RESULTS

Identifying reviews conducted using these eight methods was challenging. There were many cases where the used method of qualitative synthesis was not explicitly articulated within the review title and/or abstract. Terminology was often used inconsistently and interchangeably. The term “meta-synthesis” was frequently used as a synonym for qualitative synthesis regardless of which method was used. “Meta-synthesis” could also refer to the whole review process including study identification, quality assessment and data analysis or specifically refer to the data synthesis element. Nonetheless, across the eight methods for synthesizing qualitative studies, we identified 107 reviews (Table 4, Supplementary Table 1).

Table 4.

Summary of Methods for Synthesizing Qualitative Research and Related Reviews (For Details Including Full References, See Supplementary Table 1)

* Reviews were identified which reported using meta-ethnography, meta-study, meta-summary, thematic synthesis, grounded theory synthesis, meta-interpretation, critical interpretative synthesis, or cross-case analysis in title or abstract.

** Some reviews are reported in more than one paper by the same authors.

Use of the different methods across these 107 reviews varied. Table 4 demonstrates that reviews using critical interpretive synthesis, meta-interpretation, qualitative cross-case analysis and grounded theory synthesis were found infrequently. As the potential of these methods has yet to be fully established in the context of HTA, these are not discussed further here and readers are referred elsewhere for details (Reference Ring, Ritchie, Mandava and Jepson18). Meta-ethnography (Reference Noblit and Hare16), thematic synthesis (Reference Thomas and Harden22), and meta-study (Reference Thorne23,Reference Thorne24) were the most used methods (Table 4), especially on topics relating to health services/technologies and, whereas the number of reviews using meta-summary (Reference Sandelowski and Barroso20) were small, this method is of relevance to HTA. As such, we consider meta-ethnography, thematic synthesis, meta-study, and meta-summary to have a role in HTA, and these are outlined below with illustrative examples. Readers are referred to Supplementary Table 1 for references of individual reviews and the related theoretical guidance are recommended reading (Reference Ring, Ritchie, Mandava and Jepson18).

Meta-ethnography—derived from the work of Noblit and Hare (Reference Noblit and Hare16)—is currently the most common method for synthesizing qualitative health(-related) research—reflected in our finding of 63 reviews using this method. Meta-ethnography involves seven steps which, by bringing together findings from individual interpretive accounts, produces a new interpretation (Reference Noblit and Hare16). To synthesize these findings, meta-ethnography uses a variety of approaches including “reciprocal translation analysis” and “line-of-argument synthesis” (Reference Noblit and Hare16).

Most reviews using meta-ethnography were conducted in the 2000s by various healthcare disciplines. Twenty-nine reviews explored perceptions and experiences of disease and other clinical issues including eating disorders, diabetes, pregnancy, and childbirth. Twenty-one reviews explored views on health services/technologies including a range of medications/therapies. A further 13 reviews focused on health/social topics such as poverty and domestic violence (Supplementary Table 1).

One example (Reference Malpass, Shaw and Sharp12) of interest to HTA reviewers aimed to enhance understanding of patients’ experience of anti-depressants. Sixteen papers met the inclusion criteria and were quality assessed before synthesis. Two groups of papers were identified—those which focused on the patient-doctor decision-making process and those with a focus on patients’ self-concept and notions of stigma (Reference Malpass, Shaw and Sharp12). The reviewers found that the initial diagnosis of depression may be associated with feelings of relief or despair and that patients’ preferences for engaging in decision making during their illness is a dynamic process. To improve adherence to advice regarding medication use during depression, it was suggested that GPs should take into account patients’ changing situations and sense-of-self during their treatment (Reference Malpass, Shaw and Sharp12).

Thematic synthesis is often (but not exclusively) used for analyzing qualitative data alongside quantitative data synthesis. Initially developed by researchers from the EPPI-Center, it addresses questions around “what works,” primarily in relation to health promotion interventions. Reviewers can synthesize qualitative and quantitative research separately then integrate their findings (Reference Thomas, Harden and Oakley21). Thomas and Harden's (Reference Thomas and Harden22) method develops analytical themes through descriptive synthesis and finding of explanations relevant to a particular review question (Reference Thomas and Harden22). This method was developed to address specific review questions about need, appropriateness, and acceptability of interventions, as well as effectiveness. Peoples’ views and experiences are taken into account, and hypotheses that could be tested against the findings of qualitative studies are generated. We identified 15 reviews—10 of these investigated topics relating to health services/technologies including health promotion, decision making in chronic kidney disease, and the patient–doctor relationship. Five reviews focused on understanding disease/clinical conditions such as lay understanding of cancer risks and mothers experiences of bottle-feeding (Supplementary Table 1).

An example of thematic synthesis of relevance to HTA is a study on adolescent experiences following organ transplantation (Reference Tong, Morton, Howard and Craig25). This study aimed to improve understanding of the psychosocial impact of transplantation from a wide body of research from different time periods, locations, and contexts. Eighteen papers met the inclusion criteria and data were extracted and synthesized from individual results sections to identify descriptive themes. Five themes were identified: redefining identity; family functioning; social adjustment; managing medical demands; and, attitude to the donor (Reference Tong, Morton, Howard and Craig25). The reviewers concluded that to ensure adolescent patients are able to deal with the multifaceted nature of their psychological response to transplantation, healthcare teams need to provide them with relevant skills and support (Reference Tong, Morton, Howard and Craig25).

Meta-study involves critical interpretation of existing qualitative research (Reference Bondas and Hall6). Before synthesis can take place and a new interpretation obtained, three analytical phases are completed—“meta-theory, meta-method, and meta-data analysis” (Reference Thorne24). These phases, equating to the analysis of theory, analysis of methods, and the analysis of findings (Reference Barnett-Page and Thomas5), can be conducted concurrently to “provide a unique angle of vision from which to deconstruct and interpret” a body of qualitative literature (Reference Bondas and Hall6). Once these analytical processes have been completed, meta-synthesis “brings back together ideas that have been taken apart,” creating a new interpretation of the phenomenon under investigation (Reference Bondas and Hall6). Our search identified thirteen reviews mostly published in nursing journals. Seven of these studies were potentially relevant to HTA reviewers for example, those focusing on shared decision making and spirituality in end-of-life care (Supplementary Table 1).

Edwards et al. (Reference Edwards, Davies and Edwards8) used meta-study to understand why shared decision making in healthcare was not being implemented and to develop a model of external influences on health information exchange. These reviewers identified three themes: internet information use; cultural differences in intercultural consultations; and the influence of the women's health movement (Reference Edwards, Davies and Edwards8). A range of external influences relating to the practitioner and/or patient were included in the model. These influences included receptiveness to patient empowerment, expression of cultural identity and role expectations while health information exchange was assisted where patients already had skills in sourcing and understanding basic health information (Reference Edwards, Davies and Edwards8).

Meta-summary uses “quantitatively oriented aggregation of qualitative findings that are themselves topical or thematic summaries or surveys of data” (Reference Sandelowski and Barroso20). Such summaries can be conducted on their own, or in association with more traditional qualitative synthesis, and can include qualitative and quantitative descriptive findings. Although the least used method presented here—we only identified five reviews on topics such as compliance with hypertension therapy, motherhood and HIV infection (Supplementary Table 1)—it is of relevance to HTA as meta-summary reflects a quantitative perspective and can be used with survey data which is often excluded from some qualitative synthesize due to lack of conceptual depth and richness.

Meta-summary has been used to explore the experiences of parents receiving a positive prenatal test result indicating pregnancy could result in a child with a disability (Reference Sandelowski and Barroso19). Seventeen qualitative studies of American women and/or their partners who received a positive prenatal test result were synthesized. Once data were extracted, findings were reduced to 39 meta-findings, which were ordered by their relative presence or prominence. The most prominent finding was that positive diagnosis is depicted as the experience of termination with an emphasis on parental choice and decision making (Reference Sandelowski and Barroso19). The reviewers suggest that healthcare services should support couples receiving a positive prenatal diagnosis to develop a “narrative of choosing” that can allow them to come to terms with their decisions post-testing (Reference Sandelowski and Barroso19).