Understanding delayed access to antenatal care: a qualitative study Rosalind Haddrill
Formulating the review question, defining late booking
Yüklə 0,62 Mb.
|
- Bu səhifədəki naviqasiya:
- 2.2.2 Searching the literature and sampling
- 5 key concepts for literature search, and search terms used 1. Pregnancy
- 5. (antenatal/late) Booking
- Criteria for inclusion in synthesis Published in English
- 2.2.3 Determination of quality
- Figure 2.2.2 Literature search summary
- 151 abstracts read, 122 non-relevant studies removed
- 29 studies remaining
- 66 studies checked for quality; 12 discarded as methodologically weak, having limited relevance or limited presentation of findings
|
2.2.1 Formulating the review question, defining late booking Not tightly focused or defined in advance, the aim of CIS is to allow a definition of the phenomenon being studied to emerge from the analysis of the literature through an iterative process, modifying/refining the question in response to search findings; the question “more as a compass than as an anchor” (Eaking and Mykhalovskiy 2003: 190). Traditional systematic reviews emphasise precise review questions, which lead to narrower, less iterative and inclusive results, more suitable for aggregative reviews rather than interpretive ones (Dixon-Wood et al 2005, Greenhalgh et al 2005a). The iterative, flexible approach reflects Noblitt and Hare’s (1988) meta-ethnographic approach, which involves “identifying an intellectual interest that qualitative research might inform” (p26) rather than a specific hypothesis or fixed meanings for concepts. Concepts may emerge late in the review and be a product of the review itself, particularly when the focus is hard to specify, such as access (Dixon-Wood et al 2005, Flemming 2010). The question for the review evolved and was refined throughout the initial phases of literature searching and data extraction. The question became “what perceptions and beliefs do pregnant and postnatal women express about delayed access to/initiation of antenatal care (‘late booking’)?” There was a particular focus on barriers to early initiation, a common theme amongst all the literature. Defining late booking/delayed access is challenging as there is no consensus of opinion, with definitions varying across the literature, from after the first trimester (Roberts et al 1998, Sunil et al 2010), 20 weeks gestation (Johnson et al 2003) to the third trimester (Harvey and Faber 1993, Nothnagle 2000, Heaman et al 2014). The Kotelchuck Adequacy of Prenatal Care Utilization (APNCU) Index, used in several US studies, defines inadequate care as booking after the 4th month of pregnancy (or receiving less than 50% of the American College of Obstetrics and Gynecology recommended number of antenatal visits) (Cook et al 1999, Johnson et al 2007). As all the studies define late initiation as being at or after the current NICE guidance definition of the limit of early booking as 12 weeks and 6 days gestation (NCCWCH 2010), and as the focus of the synthesis is perceptions and beliefs, I have chosen to accept the different definitions offered by the authors. 2.2.2 Searching the literature and sampling The purpose of the literature search is to identify potentially relevant papers or ‘candidate studies’ to provide a sampling frame, from which evidence can be chosen to maximise relevance and theoretical contribution. Again reflecting Grounded Theory, other authors have identified that for a Critical Interpretive Synthesis literature searching methods may be iterative and purposive, using theoretical and purposive sampling based on key concepts identified early on, and aiming for conceptual robustness and to reach a theoretical saturation of literature, rather than identify all relevant studies (Dixon-Woods et al 2006a, Barnett-Page and Thomas 2009, Noyes and Lewin 2011). They suggest that the emerging theoretical framework, the development of inclusion and exclusion criteria and thematic categories, will inform the ongoing literature selection, accessing work that is not directly or obviously relevant (Dixon-Woods et al 2006a, Kazimierczak et al 2013). They propose a process which starts with highly structured searches of a range of electronic databases (for including nursing, medicine, sociology, psychology) using broad search terms to avoid missing relevant materials, stressing the importance of ‘adjacent literature’ and its important contribution to any theoretical arguments. This then would develop into a more organic process to fit the emergent/exploratory nature of the questions/concepts, with techniques such as website searching, reference chaining, hand searching of journals and expert contacts. As sampling is concurrent with theory generation, the focus in an interpretive synthesis is on the development of concepts and theory rather than an exhaustive summary of data. Other authors have used a low threshold to maximise the inclusion and contribution of a wide variety of papers at the level of concepts, to give a maximum variation sample. This has included qualitative, quantitative, mixed method, editorial, review, and theoretical evidence, prioritising papers that appear to be relevant rather than by method/study type, limited by practical considerations. The notion of theoretical saturation for sampling could be perceived as problematic but reflects the ‘authorial voice’ and individualistic nature of a CIS, and reflects an approach which remains grounded in the evidence (Dixon-Woods et al 2006a,b, Kazimierczak et al 2013). As the starting point for the synthesis the databases MEDLINE (via EBSCO), CINAHL Complete (via EBSCO), PsycInfo and Maternity and Infant Care (via OVIDSP) were searched using five key concepts, individually and in combination, using truncation and Boolean operators. The databases were chosen to represent medicine/health, nursing/midwifery and psychology. The concepts and subsequently the search were refined iteratively using key terms identified from papers in the initial search, which were incorporated into the list below as the search progressed:
As mentioned previously, several authors have identified the lack of UK-based research in this area (Downe et al 2009, Callaghan et al 2011). I decided therefore to search for studies from other countries with similar economic and cultural situations and levels of healthcare provision. The criteria for inclusion of evidence were as follows:
The year 1980 was identified as a limit as, as Downe et al (2009) comment, this coincides with the publication of the Black Report and subsequent interest in health inequalities and access to healthcare. The broad inclusion criteria for the synthesis necessitated a combination of different search methods. Other authors have identified the challenges of finding qualitative studies, due to inadequate indexing and titles which make it difficult to establish relevance (Dixon Woods et al 2007, Ring, Jepson and Ritchie 2011). In their systematic review of complex evidence, Greenhalgh and Peacock (2005b) found that only 30% of their primary sources were found by the traditional method of using a predefined search strategy, and recommended using other techniques. Twenty three (43%) of the studies included in this synthesis were found using the search terms and databases identified above. Database searching was supplemented with more informal snowballing methods of reference tracking, citation tracking and website searching (Greenhalgh et al 2005a, Greenhalgh and Peacock 2005b). Websites searched included NHS Evidence and NHS England, National Institute for Health and Care Excellence (NICE), The Cochrane Library and the National Perinatal Epidemiology Unit (NPEU), in addition to the use of general internet search engines such as Google and Google Scholar. As the literature search progressed I developed a theoretical sampling approach to try to achieve ‘saturation’ on the subject of late booking, looking for evidence in relation to the range of women identified as most likely to book late: black and minority ethnic (BME) women, teenagers, substance misusing women, homeless women, women experiencing abuse, low income women in both urban and rural situations (Dartnall et al 2005, Lavender et al 2007, Raatikainen et al 2007). Studies that included women who received no antenatal care at all were included, however studies whose focus was a specific ethnic group in a specific location were excluded. The aim was to maximise the range of studies rather than including multiples of similar studies. In order to cover the complete breadth of relevant research and research methodologies, qualitative, quantitative and mixed method studies were included, published in peer reviewed journals as primary research, as well as those published as policy documents. Quantitative and mixed method studies were included if they contained significantly detailed surveys of women’s perceptions and beliefs in relation to antenatal care, as opposed to one or two questions and socio-demographic data collection only. Reflecting the broad inclusion criteria for a CIS I also included systematic literature reviews/analyses/syntheses which met the inclusion criteria above, some of which had been undertaken as part of national clinical guidelines development and epidemiological reports. These reflect a detailed and considered view of the evidence relating to delayed access to antenatal care. Evidence that was UK based was prioritised for maximum relevance, even when methodologically weak. It was also decided to include studies identified in the search which contained both women’s and practitioners’ views, for further breadth. The search process is summarised in figure 2.2.2 and the database search histories can be found in Appendix 2.1. 2.2.3 Determination of quality Dixon-Woods et al (2006a) argue that there is little consensus on how/whether to appraise papers for inclusion in interpretive reviews, and that there is no hierarchy of study design for qualitative studies, as there is for randomised controlled trials. Others agree: Barnett-Page and Thomas (2009) argue that meta-ethnography is less committed to the concept of quality appraisal, whereas Flemming (2010) suggests that the quality appraisal of qualitative research remains contentious. Dixon-Woods et al suggest that for an interpretive synthesis, evidence should be prioritised by relevance (likely contribution to theory development), the content and utility of findings, rather than methodology, with a low threshold for inclusion to maximise the variety of papers at concept level. They also suggest that the threshold of inclusion should be informed by expertise and instinct rather than being rigidly articulated a priori. Sandelowski, Figure 2.2.2 Literature search summary 7444 potential titles identified from four database searches (Medline: 3292, CINAHL: 872, Psychinfo: 1389, Maternity and Infant Care: 1891) Titles read, duplicates and non-relevant studies removed 151 studies remaining 151 abstracts read, 122 non-relevant studies removed: Containing/predominantly demographic data only 33 Focus on antenatal care in general 29 Limited to a specific ethnic group in a specific community 16 Focus on outcomes/intervention/service evaluation 13 Focus on pregnancy intention/behaviour change only 11 Review/discussion/opinion/commentary only 9 Satisfaction studies/focus on number of visits 3 Antenatal screening/testing/information giving only 3 Not in high income countries 3 Provider or partner views only 2 29 studies remaining 37 additional titles identified from website searching, citation tracking and reference chaining 66 studies remaining 66 studies checked for quality; 12 discarded as methodologically weak, having limited relevance or limited presentation of findings 54 studies remaining 54 studies included in synthesis: 14 qualitative 19 quantitative 14 mixed method 7 systematic review, analysis or synthesis of literature Docherty and Emden (1997) have argued previously that such expertise is essential to distinguish between ‘surface mistakes’ and more serious errors. Kazimierczak et al (2013) similarly argue that, though appraised for both, priority should be given to theoretical relevance, as opposed to methodological rigour. Dixon-Woods et al point out that the process of synthesis involves ongoing critical judgements/interpretations of credibility and contribution anyway, suggesting that limited formal appraisal of methodological quality is appropriate. They identify that even some weak papers methodologically could have high relevance in terms of theoretical contribution, demonstrating comprehensive summary or breadth of evidence. As Talseth and Gilje (2011), in their CIS of nurses’ responses to suicide, comment: “all studies in the sample could contribute to an interpretive synthesis, even though they varied somewhat in their reported reliability/validity, trustworthiness and/or Jadad score” (p 1653) In this particular study the authors argued that quality scales such as Jadad have been criticised as being simplistic. They decided to weight all the reviewed studies equally, critically reflected on and then synthesised them into a conceptual understanding. In another the author used a quality appraisal checklist to score methodological quality but low scoring papers with greater relevance were included (Flemming 2010). Dixon-Woods et al (2005) suggest a need to consider and reflect on the credibility of the evidence (for example older research), both in terms of the quality (method) and their contribution to theory, but to exclude only those which are fatally flawed in methodology. They suggest this as a defining characteristic of CIS: a critical approach to the literature in terms of deconstructing research traditions or theoretical assumptions. In this CIS I have tried to merge these different perspectives, utilising a transparent and pragmatic approach to balance relevance with rigour. I have used two methods of quality appraisal. The first involves the use of the UK Critical Appraisal Skills Programme Critical Appraisal Checklists for (1) qualitative research and (2) systematic reviews (CASP 2013), and (3) Boynton and Greenhalgh’s (2004) Quality Checklist for Questionnaire Surveys. These were used for (1) the qualitative studies and mixed method studies with a qualitative focus; (2) the literature reviews/analyses and (3) the quantitative studies and mixed method studies with a quantitative focus. These are presented in appendix 2.2. Secondly, I have used the criteria proposed by Dixon-Woods et al (2005) and Annandale et al (2007) for assessing all empirical papers, regardless of study type. These have been adapted from the National Electronic Library for Health criteria for the evaluation of qualitative research and are as follows:
Yüklə 0,62 Mb. Dostları ilə paylaş:
|