 Okay, so the second talk I'm giving today is an introduction to meta-ethnography. And I thought I would just roll back for this talk to talk about the nature of qualitative research before we think about the ways in which we might synthesize it. So qualitative research is that which explores people's subjective understandings of their everyday lives. And it can include to collect data, observations, interviews, group interviews, analysis of texts, or even analysis of behavior using videotape. And it involves the application of logical planned and thorough methods of data collection, careful, thoughtful, and above all, rigorous analysis and that's taken from Pope and Mays in 2006. So the sorts of questions that qualitative research might ask are things like what do people think about having this condition or what's their experience of receiving a particular intervention, what do people think works about an intervention, how does it work, why do these aspects matter or don't matter, how do we best implement something. So they tend to be the why, how, what do you think, what's your experience, what's your belief, what's your practice in different activities. And synthesis of qualitative research came started to gain prominence a fair bit later than synthesis of quantitative research. So it's worth rethinking about why we might want to do it. And I think it's worth saying that there, at least in the beginning, was a certain amount of strategic thinking driving why people wanted to synthesize qualitative research. There was a danger within the evidence-based medicine field that with systematic reviews, primarily of randomized control trials, qualitative research, which had always been a little bit marginalized, was likely to become more and more marginalized through the evidence-based practice movement. And I don't know if any of you are familiar with the hierarchy of evidence pyramid, which the evidence-based medicine movement uses to show where the most robust evidence sits. But it's interesting to note that on many versions of that, there is no place for qualitative research. So you have systematic reviews at the top, randomized control trials, and then down at the bottom is expert opinion and qualitative research doesn't fit anywhere. So there was a drive for people within qualitative research in healthcare to think about, well, what could we do to make qualitative research amenable to systematic review and synthesis methods? And then there were sort of less strategic and more meaningful drivers as well that thought that it's less wasteful. We often find that there is very poor referencing between qualitative research studies, which are extensively about very similar topics. So there is quite a lot of repetition, and despite the impression that qualitative research can be very local or very locally relevant only, that we find similar findings across different studies. And actually, we might be being wasteful in terms of not making use of the previous research on similar topics. And then the ones that I think are very powerful, that one way of thinking about qualitative synthesis is that it creates better explanations. It creates higher order and conceptualization. It may produce broader, more all-encompassing theories, and that this will lead to better truths. Truths that are more socially relevant, more complete, or better. And there was a question earlier to one of your previous speakers about transferability. And I think this is quite important with synthesis that, for me, the level at which qualitative evidence becomes more transferable to different settings is when it becomes more abstracted. So theories developed out of qualitative research are more likely to be transferable across contexts and understandings than individual studies or very descriptive studies. Oops, gone to sleep. Sorry. Oh, there we go. Okay. So what do findings from qualitative research and reviews look like? They can be very different things. So it might be that you end up defining a new concept or a new theory. You might end up with a rich description of a particular phenomenology. You might create a new typology, and I'll be presenting an example of that later in the talk. It might be a description of process. It might be more explanations or theories, and it might be developments of strategy. So there's a whole variety of different outputs of qualitative research and of their reviews and synthesis. And sometimes I think people are a bit unclear about what they're hoping to develop, and we could do a better job about that. So just to return to this diagram from the epicenter, metre ethnography you can see is on the far side of the configuring and interpretative approach, and it's generally thought that it's appropriate to generate theory. And part of the reason I wanted to go back to the beginning and say, well, what is qualitative research is that metre ethnography is particular in the way that it thinks about which level of data it works at. So after Schultz it looks at different levels of interpretation about how do we make sense of the world. And there are first order constructs, which are our everyday ways of making sense of our world. And then there are what he called second order constructs, which are social science researches interpretations of this common sense world into academic concepts and theories. And so in a research paper, and I've got some examples of this at the end if anybody wants me to go through it in more detail, the first order constructs would usually be the participant quotes when you're interviewing somebody, the way in which they describe how they understand what's happening to them or their experiences or beliefs are first order constructs and in a paper, those would be presented as the quotes and there may be descriptive author findings around them as well. And then the second order constructs are the academic author interpretation. So when you're looking at a research paper, these are the explanations, the headers, the concepts, the diagrams, any of that stuff. And then our job as systematic reviewers is to produce third order constructs, which are the reviewer interpretations of this level of data. There are some people who say, well, there's still just second order constructs because they're academic interpretations. But I think just in terms of thinking about how you present and lay out data, it's useful to keep the one, two, three. So you find lots of different sorts of qualitative data and qualitative interpretation. And this is adapted from Sandalowski and Borosa's paper of 10 years ago, where they talk about this continuum of interpretation. So exploratory thematic surveys, where it might just be very descriptive. And the second order constructs are very close to the first order constructs. And then moving up to more descriptive and more explanatory interpretations, where the level of abstraction or critical interpretation gets further away. And she says, well, if you're not at least exploring thematic surveys, you're not really doing qualitative research. And then qualitative research may operate in this continuum of close to the data versus further away from the data. And that has implications for the sorts of qualitative evidence synthesis that you can do. So returning to what meta ethnography is and does, there was a key text in 1988 written by Noblett and Hare about synthesizing qualitative studies. It's a really good read. It's really thin. But it's important to note that this is a method of synthesis. It didn't come out of the evidence-based practice systematic review word at all. So there's no guidance in that book about searching for studies or inclusion criteria or quality appraisal tools. And Noblett himself has been quoted as saying that he's amazed that this method is being used mostly in fields of professional practice as an evidence-based practice. It wasn't their interpretation. It wasn't their intention at all. They developed it because they'd done or his team had done a set of ethnographies about different schools and they'd been trying to develop some universal findings from those individual pieces of work. So it's come from a very different world, if you like. And it is very explicitly interpretative. So it sits on that far end of the configuring and interpreting things, defining a synthesis as a product of an activity where some set of parts is combined and integrated into a whole and it involves conceptual innovations, so new interpretations and new thinking. It was picked up in the evidence-based medicine world in systematic review and synthesis, you know, good many years later in 2002. And this worked example using meterethnography to synthesize qualitative research in healthcare, written by my colleague Nicky, Britain and colleagues, was the first real attempt to try and use it in this context. And they demonstrated that it was possible to apply it to different sorts of activities. But again, this was a worked example. So the other stages of systematic review have really just been developed through case law, through people doing it in the world. There's no guidance there elsewhere. The book itself describes seven stages of synthesis for meterethnography. And there are a couple of differences to traditional systematic reviews, even within qualitative systematic reviews. So one of them is that Noblett and Herr suggest that exhaustive searches and trying to generalize from all the studies of a particular setting is a bad thing, not a good thing, and that you may end up with trite conclusions if you try and synthesize too many papers. So this is quite different to what most people think, although it's also unhelpful because the number of papers which has been described as too many varies from six to about 40. So it doesn't really help us think about what might be too many. And the other thing is that it talks about quality of papers in the synthesis as being better quality if they contribute more to the synthesis. So you make a judgment through the process of synthesis and those papers which are more conceptually well developed which allow you to make sense of other papers, get a stronger rating. So it doesn't have, again, the traditional go through each paper and assess for bias, although many people who do it will do that. So the key mechanism of synthesis itself is thought of as translation and this just means that you can read and reread and compare constructs between the different papers and think about whether one case is like another or not. So you're permanently testing ideas against each other to say one is like another except that. And as I've said before, the important thing about the translation is that it happens at the conceptual level. So they're explicit about translation being between second order constructs rather than first order constructs. Yeah, so just a reminder of what the first and second order constructs are and we're operating at a second level. This is an example to show you the sort of thing that I mean. So this is from CVD as cardiovascular disease prevention programs. And this is just looking at how you build up from first order to third order interpretation. So these are quotes from participants in different papers. This is about the benefits of using food sampling which gave people to feel relaxed and ask questions. This is also about having a person who's able to give information and advice. This is how the original researchers have interpreted this kind of information. So the first one was saying practical demonstrations and maybe more effective than written information. This one is talking about the value of program champions. And how we've developed that is to say, well, both of these things and these things are actually about the ability to develop a relationship with somebody who's delivering a program. And so we've said actually these are, although they're giving different examples, these are actually about the same thing. We've turned them into a third order interpretation to reflect that. So this is the first kind of translation. It's reciprocal translation where you're looking for similarities. So you read all of the papers, you translate the themes into the terms, so into the constructs of the others and pay attention to the different metaphors and organisers. Yes, sorry, so I've just said that. It's very similar to a constant comparative method for those of you who are familiar with that kind of approach. And it looks for also about whether some concepts are better and this might be because they are more sophisticated or higher level. So they cover more of the findings in another study. They offer greater explanatory power. And just to remind you that the interpretation is crucial. So you are as the reviewer trying to make sense of these things. And there are lots of different ways of juxtaposing concepts. So you might be coding in Envivo, but equally you might be tabulating findings using malimups and colour coding and text descriptions to try and think about those things. Can you see this or is this too tiny? It's okay, cool. So this is taken from a review that I did about experiences of women of heavy menstrual bleeding without any sort of malign cause. And so this is lined up. These are the four studies. These are the findings within those studies. So this is what the authors have said. This is my label for those and this is my interpretation. So this is a kind of reciprocal translation. And this is a similar one that's taken, this is taken from this Britain paper where the concepts within the primary studies and their second order interpretations are then collapsed into these higher order, third order interpretations. And you can see that tabulating these things make it relatively clear about how you build these things up. So the second kind of translation that Noblett and Hare talk about is refutational translation. And this is where you're looking for oppositional or counter argument findings. So it's a bit like looking for disconfirming cases. You're specifically looking for ideas that refute some of urging patterns in your initial translation. And again, this is an example from that review, the meta ethnography of heavy menstrual bleeding. And this was the refutational findings where only two studies talked about this and they came to opposing conclusions. So one study found that friends, colleagues, families were important to encourage people to recognize that they had a problem and that it wasn't something they should just put up with and they called this entering the sick role. Whereas this other paper about Elston found that other people belittled women's experiences of suffering very heavy bleeding. So it was very opposite. And we didn't have any way of explaining why this might be the case. They were very similar participant groups. It was very similar analysis. There wasn't any obvious way of explaining why these different findings were coming about. So this was a refutational finding. Oh, and I even did it bigger, so you can read it, sorry. Okay, and then the final type of translation is a line of argument. And this is a kind of very holistic way of saying, well, what can we say about the whole? What do we know about all of the findings that we've got here? And this may particularly lead to the development of a new model or theory or new understanding. And many metro ethnographies try to get to this stage of producing a line of argument. This is an example of a diagrammatic representation of a line of argument. And this is, again, from that Britain review, it's about adherence to medicine taking and they developed a new typology of patients. So rather than just having the dichotomous idea that some people are adherent to their prescribed medication regimes and some people are not, they developed this idea that there were actually four groups of people and they defined them as passive acceptors, active acceptors, rejectors and active modifiers. And that these different groups of people had different worries, which they resolved in different ways. So this developed a new typology of medicine taking. And again, this is an overarching synthesis, a line of argument from that review about heavy menstrual bleeding, which proposed that there was a medical disease model and a patient illness model, which were very different and that the medical disease model, which focused on blood loss, failed to capture the experience of women and was very problematic. And that's my last slide. So thank you. Thank you. Thank you. Thank you. Thank you. Thank you.