Meta-synthesis method for qualitative research - Wiley Online Library

M E T H O D O L O G I C A L I S S UE S I N N U R S I N G R E S E A R C H

Meta-synthesis method for qualitative research: a literature review Denis Walsh

MA PGDipEd RM RGN DPSM

Senior Lecturer, Department of Midwifery, Faculty of Health, University of Central Lancashire, Preston, UK

Soo Downe

BA MSc PhD RM

Professor, Midwifery Studies Research Unit, Faculty of Health, University of Central Lancashire, Preston, UK

Accepted for publication 25 May 2004

Correspondence: Denis Walsh, Department of Midwifery, Faculty of Health, University of Central Lancashire, Preston PR1 2HE, UK. E-mail: [email protected]

Journal of Advanced Nursing 50(2), 204–211 Meta-synthesis method for qualitative research: a literature review Aim. This paper discusses the purpose and stages of meta-synthesis and the epistemological status of knowledge generated from the technique. Particular attention is paid to exploring the contested areas of the method that remain. Background. There is a growing interest in meta-synthesis as a technique for generating new insights and understanding from qualitative health care research. An increasing number of meta-synthesis papers are appearing in the nursing and midwifery literature. Methods. Literature on the technique of meta-synthesis and examples of metasynthesis papers were searched and reviewed. A meta-synthesis exercise was undertaken, and this informed reflection and critique of the method. Findings. Meta-synthesis attempts to integrate results from a number of different but inter-related qualitative studies. The technique has an interpretive, rather than aggregating, intent, in contrast to meta-analysis of quantitative studies. Examples from the literature indicate that some aspects of the technique are not yet fully established. Conclusion. Despite the contingent nature of evidence gleaned from meta-synthesis and current lack of consensus about some of its aspects, meta-synthesis is an important technique for qualitative researchers and can deepen understanding of the contextual dimensions of health care.

WALSH D. & DOWNE S. (2005)

Keywords: midwifery, literature review, meta-synthesis, qualitative research, interpretive, contextual

Introduction Meta-synthesis is a relatively new technique for examining qualitative research (Jensen & Allen 1996). It has been applied in areas as diverse as transformational leadership (Pielstick 1998), experience of chronic illness (Thorne et al. 2002), diabetes (Paterson et al. 1998, Campbell et al. 2003), concepts of caring (Sherwood 1997a), adaptation to motherhood (Beck 2002, Clemmens 2003) and midwifery care (Kennedy et al. 2003). Stern and Harris (1985) were the first to coin the phrase ‘qualitative meta-synthesis’ with reference to the 204

amalgamation of a group of qualitative studies. Their aim was the development of an explanatory theory or model which could explain the findings of a group of similar qualitative studies. This highlights one of the key differences between this method and meta-analysis of quantitative studies. The latter aims to increase certainty in cause and effect conclusions in a particular area, while the former is more hermeneutic, seeking to understand and explain phenomena. ‘Systematic review’ must also be distinguished from meta-synthesis. The systematic review is a robust way of comparing quantitative research and proceeds according to 2005 Blackwell Publishing Ltd

Methodological issues in nursing research

well-determined steps, which include statistical analysis of the pooled results of studies. This statistical analysis is more accurately called meta-analysis, although this phrase has become interchangeable with systematic review. With the micro ethnographies replacing grand anthropological ethnographies of whole cultures, more and more studies have been undertaken where overlap of milieu occurs, and this has led to attempts to summarize existing knowledge in specific fields. The seminal publication of Noblit and Hare (1988) presented a synthesis of linked ethnographies, describing the method as meta-ethnography, and their work has become something of a template for subsequent endeavours (Jensen & Allen 1994, Britten et al. 2002). This turn to synthesis has not been without its critics. Interpretive research approaches have always endorsed a more circumspect attitude to knowledge generation than positivist approaches, which seek to establish objective truth claims. Philosophically, interpretivists endorse constructionist and postconstructionist thought, which emphasize the contingent, meaning-making nature of knowledge production (Kent 2000). Thus, the synthesis of tentative, somewhat equivocal findings of qualitative research methods into some kind of more comprehensive understanding or even explanatory theory of phenomena is viewed with suspicion. As Sandelowski et al. (1997) say: To summarize qualitative findings is to destroy the integrity of the individual projects on which such summaries are based, to thin out the desired thickness of particulars…and ultimately to lose the vitality, viscerality and vicariism of the human experiences represented in the original studies. (p. 366)

Both Sandelowski et al. (1997) and Campbell et al. (2003) comment on the postmodernist critique of knowledge generation, which is deeply cynical of any one coherent theory as a singular explanation of phenomena. Countering these arguments is the claim that qualitative researchers risk further marginalization from policymakers and clinicians if their work remains isolationist and esoteric, and seemingly incapable of influencing either strategy or practice (Silverman 1997). The argument is that, if they continue to produce non-reconcilable islands of knowledge around the same phenomena, they are doomed to irrelevant speculation and to reinventing the wheel. In the same vein, Statham (1988) coin the phrase ‘analytic interruptus’ to criticize qualitative researchers for not going far enough in interrogating their data and analysing their work. Metasynthesis of qualitative studies does attempt such deeper interrogation and analysis. In response to the postmodernist critique that synthesis is reductionist, it may be helpful to view the process as

Meta-synthesis method for qualitative research

opening up spaces for new insights and understandings to emerge, rather than one in which totalizing concepts are valued over richness and thickness of description. This would move the debate away from assumptions that the essence of phenomena has been revealed in a final, unarguable summary, and towards an appreciation that synthesis is an ever-expanding, boundary-breaking exercise (Sherwood 1997b). This focus then respects the multi-layered contexts which can be peeled back to reveal generative processes of phenomena not glimpsed in standalone studies (Sandelowski et al. 1997). Alongside the problematic quest for essence, some authors (Jensen & Allen 1996) urge us to seek the ‘truth value’ of what is being examined. This is unhelpful language because it suggests a stable, fixed reservoir of knowledge waiting to be tapped. A more useful approach lies in understanding that explanatory efforts and theory generation are transitory and always open to revision, because phenomena are mutable and fluid. The temporary and partial knowledge claims of theory generation within this method are termed ‘mid range theory’ by several authors (Estabrooks & Field 1994, Britten et al. 2002), perhaps mirroring Glaser and Strauss’s (1967) datadriven ‘substantive theory’. Further research on the basis of substantive theory may or may not generate ‘formal theory’ in Glaser and Strauss’s terms. Given the increasing appreciation of the forever-contingent nature of knowledge in a postmodern age, this may or may not be a goal of future meta-synthesis. In summary, bringing together qualitative studies in a related area enables the nuances, taken-for-granted assumptions, and textured milieu of varying accounts to be exposed, described and explained in ways that bring fresh insights. The ‘how’ of achieving this, given the limitations already referred to, will now be explored.

Search methods A conventional literature search was undertaken using terms ‘meta-synthesis’ and ‘meta-analysis’. The databases searched were ASSIA, CINAHL, PsychInfo and Medline. Papers were included if they were published since 1980, in English and had a qualitative research focus. Papers with a quantitative research focus were excluded. Other relevant literature, specifically books and book chapters, were identified through an iterative process based on the journal papers. This ‘berrypicking’ approach (Bates 1989) to literature searching is discussed below. Because meta-synthesis is a relatively new technique, judgements were not made about the quality of records found but everything relevant to the topic was read.

2005 Blackwell Publishing Ltd, Journal of Advanced Nursing, 50(2), 204–211

205

D. Walsh and S. Downe

Findings The rest of this paper explores current understandings of the process of meta-synthesis. These stages are illustrated with examples from the literature and, in the process, areas which are still subject to debate are highlighted. In undertaking this interpretation of the current state of the art, we are aware of a number of seminal publications, which are influencing the developing debate in this area (Murphy et al. 1998, Spencer et al. 2003).

Framing a meta-synthesis exercise As with other research activity, an appropriate research question, purpose or aim frames a meta-synthesis. To date, published meta-synthesis accounts have been framed rather broadly. For example, Sherwood’s (1997a) aim was simply to synthesize work on ‘caring from the perspective of the client’. Pielstick’s (1998) account was of ‘transformational leadership’. Britten et al. (2002) adopted a more specific research question relating to the perceived meanings of medicines and their effect on medicine-taking behaviour and communication with health professionals. Clemmens (2003) interest was in the experience of adolescent motherhood and Kennedy et al.’s (2003) was midwifery care. The scope of a meta-synthesis remains a topic for debate, with advocates both of a narrower, more precise approach, and a broader, more inclusive stance. The issue of framing is crucial since, as in meta-analytic techniques, this ultimately determines the manageability of the process, and delimits the transferability of the findings. There is, therefore, a case for a tighter focus at the outset. Indeed, in some examples the breadth of focus and consequent generation of large quantities of papers has led researchers to restrict those included in the review in a way which may seem somewhat arbitrary (Campbell et al. 2003). Significantly, the search must allow for papers which Noblit and Hare (1988) term ‘refutational’, or those which come to oppositional conclusions from the main body of the work in a particular area.

Locating relevant papers While there has been a significant concentration in the literature on methods of synthesis, there appears to have been far less concentration of methods of identifying studies. In contrast to the approach in the majority of meta-syntheses, the process of location and selection of papers is described exhaustively in many meta-analytic accounts. Once the principal of synthesizing constructed accounts is accepted, there does not seem to be any a priori reason why a similar 206

approach should not be used in meta-synthesis. Indeed, very recently authors have published reviews which do adopt this method of locating studies (Barroso et al. 2003, Nelson 2003). As Barroso et al. (2003) state: For researchers conducting qualitative meta-synthesis projects, the ideal goal is to retrieve all of the relevant studies in a field – not simply a sample of them. (p. 153)

It is therefore good practice to undertake a robust search on the topic area as one would do in the early stages of undertaking a systematic review; however, this may well be inadequate in identifying all the relevant literature. The process is not in fact as straightforward as it has become in meta-analysis. The established databases commonly used for meta-analytic searching tend to record studies published in journals. Due to the epistemological stance of many qualitative researchers, papers in this genre are lengthy, since they need to include quantities of data quotations and contextual description. As well as being reported in journal papers, they may be produced as books (which can sometimes be found in searches using key words) or book chapters (which usually cannot be found this way). Additionally, the depth of synthesis demanded by the review may require access to full reports or unpublished theses, which are also sometimes hard to obtain. As a result of the non-availability of many qualitative accounts, search strategies must augment electronic searches with more traditional methods of reviewing, including back-tracking of references and appeals to known authorities in the area for advice about the existence of more obscure publications which might otherwise be missed. Beyond the practical issue of accessing data for the review, there is a question about the prospective nature of the search strategy. The assumption in classic meta-analysis is that the strategy is fully formed before formal searching begins, and that it is only valid if it leads to a linear process of decisiondescription-search-location. In contrast, Barroso et al. (2003) refer to Marcia Bates ‘berrypicking model’, which was first described in a seminal paper in 1989. Bates critiques the reality of what she terms the ‘classic information retrieval model’, which is fixed and linear, and suggests that in fact users of most information technology and manual searches: begin with just one feature of a broader topic, or just one relevant reference, and move through a variety of sources. Each new piece of information they encounter gives them new ideas and directions to follow and, consequently, a new conception of the query…Furthermore, at each stage, with each different conception of the query, the user may identify useful information and references. In other words, the query is satisfied not by a single final retrieved set, but by a series of selections of individual references and bits of information at each


Methodological issues in nursing research stage of the ever-modifying search. A bit-at-a-time retrieval of this sort is here called berrypicking…One could do berrypicking of information without the search need itself changing (evolving). (Bates 1989)

Acknowledgement of the potential that search directions may be divergent, rather than linear in meta-synthesis, as in the paper of Barroso et al. (2003), is a challenge to the classic prospective approach of meta-analytic searching. However, at least in the case of the development and refinement of a search strategy, divergence (and not linearity) probably does reflect real life processes. In a recent meta-synthesis we undertook, iteration around the scope of the review continued until the final stages of the synthesis. This iterative process can be used to create a more or less precise research question, to determine which databases and other sources to search, and decide the inclusion and exclusion criteria for the studies located. These aspects are yet to be determined in the meta-synthesis literature. The crucial aspect, however, is that of transparency, and reviewers must state how they arrived at their strategy, and how this evolved over time. This maximizes the authenticity of the final account. Such transparency includes a flow chart of the number of studies located using the strategy, and the stages to the final list of papers reviewed, with a rationale for those excluded.

Deciding what to include One specific aspect of inclusion which has generated debate is that of combining studies which used different methodological approaches. This debate is less prominent in classic metaanalytic accounts, since the certainties required only permit the inclusion of ‘good quality’ controlled (and preferably randomized) studies (Chalmers 1996). However, in the qualitative paradigm, which sees truths as multiple, and knowledge as constructed, it is legitimate to include a variety of approaches in a meta-synthesis. The question then arises of how, or if, these multiple approaches can be synthesized. Noblit and Hare (1988) propose the use of meta-synthesis in the context of ethnographic research, and their contention was that it could only be applied to papers embracing the same method. The reasoning is that their particular approach requires one study to be translated into another, literally exchanging metaphors, ideas and concepts from one to the other. In this way the important contextual signifiers of each are respected and therefore meaning can be preserved, although new meanings could also emerge. Ethnographies on similar topics are suitable for such synthesis, hence their term ‘meta-ethnography’. Mixing phenomenology and grounded theory with ethnography would make this translation impossible, in the view of Noblit and Hare, because of


the contrasting ways in which these approaches handle data analysis. Descriptive phenomenology uses small sample sizes to elucidate the detail lived experiences and attempts to tease out an essence in phenomena, uncontaminated by the researcher’s presuppositions. Ethnography is concerned with how shared experience is constructed within an environment, and the assignment of meaning within the culture. Grounded theory focuses more on the inductive processes of building explanations for phenomena. Sandelowski et al. (1997), on the other hand, proposes an approach which accounts for different methodologies through explicit recognition of them prior to and during the analytic stage. Beyond this, some exponents of metasynthesis have gone to the extreme of using more generic headings such as ‘descriptive’ or ‘explanatory’ to classify studies, and then amalgamating them without reference to their origins (Sherwood 1997b). Recent exemplars of metasynthesis include studies using different qualitative methods but, in a similar approach to that advocated by Sandelowski, they group them together first for initial examination before attempting any sort of synthesis between methods (Campbell et al. 2003). In parallel with the issue of mixing methods is that of mixing investigators. One of the tenets of constructivist research is that the interpretation is constructed by a single investigator, or by a team of investigators. The implication is that different investigators will generate different interpretations of the phenomena. Sandelowski et al. (1997, p. 367) acknowledges this issue, and proposes that there may be three different approaches to meta-synthesis, namely: • integration of findings of one investigator’s multiple studies in a related field; • synthesis of studies by different investigators in a related field; • quantitative summary of key elements across qualitative studies. The first approach may be seen as the ‘purest’, since it acknowledges this central tenet of constructed knowledge. The disadvantage is that it does not allow for the exploration of multiple viewpoints. The second is the most commonly adopted approach to date. The third is actually a methodological rather than a heuristic point and, arguably, this technique could be used in any meta-synthesis, although it would be a superficial one. It is likely that the optimal approach to all of these epistemological issues is firstly to acknowledge, and then to choose, an approach which is resonant with pre-existing beliefs, and to be transparent about the reasons for and implications of this choice for the results of the review.


207


Appraisal of studies While methods for the synthesis of qualitative studies have been well described (Noblit & Hare 1988, Paterson et al. 2001), methods for appraising the rigour of individual studies has not been so widely discussed. Indeed, in some cases this whole issue had been by-passed with the justification that the rigour of individual studies is less important than the attempt to be as inclusive as possible. Sherwood (1997a) justifies this on the grounds that it is more important not to miss relevant studies. This is a line which we have also taken with respect to a summary review of quantitative studies of the outcomes of freestanding midwife-led units, with the rationale that, although good quality studies in the specific area we were reviewing did not exist, the review was still worth doing to highlight the need for quality studies in the future. Britten et al. (2002) did not apply qualitative criteria for judging papers because their purpose was to illustrate a worked example. In a later publication, however, the same team (Campbell et al. 2003) argue strongly for evaluating studies first. If one is urging the research community to embrace qualitative meta-synthesis, then the use of robust quality markers to judge qualitative research is crucial for the credibility of this method. Where formal criteria have been used for evaluating studies for inclusion in meta-synthesis reviews, two or three different checklists for have been drawn up (Murphy et al. 1998, Mays & Pope 2000). Although there is a great deal of overlap between these checklists, some are more prescriptive and comprehensive than others [see, for example, Campbell et al.’s (2003) use of Critical Appraisal Skills Programme (CASP) criteria]. In our own recent meta-synthesis I we used a checklist derived from seven different published lists (Murphy et al. 1998, Popay & Rogers 1998, CASP 1999, Mays & Pope 2000, Yardley 2000, Cesario et al. 2002, Spencer et al. 2003) located in a systematic research by one of us (DW). This remains a somewhat contested area in qualitative research. While we acknowledge that, for heuristic reasons, agreement on the quality of research is much harder to reach in the interpretivist than in the positivist field, we strongly believe that this is a crucial area for debate, not only for meta-synthesis, but also for qualitative research in general. Only a few published meta-syntheses have included descriptive tables of the studies under review (Beck 2002, Campbell et al. 2003). However, in our view such tables are essential, since they summarize at a glance the methodological and theoretical basis of individual studies. In addition, the criteria used to appraise studies should be included, but this was done in only one paper (Campbell et al. 2003). While this may be due to the space restrictions of journals, these 208

tables have become a regular feature of meta-analysis papers, either on their associated web sites or in the journals themselves. We believe that this tendency should be fostered in the qualitative literature. The next stages concern the synthesis of included studies. In this, we largely follow the classic method of Noblit and Hare (1988).

Analytic technique Determine how studies are related, or dissonant, through a compare and contrast exercise With the definitive studies identified, a compare and contrast exercise is undertaken. Classically, the original author’s understanding of key metaphors, phrases, ideas, concepts, and relations in each study is identified, and usually tabulated. While this may seem straightforward, our experience is that it is often at this stage that questions are raised about the adequacy of the accounts given in papers. In some cases, these are descriptive rather then interpretive, sometimes because the author is resisting a totalizing temptation. In these cases, the study may be summarized loosely to draw inferred themes and concepts from the narrative account; or it may be left out of the summary altogether. In each case, the rationale and consequences of the decision must be explained. In some cases, the published do not seem to be adequate for the conclusions drawn. If the quality criteria used at the outset do not exclude these accounts, this raises further methodological and heuristic questions. In our recent study, our inability to understand the context and conclusions in some published papers led us to seek the original theses and reports. Whatever level of analysis is pursued, the process requires the preservation of meaning from the original text (as interpreted by the authors, or as raw data) as far as possible. Any temptation to force a fit in the interests of illustrating homogeneity should be resisted, as the links between studies may be reciprocal, complementary or conflicting. This process is not to do with distilling out a core meaning or reducing down related categories so that they can placed under an umbrella of some all-encompassing theory or explanation. Hermeneutic intent needs to be preserved so that the richness and intricacies of meaning are revealed. Relationships between categories may then throw up new understandings and perspectives, making the whole greater than the sum of its parts. Practically, this process begins with the first reading of the studies and is completed with the creation of a grid of key concepts. These findings are then juxtaposed to both identify homogeneity of categories/codes/themes and, crucially, to


Methodological issues in nursing research

note discordance and dissonance. Jensen and Allen (1996) identify two processes here. The first, or hermeneutic, aspect consists of accurately capturing individual or cultural constructions. This is the art of respecting and representing context as intended through the original research. Second is the dialectic aspect, which relates the studies to each other. This can be achieved through juxtaposition of the parameters identified by the compare and contrast exercise. Reciprocal translation The next phase begins the translation of one study’s findings into another, using metaphors and concepts that could be applied to both. Sometimes commonalities will be only too apparent and reciprocal translations are relatively straightforward. At other times they will stand in opposition as a ‘refutational translation’ (Noblit & Hare 1988), or they may overlap without being substitutional. Refutation and overlap may contribute to another, emergent, category or understanding which has not been identified in the original accounts. It is crucial to the method that differences are not glossed over or subsumed. Deviant data may be the raw material of another perspective, making a new space for understanding. In fact the absence of divergent or deviant data may arouse suspicion as to the rigour of the reciprocal translation, as qualitative inquiry rarely results in complete congruence of meaning.

Synthesis of translation The final phase is synthesizing the translations to elucidate more refined meanings, exploratory theories and new concepts. Clusters of metaphors become progressively more refined and a consensus emerges as to core themes or explanatory, mid-level, or substantive theory (Sherwood 1997b, Strauss & Corbin 1998, Campbell et al. 2003). The synthesis needs to reflect the tension between contradictory or alternative explanations if reciprocal translations suggest a lack of congruence. Ultimately, the final synthesis will be the grounds on which the value of meta-synthesis is judged and it therefore needs to convey explicitly how the whole is greater than the sum of the constituent parts.

Discussion Because meta-synthesis is in its infancy, little has been written about how rigor in analytical technique should be applied. Jensen and Allen (1996) suggest that the criteria of credibility, auditability, fittingness and confirmability should be used. They link credibility to faithfulness in handling the data so that it remains true to its source, suggesting that the original


research participants should be able to recognize their experiences in the meta-synthesis. They do not elaborate on whether that should be achieved literally and the complex logistics of even attempting it. Thorne et al. (2002) suggest returning to the original researchers and asking them if the integrity of their original research is intact following metasynthesis. Presumably, this would enable them to pass judgement on whether their work has been misinterpreted or extrapolated beyond the limits of the data. For Thorne and colleagues, this was easy to achieve because the research they were synthesizing was predominantly their own. Paterson et al. (1998) prefer to use the criterion of trustworthiness, established by having other researchers independently check the process at each step. In an earlier paper, Jensen and Allen (1994) said that a synthesis could be internally validated by original respondents’ quotes and externally validated by comparison with the theoretical literature. Estabrooks and Field (1994) suggest that the recurrence of themes between compared studies adds to validity, but this begs the question of how discordant data should be interpreted. Triangulation is another technique said it ensure soundness in analysis. Pielstick (1998) understands this as using multiple studies (meta-synthesis does this by definition), multiple data sources and independent review by two recognized experts. Only the use of experts seems worthwhile, because this complex, essentially inductive and reflective process requires the finely honed skills of experienced qualitative researchers (Fingeld 1999, Thorne et al. 2002). Stern and Harris (1985) are unequivocal about the difference between aggregating studies to further analyse them statistically, and aggregating them with further interpretive intent. The interpretive approach leads one to preserve context but, at the same time, to strive for a holistic perspective of the phenomena under review. The goal is to increase understanding, leading to greater explanatory effect, rather than to aggregate and merge findings in a kind of averaging process. Interpretive intent can stimulate new insights and germinate tentative theory from previous studies that stopped short of theory generation (Jensen & Allen 1994). Thus, new knowledge can emanate from such endeavour. The meta-synthesis turn has been criticized on theoretical, epistemological, heuristic and practical grounds. Despite this, it is gaining momentum. We have identified both some of the generally accepted aspects of the technique, which are mainly the approaches to synthesis, and some of the more contentious aspects, such as searching for relevant studies, the process of individual study appraisal and establishing inclusion criteria. We believe that meta-synthesis of qualitative


209


Author contributions

What is already known about this topic • Meta-synthesis of qualitative research is a parallel technique to meta-analysis of quantitative research but has important differences. • The aim of meta-synthesis is interpretive rather than deductive. • Quantitative meta-analysis aims to increase certainty in cause and effect conclusions, whereas qualitative metasynthesis seeks to understand and explain phenomena.

What this paper adds • An exploration of the preliminary stages in undertaking a meta-synthesis. • An explanation of the method’s specific analytic technique is explained. • Exploration of the remaining areas of contestation in the method. research has the potential to enrich understanding of complex, multi-faceted health experiences and healthcare practice environments. In the methodological debates that surround it, there is also the potential to inform contentious aspects of qualitative research such as quality criteria. Although in its infancy, we believe that there is enough of a reservoir of studies on related topics to engage with the method much more than is found in the current literature.

Conclusion As a consensus develops on how meta-synthesis is done and how it should be evaluated, it is hoped that the few health care-related examples cited in this review will expand exponentially. This is not to say that the technique is easy or straightforward, as it requires advanced skills in understanding and doing qualitative research and a highly developed ability to be reflexive. Challenging assumptions and interrogating both consistent and familiar data and the dissident and strange play a part in expanding the boundaries of understanding and explanation. Openness and transparency are additional hallmarks of the method. The future of qualitative meta-synthesis will also be decided by its acceptance by the proponents of the evidence-based practice agenda, and, ultimately, its utility for practice. Both of these show promise as the traditional antipathy towards qualitative evidence is thawing and clinicians are beginning to appreciate its significance for informing the contextual dimensions of research settings.

210

Study conception and design/Critical revision of the manuscript – DW, SD; Drafting of manuscript – DW; Supervision – SD.

References Barroso J., Gollop C., Sandelowski M., Meynell J., Pearce P. & Collins L. (2003) The challenge of searching for and retrieving qualitative studies. Western Journal of Nursing Research 25(2), 153–178. Bates M. (1989) The design of browsing and berrypicking techniques for on-line search interface. Online Review 13(5), 407–431. Beck C. (2002) Mothering multiples: a meta-synthesis of qualitative research. Maternal and Child Nursing 27(4), 214–221. Britten, N., Campbell, R., Pope, C., Donovan, J., Morgan, M. & Pill, R. (2002) Using meta ethnography to synthesise qualitative research: a worked example. Journal of Health Services Research and Policy 7(4), 209–215. Campbell R., Pound P., Pope C., Britten N., Pill R., Morgan M. & Donovan J. (2003) Evaluating meta-ethnography: a synthesis of qualitative research on lay experiences of diabetes and diabetes care. Social Science and Medicine 56(4), 671–684. Cesario S., Morin K. & Santa-Donato D. (2002) Evaluating the level of evidence in qualitative research. Journal of Obstetrics, Gynaecology and Neonatal Nursing 31(6), 531–537. Chalmers I. (1996) Systematic Reviews and Meta-Analysis. BMJ Publishing, London. Clemmens D. (2003) Adolescent motherhood: a meta-synthesis of qualitative studies. American Journal of Maternal and Child Nursing 28(2), 93–99. Critical Appraisal Skills Programme (CASP) (1999) 10 Questions to Help You Make Sense of Qualitative Research. CASP, Oxford. Estabrooks C. & Field P. (1994) Aggregating qualitative findings: an approach to theory development. Qualitative Health Research 6(4), 553–560. Fingeld D. (1999) Courage as a process of pushing beyond the study. Qualitative Health Research 9, 803–814. Glaser B. & Strauss A. (1967) The Discovery of Grounded Theory: Strategies for Qualitative Research. Aldine Press, Chicago, IL. Jensen L. & Allen M. (1994) A synthesis of qualitative research on wellness-illness. Qualitative Health Research 4(4), 349–369. Jensen L. & Allen M. (1996) Meta-synthesis of qualitative findings. Qualitative Health Research 6(4), 553–560. Kennedy H., Rousseau A. & Low L. (2003) An exploratory metasynthesis of midwifery practice in the United States. Midwifery 19, 203–214. Kent J. (2000) Social Perspectives on Pregnancy and Childbirth for Midwives, Nurses and the Caring Professions. Open University Press, Buckingham. Mays N. & Pope C. (2000) Qualitative research in health care: assessing quality in qualitative research… first in a series of three articles. British Medical Journal 320(7226), 50–52. Murphy E., Dingwall R., Greatbatch P. & Watson P. (1998) Qualitative research methods in health technology assessment: a review of the literature. Heath Technology Assessment 2(16), 1–209. DOH, York.


Methodological issues in nursing research Nelson A. (2003) Transition to motherhood. Journal of Obstetrics, Gynaecology and Neonatal Nursing 32(4), 465–477. Noblit G. & Hare R. (1988) Meta-Ethnography: Synthesising Qualitative Studies. Sage, Newbury Park. Paterson B., Thorne S. & Dewis M. (1998) Adapting to and managing diabetes. Image Journal of Nursing Scholarship 30(1), 57–62. Paterson B., Thorne S., Canam C. & Jillings C. (2001) Meta-Study of Qualitative Health Research. Sage, London. Pielstick C. (1998) The transforming leader: a meta-ethnography analysis. Community College Review Winter, 1–14. Popay J. & Rogers A. (1998) Rationale and standards for the systematic review of qualitative literature in health services research. Qualitative Health Research 8(3), 341–352. Sandelowski M., Docherty S. & Emden C. (1997) Qualitative Metasynthesis: Issues and Techniques. Research in Nursing and Health 20, 365–371. Sherwood G. (1997a) Meta-synthesis: merging qualitative studies to develop nursing knowledge. International Journal for Human Caring 3(1), 37–42. Sherwood G. (1997b) Meta-synthesis of qualitative analyses of caring: defining a therapeutic model of nursing. Advanced Practice Nursing Quarterly 3(1), 32–42.

Meta-synthesis method for qualitative research Silverman D. (1997) Towards an aesthetics of research. In Qualitative Research: Theory, Method & Practice (Silverman D., ed.), Sage, London, pp. 239–253. Spencer L., Ritchie J., Lewis J. & Dillon L. (2003) Quality in Qualitative Evaluation, A Framework for Assessing Research Evidence. Government Chief Social Researcher’s Office, London, Occasional Papers Series 2. Available at http://www.policyhub. gov.uk. Statham A. (1988) Women’s approach to work: the creation of knowledge. In The Worth of Women’s Work: A Qualitative Synthesis (Statham A., ed.), State University of New York Press, Albany. Stern P. & Harris C. (1985) Women’s health and the self-care paradox: a model to guide self-care readiness – clash between the client and nurse. Health Care for Women International 6, 151–163. Strauss A. & Corbin J. (1998) Basics of Qualitative Research Techniques and Procedures for Developing Grounded Theory. Sage, Thousand Oaks, CA. Thorne S., Paterson B., Acorn S., Canam C., Joachim G. & Jillings C. (2002) Chronic illness experience: insights from a meta-study. Qualitative Health Research 12(4), 437–452. Yardley L. (2000) Dilemmas in qualitative health research. Psychology and Health 15, 215–228.


211