patients, peoplelike those who we're trying to encourageto return their questionnaire.It should have involved those more, people like them,

  • 09:57

    SHAUN TREWEEK [continued]: in the design of content.So we had a little tiny bit of knowledge.So the think a message to ourselvesand others would be, we might get a better effecthad we involved people who were like those people whowere taking part in the trials to gettheir views on what the content of that letter would be.What would motivate you to return this question?

  • 10:19

    SHAUN TREWEEK [continued]: What would be important to you?What can we do to make it easier for youto return that questionnaire?So that would be a key learning point.But overall, really pleased with that particular SWAT.Cheap as chips, small benefit.What's not to like?[Are there any tools or resources you would recommend?]

  • 10:39

    SHAUN TREWEEK [continued]: There's actually quite a lot thatis becoming available to help people with SWATs.I'm going to talk about to.So I lead an initiative called Trial Forge.And SWATs are really right now at the heart of what we do,because they're a good way of generating evidence.Also, recognize that SWATs are quite new to many, many people.

  • 10:59

    SHAUN TREWEEK [continued]: So what we've done within Trial Force is developsome guidance that enables people who've neverheard about SWATs to at least get a sense for what they areand the sorts of challenges they might imagine exist, and givesome practical tips to try and reducethe size of those challenges.So in the journal trials, we've published something

  • 11:21

    SHAUN TREWEEK [continued]: called Trial Force Guidance One, which talks about,what is a SWAT?What are the challenges?What are the the things you might need to know?What are the things that are actually different about SWATsthan what you might expect?So we give advice about costs.We give advice about sample sizes.We give advice about ethics, for example.We also have a second guidance document,

  • 11:41

    SHAUN TREWEEK [continued]: which helps people make decisions aboutwhether a new evaluation of a SWAT interventionis actually needed, so that we wantto develop some criteria so that peoplemake sensible choices about which interventionsneed evaluation.Other things that I can say is, thereis a Trial Forge SWAT Centre based

  • 12:03

    SHAUN TREWEEK [continued]: at the University of York.And they have an enormous amount of material on SWATs,as well as a great willingness to help with SWATs, you can contact them,and they'll give you advice and help.[What advice would you have for students looking to do SWATs?]Do one.

  • 12:25

    SHAUN TREWEEK [continued]: Students actually have a great source of potential SWATs.SWATs are often very small.They're often very simple.They're a great way of getting a short publicationearly in your career.So I would recommend that any student whois interested in health research thinkabout doing a SWAT if they're involved with a trial.

  • 12:45

    SHAUN TREWEEK [continued]: You can run the entire study yourself.You have your own study.You have your own publication.It's a great way to get started in research.[MUSIC PLAYING]


Shaun Treweek, PhD, Chair of Health Services Research Unit, University of Aberdeen, discusses studies within a trial (SWAT), including suitability, data collection and analysis, process for conducting, tools and resources, and advice for students.

Looks like you do not have access to this content.

An Introduction to Studies Within a Trial (SWAT)

Shaun Treweek, PhD, Chair of Health Services Research Unit, University of Aberdeen, discusses studies within a trial (SWAT), including suitability, data collection and analysis, process for conducting, tools and resources, and advice for students.

Copy and paste the following HTML into your website