The Use of Ethnography in Maternity Care.

While the value of ethnography in health research is recognized, the extent to which it is used is unclear. The aim of this review was to map the use of ethnography in maternity care, and identify the extent to which the key principles of ethnographies were used or reported. We systematically searched the literature over a 10-year period. Following exclusions we analyzed 39 studies. Results showed the level of detail between studies varied greatly, highlighting the inconsistencies, and poor reporting of ethnographies in maternity care. Over half provided no justification as to why ethnography was used. Only one study described the ethnographic approach used in detail, and covered the key features of ethnography. Only three studies made reference to the underpinning theoretical framework of ethnography as seeking to understand and capture social meanings. There is a need to develop reporting guidelines to guide researchers undertaking and reporting on ethnographic research.

Introduction

While there is universal acknowledgment that clinical care should be informed by the best available evidence, this does not always occur. In Australia and internationally, policymakers, health service managers, clinicians, and researchers struggle to implement evidence into routine practice (Braithwaite et al., 2018; Curtis et al., 2017; Holmes et al., 2017). The implementation of new practices is a complex process that is associated with a wide range of individual and organizational factors and processes (e.g., social, behavioral, economic, management) that may hinder or support implementation (Curtis et al., 2017; Greenhalgh, 2018). To understand and address the challenges associated with evidence implementation, over the last 15 years, a growing body of literature has studied the contextual factors and underlying mechanisms that affect implementation (Fogarty International Center, 2013; Greenhalgh, 2018). While this body of research is not associated with any particular research methods, methodological approaches that are most suitable tend to be social science informed approaches such as ethnography (Best et al., 2012; Braithwaite et al., 2018; Hawe et al, 2009; Wall, 2015).

Despite common critiques of ethnography as being unscientific and too limited to allow generalization (Herbert, 2000), in recent years, there has been a growing interest in the use of ethnography in healthcare research (Bunce et al., 2014; Eccles et al., 2012; Greenhalgh, 2018; Greenhalgh & Swinglehurst, 2011), including maternity care. Ethnography is a methodology used in maternity settings that can promote creativity, rigor, and reveal complex phenomena with a capacity to formulate understanding (Newnham et al., 2016, 2021), hence this setting was chosen for this study. In addition, maternity was selected as the setting for this study as the authors of this paper are part of an academic research group whose remit is maternity care, and the second author is a midwifery academic.

Ethnography is an accessible, flexible, and pragmatic qualitative method that can provide a nuanced understanding of the contextual and cultural factors that shape healthcare practice (Braithwaite et al., 2018; Bunce et al., 2014; Greenhalgh, 2018; Holmes et al., 2017; Tomoaia-Cotisel et al., 2013). Ethnography uses naturalistic observation and face-to-face interaction, that is, what is seen, heard and experienced, to understand how a cultural group functions and the rules that guide behaviors (Adler & Adler, 1994; Atkinson & Hammersley, 1994; Denzin, 1970; Hammersley & Atkinson, 2007; Liamputtong, 2011; Liamputtong & Ezzy, 2005; Lofland & Lofland, 1971). The focus of ethnographic research is usually on a few cases, generally fairly small-scale to facilitate in-depth study (Hammersley & Atkinson, 2007). Hallmarks of ethnography include an inductive and cyclical approach to data collection and analysis, cultural immersion (i.e., spending extended time emerged in the lives of the group/s being studied), fieldwork using a range of methods including observation, reflexivity, and consideration of the researcher’s inherent subjectivity (Adler & Adler, 1994; Atkinson & Hammersley, 1994; Denzin, 1970; Geertz, 1998; Hammersley & Atkinson, 2007; Liamputtong, 2011; Liamputtong & Ezzy, 2005; Lofland & Lofland, 1971; Willis, 2007).

More specifically, the starting point of ethnographic research is a set of questions or problems from which the fieldwork begins, rather than a set of preconceived ideas to be proven or tested (Hammersley & Atkinson, 2007; Mackenzie, 1994). Data collection tends to be relatively unstructured, and does not involve following a fixed research design specified at the start (Hammersley & Atkinson, 2007; Liamputtong & Ezzy, 2005). While usually a number of methods are used, immersion in the field through observation is the mainstay of ethnographic research (Atkinson & Hammersley, 1994; Hammersley & Atkinson, 2007; Herbert, 2000). Spending time in the cultural setting, listening, observing, and asking questions to gain insight into the day to day relationships that influence behavior, is fundamental (Francis, 2013; Hammersley & Atkinson, 2007). Observation methods may consist of non-participant observation, where the researcher watches and takes field notes without participating, or participant observation, where there is more active engagement with research participants (Adler & Adler, 1994; Eberle & Maeder, 2011; Morgan et al., 2016).

Ethnographic investigations encompass both intrinsic and extrinsic knowing, and consideration is given to the insider versus outsider continuum, that is, how the researcher balances objectivity versus insider knowledge (Adler & Adler, 1994; Francis, 2013). Key sources of insider knowledge include the researcher if he or she is also a member of the group being observed, or conversations with key informants (Adler & Adler, 1994; Francis, 2013; Hammersley & Atkinson, 2007; Morgan et al., 2016). Key informants are people the ethnographer has recognized as able to answer specific questions to clarify observations (Adler & Adler, 1994; Francis, 2013; Lofland & Lofland, 1971).

Regardless of where the researcher is placed on the insider-outsider continuum, consideration is given to the researcher’s inherent subjectivity through reflexivity (Denzin, 1970; Reeves et al., 2013). Reflexivity is a particular type of reflection, which involves consideration of oneself (the researcher) while planning and conducting ethnographic research (Denzin, 1970; Reeves et al., 2013). It includes self-reflection of how one’s background, values, and history, may impact on the study and interpretation of study findings (Denzin, 1970; Reeves et al., 2013).

Another key characteristic of ethnography is its cyclical approach to data collection and analysis (Francis, 2013; Reeves et al., 2013). This means that data collection and analysis are conducted concurrently; following a period of analysis the researcher returns to the field to validate or clarify previous observation or developing theories (Francis, 2013; Reeves et al., 2013). Data collection generally involves multiple methods, including observation, interviews, or informal conversations, focus groups, surveys, and document analysis (Denzin, 1970; Hammersley & Atkinson, 2007; Willis, 2007). During analysis, data derived from these multiple methods (or sources) are compared and contrasted, called data triangulation (Denzin, 1970; Reeves et al., 2013). Data triangulation is an important way for ethnographers to establish methodological rigor/quality and provide in-depth and holistic understanding (Denzin, 1970; Reeves et al., 2013).

Ethnographic research is also theory informed, allowing for study findings to be transferred to other settings (Bunce et al., 2014; Hammersley & Atkinson, 2007). Specifically, ethnographic research is influenced by symbolic interactionism, a sociological theory that regards how meaning is created and shapes social interaction and behavior (Denzin, 1970; Francis, 2013). As an outcome of research, ethnography generally produces thick descriptions of culture, as well as a contribution to theory (Denzin, 1970; Francis, 2013; Geertz, 1973; Hammersley & Atkinson, 2007).

Ethnography is increasingly positioned as an appropriate and rigorous method to examine phenomena (Bunce et al., 2014; Denzin, 1970; Greenhalgh, 2018; Morgan et al., 2016; Murphy & Dingwall, 2007; Reeves et al., 2013). The key features of ethnography provide the methodological rigor and quality required to present a more accurate perception of what is happening than other approaches allow (Bunce et al., 2014; Denzin, 1970; Morgan et al., 2016; Murphy & Dingwall, 2007; Reeves et al., 2013). Nonetheless, while the value of ethnography in health research is recognized (Greenhalgh, 2018; Greenhalgh & Swinglehurst, 2011), the extent to which it is used, and which particular ethnographic principles are used and reported, is unclear. As such, the aim of this study was to map the use of ethnography in maternity care, and identify the extent to which the key principles of ethnographies were used (or reported).

Method

A systematic review process was used following the PRISMA reporting guidelines for systematic reviews (as per the review protocol, unregistered) (Liberati et al., 2009).

Information Sources and Search

Relevant studies were identified by searching the databases PubMed, Web of Science, Scopus, and Maternity and Infant Care. The databases were searched using the terms “ethnography” in combination with the term “maternity” using the “search in abstract” only function. The term “maternity” was selected as this is a globally used term, and able to identify studies conducted by midwives (e.g., in the UK and Australia) as well as obstetric nurses (e.g., the US).

The databases were searched for the period from January 2009 to December 2018. This search identified 278 articles for screening (37 from PubMed, 65 from Web of Science, 89 from Scopus, and 87 from Maternity and Infant Care). Retrieved citations were uploaded to an Endnote database, and 144 duplicates were removed. The remaining 134 articles were screened for eligibility against the eligibility criteria (outlined below) by the first author by reading the title and abstract. This process excluded a further 83 articles. Reasons for exclusion included not being conducted within a maternity care setting (27), a review study (mostly meta-ethnography) (37), a commentary (12), or ethnography not used as a method (7). The full text of 51 studies were reviewed at which point a further eight were excluded: four were not specific to maternity care, two were poster abstracts, and two were not available in full text. The flow of papers through the review process is outlined in Figure 1.

Figure 1.

Flow of papers through review.

Eligibility Criteria

Studies were included if they:

• Stipulated an aim relevant to maternity care

• Were published in peer reviewed journal in English

• Used the term ethnography in the abstract to describe the method used

• Were empirical (commentaries and review studies were excluded)

• Published between January 2009 and December 2018

Data Extraction and Charting Process

Information relevant to the research question was extracted from each article using a purposely designed electronic data charting form. The first author (a sociologist) designed a data extraction form to extract key information about the aim and use of ethnography in each study. To the best of our knowledge, there are no publicly available tools to extract information from (or critique) ethnographic research. As such, the design of the data extraction form was informed by the methodological literature, both literature relevant to evaluating the quality of ethnographic studies (Mackenzie, 1994; Reeves et al., 2013; Richardson, 2000), and a range of seminal works (Adler & Adler, 1994; Atkinson & Hammersley, 1994; Denzin, 1970; Hammersley & Atkinson, 2007; Liamputtong, 2011; Liamputtong & Ezzy, 2005; Lofland & Lofland, 1971). The specific questions included in the data extraction form are outlined in Box 1.

Box 1.

Data extraction questions.

QUESTIONS INCLUDED IN DATA EXTRACTION FORM Use of ethnography  ○ How do the authors describe the type of ethnography used? ○ What is the rationale for using ethnography as explained by the authors? ○ Do the authors stipulate the type of ethnography used? If so, what type of ethnography was used?  ○ Not stated  ○ This study is “informed by ethnography”  ○ Classic/traditional/realist ethnography  ○ Genre ethnography  ○ Rapid ethnography  ○ Focused ethnography  ○ Feminist ethnography  ○ Critical ethnography  ○ Online ethnography  ○ Duo-ethnography  ○ Auto-ethnography  ○ Video-ethnography  ○ Institutional ethnography  ○ Other: __________________________________ • Do the authors articulate the epistemological position/theoretical framework of their chosen approach? If yes, how is this articulated?Study Aim  • What was the study aim? • Is the aim of the study clearly stipulated and appropriate? • Does the aim fit with an ethnography? • What is the rationale for using ethnography as explained by the authors? As related to the study aim?Study Sites/Cases Under Study  • Where was the study undertaken? • Setting and number of sites • What is the “case” under study? (setting, group of people, etc.) • How many cases are studied?Data collection  • How long is the study period? The data collection period? • Which data collection methods were used? • For each method, what was the aim, who were the participants, and what was the duration (e.g., number of hours of observation conducted)? • If observation was used, was participant or non-participant observation used? • Was the insider perspective included/covered? • How were observations recorded? Were field notes used? Was a structured approach to recording observations used? (e.g., none observational dimensions by Spradley). • Was an inductive approach to data collection and used?Reflexivity  • Was consideration given to the researcher’s position on the insider-outsider continuum? Where is the researcher positioned on the insider outsider continuum? • Is there evidence of the use of reflexivity? Has the relationship between researcher and participants been adequately considered? • Is the relationship between the researchers and participants made explicit? • Has there been a discussion of the researcher’s inherent subjectivity?Analysis  • What was the approach to data analysis and interpretation used? • Has an iterative and cyclical process to data collection and analysis been used? • Did the authors triangulate the data? If yes, how did the authors triangulate the data? • Was their theory development? Does the study use and contribute to theory?

Data extraction was completed by both authors, with extraction completed by the second author (an academic midwife) cross checked by the first author for consistency. Study authors were not contacted for missing information. Data charting was implemented using REDCap, a customizable informatics systems-based web software.

Synthesis of results

Collation and synthesis of the extracted information was conducted using Word 2016 and Excel 2016, and results are reported narratively and tabularly.

Results

A total of 39 studies were identified for inclusion, published in 43 papers (one study was published in three publications, and two studies were published across two publications). Thirteen studies were from the UK (14 papers), five from Australia (7 papers), five from the United States, five from Canada, two from Taiwan, and a study from the Netherlands, Brazil, Hong Kong, Malawi, Afghanistan, West Africa, Burkina Faso, Swaziland, and Tanzania. Only one study described the ethnographic approach used in detail, including all of the key features of ethnography (inductive and cyclical approach to data collection and analysis, field work, data triangulation, reflexivity, consideration of the researcher position on the insider-outsider continuum, and contribution to theory) (Table 1).

Table 1.

Methodology and Method used in Included Studies.

Author (Year), country	Type of ethnography	Methods	Number of study sites/cases	Rationale for using ethnography stated	Epistemological position/theoretical framework stated	Researcher position as insider or outsider	Were field notes used	Reflexivity	Inductive approach to data collection	Iterative and cyclical process to data collection and analysis	Data triangulation	Contribution to theory	Approach to analysis
Allen et al. (2015), Australia	Focused ethnography	Focus group; Observation; Document analysis	1	NS	Yes	Outsider	Yes	NS	Yes	NS	NS	Yes	Roper and Shapira’s framework for ethnographic data analysis
Arnold et al. (2015), Afghanistan	Ethnography	Interviews; Focus group; Observation	1	NS	NS	NS	Yes	NS	Yes	Yes	NS	NS	Thematic analysis
Aubrey et al. (2017), Canada	Focused ethnography	Interviews	1	NS	NS	NS	Yes	Yes	Yes	Yes	NS	NS	Thematic analysis
Black (2011); US	Ethnography	Interviews	NS	NS	NS	NS	Yes	NS	Yes	Yes	NS	NS	NS
Chen et al. (2015), Taiwan	Ethnography	Interviews	1	Yes	NS	NS	Yes	NS	Yes	Yes	NS	NS	Content analysis
Cheyney et al. (2014), US	Ethnography	Interviews	NS	NS	NS	NS	Yes	NS	Yes	Yes	Yes	NS	Triangulation.Grounded theory, constant comparison
Danielsen et al. (2017), Malawi	Ethnography	Interviews; Observation	1	NS	NS	NS	Yes	NS	NS	NS	NS	NS	NS
Dove et al. (2014), Australia	Critical ethnography	Interviews; Observation	≥5	NS	NS	NS	Yes	NS	Yes	Yes	NS	NS	Five stages of data collection and analysis consisting of continuous normative reflection
Finlay and Sandall (2009), UK	Ethnography	Interviews; Observation	2	NS	NS	NS	Yes	NS	Yes	NS	NS	Yes	Grounded theory, constant comparison
Flacking and Dykes (2013), UK	Ethnography	Interviews; Observation	4	NS	NS	Outsider	Yes	NS	Yes	Yes	Yes	Yes	Triangulation.Grounded theory, constant comparison
Gagnon et al. (2013), Canada	Focused ethnography	Interviews; Observation	≥5	Yes	NS	NS	Yes	NS	Yes	NS	Yes	Yes	Triangulation.Thematic analysis
Goodwin et al. (2018), UK	Focused ethnography	Interviews; Observation	1	Yes	NS	NS	Yes	Yes	Yes	Yes	NS	Yes	Thematic analysis
Grassley et al. (2015), US	Institutional ethnography	Interviews’ Focus group; Observation	1	Yes	NS	NS	Yes	NS	Yes	NS	NS	Yes	Content analysis
Harte et al. (2016), Australia	Video-ethnography	Interviews; Observation	1	Yes	NS	NS	Yes	NS	Yes	NS	Yes	Yes	Triangulation.Thematic analysis.
Higginbottom et al. (2013), Canada	Focused ethnography	Interviews	NS	Yes	NS	NS	Yes	NS	Yes	Yes	NA	Yes	Roper and Shapira’s framework for analysis of ethnographic data
Higginbottom et al. (2015), Canada	Focused ethnography	Interviews	NS	Yes	NS	NS	Yes	NS	Yes	NS	NS	Yes	Roper and Shapira’s framework for analysis of ethnographic data
Hugill et al. (2013), UK	Focused ethnography	Interviews; Observation	1	Yes	NS	Insider and Outsider	Yes	Yes	Yes	NS	NA	Yes	Thematic analysis
Hunter and Segrott (2010) and Hunter (2010), UK	Ethnography	Interviews; Focus group; Observation; Key informant; Document analysis	2	Yes	NS	NS	Yes	NS	Yes	Yes	NS	Yes	Thematic analysis
Kemp et al. (2010); UK	Ethnography	Interviews; Observation	NS	Yes	NS	NS	Yes	Yes	Yes	NS	NA	Yes	Thematic analysis
Kennedy et al. (2010), UK	Institutional ethnography	Interviews; Observation	2	Yes	NS	NS	NS	NS	Yes	NS	Yes	NS	Triangulation.Narrative analysis
Kennedy et al. (2013), UK	Institutional ethnography	Interviews; Observation	2	Yes	NS	NS	Yes	NS	Yes	Yes	Yes	Yes	Triangulation.Narrative analysis
Lange et al. (2016); West Africa	Ethnography	Interviews; Observation	≥5	NS	NS	NS	Yes	NS	Yes	NS	Yes	Yes	Triangulation.
Lee et al. (2009), Hong Kong	Ethnography	Interviews; Document analysis	1	NS	NS	NS	NS	NS	Yes	NS	NS	Yes	NS
Lindsay et al. (2012), UK	Ethnography	Interviews; Observation; Document analysis	1	NS	NS	NS	NS	Yes	Yes	Yes	Yes	Yes	Triangulation.Thematic analysis
Logsdon et al. (2017), the Netherlands	Ethnography	Interviews	≥5	Yes	NS	NS	Yes	NS	Yes	Yes	Yes	Yes	Triangulation.Grounded theory
Miltenburg et al. (2018), Tanzania	Ethnography	Observation	4	NS	NS	Insider	Yes	Yes	Yes	NS	NS	NS	Thematic analysis
Mondy et al. (2016), Australia	Video-ethnography	Interviews; Observation	2	NS	NS	Outsider	Yes	Yes	Yes	NS	Yes	Yes	Triangulation.Content analysis
Newnham et al. (2015, 2017a, (2017b), Australia	Critical ethnography	Interviews; Observation; Document analysis	1	Yes	Yes	NS	Yes	Yes	NS	NS	Yes	Yes	Triangulation.Thick description following Geertz
O’Boyle et al. (2013), UK	Feminist ethnography	Interviews; Observation; Document analysis	NS	Yes	Yes	Insider	Yes	Yes	Yes	NS	Yes	NS	Triangulation.Thematic analysis
O’Boyle et al. (2014), UK	Critical ethnography	Interviews; Observation; Document analysis	NS	NS	Yes	Insider	Yes	Yes	NS	NS	NS	Yes	NS
Olson et al. (2013), Canada	Ethnography	Interviews; Observation	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS	Thematic analysis
Østergaard (2015), Burkina Faso	Ethnography	Interviews; Focus group; Observation	3	NS	NS	NS	Yes	NS	Yes	Yes	NS	Yes	Inductive and deductive analysis
Spendlove (2018), UK	Ethnography	Interviews; Observation	1	Yes	Yes	Insider	Yes	Yes	Yes	Yes	Yes	Yes	Triangulation.Constant comparison
Taylor et al. (2009); UK	Ethnography	Interviews; Observation	1	Yes	NS	NS	Yes	NS	Yes	Yes	NS	NS	Grounded theory
Thwala et al. (2011); Swaziland	Ethnography	Interviews	NS	NS	NS	NS	Yes	NS	NS	NS	NA	NS	NS
Torres (2014, 2015) US	Ethnography	Interviews; Observation	NS	NS	NS	NS	Yes	NS	Yes	Yes	NS	Yes	Thematic analysis
Varcoe et al. (2013), US	Ethnography	Interviews; Focus group; Observation	NS	NS	NS	NS	NS	NS	Yes	NS	Yes	Yes	Triangulation.Thematic analysis
Véras and Traverso-Yépez (2011), Brazil	Institutional ethnography	Interviews; Focus group; Observation	1	Yes	NS	NS	Yes	NS	NS	NS	NS	Yes	NS
Yeh et al. (2014), Taiwan	Ethnography	Interviews; Key informant	1	Yes	NS	Insider	NS	NS	Yes	Yes	Yes	Yes	Triangulation. Thematic coding.

NS = not stated; NA = not applicable; US = United States; UK = United Kingdom.

Approaches to Ethnography

Of the 39 studies, 22 were described as ethnographies or informed by ethnography (Arnold et al., 2015; Black, 2011; Chen et al., 2015; Cheyney et al., 2014; Danielsen, 2017; Finlay & Sandall, 2009; Flacking & Dykes, 2013; Hunter, 2010; Hunter & Segrott, 2010; Kemp & Sandall, 2010; Lange et al., 2016; Lee et al., 2009; Lindsay et al., 2012; Logsdon & Smith-Morris, 2017; Miltenburg et al., 2018; Olson & Couchie, 2013; Østergaard, 2015; Spendlove, 2018; Taylor et al., 2009; Thwala et al., 2011; Torres, 2014, 2015; Varcoe et al., 2013; Yeh et al., 2014), seven were focused ethnographies(Allen et al., 2015; Aubrey et al., 2017; Gagnon et al., 2013; Goodwin et al., 2018; Higginbottom et al., 2013, 2015; Hugill et al., 2013), four institutional ethnographies(Grassley et al., 2015; Kennedy et al., 2010, 2013; Véras & Traverso-Yépez, 2011), three critical ethnographies (Dove & Muir-Cochrane, 2014; Newnham et al., 2015, 2017a, 2017b; O’Boyle, 2014), two video ethnographies (Harte et al., 2016; Mondy et al., 2016) and one a feminist ethnography (O’Boyle, 2013).

While most studies (28/39) provided some description of the methodology, outlining some key features of ethnography, the level of detail provided varied considerably across studies. Eleven studies stated that their study was ethnographic or informed by ethnography without further explanation (Arnold et al., 2015; Black, 2011; Danielsen, 2017; Finlay & Sandall, 2009; Lange et al., 2016; Miltenburg et al., 2018; Olson & Couchie, 2013; Østergaard, 2015; Thwala et al., 2011; Torres, 2014, 2015; Varcoe et al., 2013). Only five studies articulated the epistemological position/theoretical framework of their chosen ethnographic approach (Allen et al., 2015; Newnham et al., 2015, 2017a; O’Boyle, 2013, 2014; Spendlove, 2018). Most commonly ethnography was described as a methodology that:

• attempts to understand the culture of a group of people or organization (Allen et al., 2015; Aubrey et al., 2017; Dove & Muir-Cochrane, 2014; Flacking & Dykes, 2013; Gagnon et al., 2013; Goodwin et al., 2018; Higginbottom et al., 2013, 2015; Newnham et al., 2017a, 2017b; Spendlove, 2018; Taylor et al., 2009; Yeh et al., 2014)

• studies people in their natural setting (every contexts) to understand actual behavior (Allen et al., 2015; Dove & Muir-Cochrane, 2014; Flacking & Dykes, 2013; Goodwin et al., 2018; Higginbottom et al., 2013, 2015; Hunter, 2010; Logsdon & Smith-Morris, 2017; Spendlove, 2018; Yeh et al., 2014)

• provides rich and detailed descriptions of actual behavior (Flacking & Dykes, 2013; Spendlove, 2018)

• analyses cultural norms and rituals (Dove & Muir-Cochrane, 2014; Goodwin et al., 2018; Yeh et al., 2014)

• seeks to understand a culture by the researcher immersing themselves into the culture (Hunter, 2010; Spendlove, 2018; Taylor et al., 2009)

• uses multiple data sources to allow for a holistic understanding (Higginbottom et al., 2015; Kemp & Sandall, 2010; Lee et al., 2009)

• seeks to understand and capture social meanings (Flacking & Dykes, 2013; Hunter, 2010; Hunter & Segrott, 2010)

• is interpretive (Flacking & Dykes, 2013; Spendlove, 2018) and views reality as socially constructed (Spendlove, 2018)

In those studies that provided a description of the approach, focused ethnography was described as a time-limited exploratory approach focused on a discrete group of persons, social phenomenon or sub-culture (Aubrey et al., 2017; Gagnon et al., 2013; Goodwin et al., 2018; Higginbottom et al., 2013, 2015), that is problem focused and context-specific (Higginbottom et al., 2013). Institutional ethnography was described as an ethnographic approach that describes how our daily experiences, processes or current practices, in particular those of everyday work, are mediated by institutional forces such as norms, regulations, reports, procedures, and discourses (Grassley et al., 2015; Kennedy et al., 2010, 2013; Véras & Traverso-Yépez, 2011). Critical ethnographies were described as including an examination of power and power imbalance (Newnham et al., 2015, 2017a, 2017b; O’Boyle, 2013, 2014) and a recognition of the cultural impacts on practice as dynamic processes rather than fixed entities (Dove & Muir-Cochrane, 2014). One of these studies adopted critical medical anthropology, defined as encouraging the examination and critique of the power relationships that influence normalized behaviors within the medical model (Newnham et al., 2015, 2017a, 2017b). Video-ethnography was described as the use of video recordings to collect an intensive, complex, and rich data set to provide insight into a cultural setting or practice (Harte et al., 2016; Mondy et al., 2016). The description of critical feminist ethnography was limited to: able to move beyond description toward a degree of social critique (O’Boyle, 2013) (See Table 1).

Over half of studies (20/39) provided no justification or explanation as to why ethnography was the chosen approach (Allen et al., 2015; Arnold et al., 2015; Aubrey et al., 2017; Black, 2011; Cheyney et al., 2014; Danielsen, 2017; Dove & Muir-Cochrane, 2014; Finlay & Sandall, 2009; Flacking & Dykes, 2013; Lange et al., 2016; Lee et al., 2009; Lindsay et al., 2012; Miltenburg et al., 2018; Mondy et al., 2016; O’Boyle, 2014; Olson & Couchie, 2013; Østergaard, 2015; Thwala et al., 2011; Torres, 2014, 2015; Varcoe et al., 2013) (see Supplemental File S1). The 19 studies that did provide a reason for using ethnography outlined a study aim that was aligned, to varying degrees, with ethnography. The stated aim of these studies included:

• To understand the culture of an organization or setting, or understand a phenomenon within its cultural context (Goodwin et al., 2018; Higginbottom et al., 2013, 2015; Logsdon & Smith-Morris, 2017; Newnham et al., 2015, 2017a, 2017b; Taylor et al., 2009; Véras & Traverso-Yépez, 2011; Yeh et al., 2014)

• To understand how people give meaning to their experiences (Goodwin et al., 2018; Hunter, 2010; Hunter & Segrott, 2010)

• To study the forces (rules, activities, discourses) that shape a person’s world or setting (Grassley et al., 2015; Kennedy et al., 2010, 2013)

• To describe “real life” experiences(Chen et al., 2015; Hugill et al., 2013; Hunter, 2010; Hunter & Segrott, 2010)

• To gain an insider perspective (Taylor et al., 2009)

• To include observation as a method (Harte et al., 2016; O’Boyle, 2013)

• To allow for data triangulation (O’Boyle, 2013)

• To provide a holistic or in-depth understanding (Hugill et al., 2013; Kemp & Sandall, 2010; Logsdon & Smith-Morris, 2017)

• To allow the researcher’s identity as a midwife and researcher to be reflexively incorporated into the process of data collection and analysis (Spendlove, 2018)

• Because of the linguistic, cultural, and social diversities of the populations examined(Gagnon et al., 2013)

Of the 20 studies that did not stipulate a reason for using ethnography, most (15/20) were relevant to ethnography to the extent that they either regarded the study of a culture or specific setting (Allen et al., 2015; Danielsen, 2017; Dove & Muir-Cochrane, 2014; Finlay & Sandall, 2009; Flacking & Dykes, 2013; Lange et al., 2016; Lindsay et al., 2012; Mondy et al., 2016; O’Boyle, 2014; Østergaard, 2015; Torres, 2014, 2015; Varcoe et al., 2013) and/or used observation was a key method (Allen et al., 2015; Arnold et al., 2015; Danielsen, 2017; Dove & Muir-Cochrane, 2014; Finlay & Sandall, 2009; Flacking & Dykes, 2013; Lange et al., 2016; Lindsay et al., 2012; Miltenburg et al., 2018; Mondy et al., 2016; O’Boyle, 2014; Olson & Couchie, 2013; Torres, 2014, 2015; Varcoe et al., 2013). The relevance of ethnography in the remaining five studies was unclear (Aubrey et al., 2017; Black, 2011; Cheyney et al., 2014; Lee et al., 2009; Thwala et al., 2011), as the stated aims were not clearly aligned with ethnography (e.g., to examine clinicians’ perspectives on a specific issue using interviews(Aubrey et al., 2017)) (see Supplemental File S1).

Data Collection

Sixteen studies were conducted at one site only, five at two sites, one at three sites, two at four sites, four at five, or more sites, and the number of sites were not defined in 11 studies (Table 1 and Supplemental File S1).

The majority of studies used more than one data collection method (30/39) and adopted an inductive approach to data collection (33/39). While most studies described participant groups that were key stakeholders, only two studies used the term key informants (Hunter, 2010; Yeh et al., 2014). Regardless of the method used, the majority of studies used field notes or memoing to record observation and non-verbal communication (33/39).

In terms of the specific methods, 37 studies used interviews, eight of which used interviews as the only method (Aubrey et al., 2017; Black, 2011; Chen et al., 2015; Cheyney et al., 2014; Higginbottom et al., 2013, 2015; Logsdon & Smith-Morris, 2017; Thwala et al., 2011). Seven studies used focus groups, of which six did not stipulate the duration, and one was very short (10 minutes) (Grassley et al., 2015). Seven studies used document analysis, described as an analysis of relevant documents, such as minutes of meetings and hospital (Allen et al., 2015; Hunter, 2010; Lee et al., 2009; Lindsay et al., 2012; Newnham et al., 2017b; O’Boyle, 2013, 2014). A total of 29 studies used observation, one of which used observation as the only method (Miltenburg et al., 2018).

In terms of the observation approach, 19 studies used participant observation, five used non-participant, two semi-participant observation and three did not specify. By and large, what was meant by participant versus non-participant observation was left undefined. Of the 10 studies (out of 29) that described the approach to observation used, descriptions were mostly brief and superficial. Only two studies, one that defines semi-participant observation (Flacking & Dykes, 2013) and one participant observation (Arnold et al., 2015), provided an adequate description (see Supplemental File S2).

The duration of the observation was also left largely undefined. Of the 29 studies, 10 did not provide any information in relation to the observation period, and eight provided vague descriptions such as “6 weeks of observation”(Arnold et al., 2015), “3 unit observations”(Grassley et al., 2015), or “6 months”(Kennedy et al., 2010). The remaining 11 studies stipulated the exact number of observation hours conducted, with the observation period overall brief (Supplemental File S2).

In relation to how observations were recorded, while most studies (24/29) stipulated that field notes were used, the level of detail was minimal, and only one study made reference to a framework/structure used to guide note taking (Flacking & Dykes, 2013). This study used the observational dimensions by Spradley consisting of: physical layout of the place, range of people involved, the activities that occur, the physical things that are present, actions people undertake, activities that people undertake, the sequencing of events that occur, things that people are trying to accomplish, emotions felt and expressed). Five studies did not state how observations were recorded (Harte et al., 2016; Kennedy et al., 2013; Lindsay et al., 2012; Olson & Couchie, 2013; Varcoe et al., 2013) (see Supplemental File S2).

Reflexivity

Nine studies stipulated the position of the researcher on the insider-outsider continuum (Allen et al., 2015; Flacking & Dykes, 2013; Hugill et al., 2013; Miltenburg et al., 2018; Mondy et al., 2016; O’Boyle, 2013, 2014; Spendlove, 2018; Yeh et al., 2014). In five of these studies the researcher was described as an insider (Miltenburg et al., 2018; O’Boyle, 2013, 2014; Spendlove, 2018; Yeh et al., 2014), in three as an outsider(Allen et al., 2015; Flacking & Dykes, 2013; Mondy et al., 2016), and one study commented that the team included insiders and outsiders (Hugill et al., 2013). This study described the position of the researchers as follows:

The lead author is an experienced neonatal nurse and father of preterm children and was the only one who had previously worked on the study unit. The other authors are all female and a midwife, neonatal nurse, and sociologist who brought their own experiences, perceptions, and views to the study, analysis, and representation of the data. This broad base of experience helped to balance insider and outsider perspectives (Hugill et al., 2013).

Eleven studies considered the relationship the researcher and participants and made reference to reflexivity, albeit to varying degrees (Aubrey et al., 2017; Goodwin et al., 2018; Hugill et al., 2013; Kemp & Sandall, 2010; Lindsay et al., 2012; Miltenburg et al., 2018; Mondy et al., 2016; Newnham et al., 2015; O’Boyle, 2013, 2014; Spendlove, 2018; Yeh et al., 2014). The depth of discussion ranged for a brief comment that “reflexive practice” was used (e.g., a reflexive journal was used or “issues of reflexivity were addressed”) (Aubrey et al., 2017; Lindsay et al., 2012; Miltenburg et al., 2018; Mondy et al., 2016), while other studies provided a more detailed discussion (Goodwin et al., 2018; Hugill et al., 2013; Kemp & Sandall, 2010; Newnham et al., 2017b; O’Boyle, 2013, 2014; Spendlove, 2018). For example:

Reflexivity took the form of critical self-reflection through that we aimed to account for our subjectivities and research conduct in a transparent manner, a sometimes challenging but ultimately productive process (Hugill et al., 2013).

Reflexive accounts were written and shared with a project support group to ensure that all potential personal and interpersonal influences were explored and considered appropriate (Goodwin et al., 2018).

Understanding one’s position is fundamental to the location of the frameworks of power in critical research. In this way, [the researcher] maintained a reflexive position, journaling thoughts and identifying potential analytic bias, repeatedly returning to the data, and following up discongruencies, which can lead to deeper research insights (Newnham et al., 2017b).

At the very least an attempt must be made to be transparent and reflective about ones subjectivity and the reflexive nature of all interactions in and with the field. I will endeavour then to include in this paper aspects of my presence in the scene as a midwife with my own professional interpretations and concerns (O’Boyle, 2014).

Data Analysis

An iterative and cyclical process to data collection and analysis was used in 18 studies (Table 1).

Data triangulation was reported by 15 studies, with overall little detail about the process of data triangulated provided (Cheyney et al., 2014; Flacking & Dykes, 2013; Gagnon et al., 2013; Harte et al., 2016; Kennedy et al., 2010, 2013; Lange et al., 2016; Lindsay et al., 2012; Logsdon & Smith-Morris, 2017; Mondy et al., 2016; Newnham et al., 2017b; O’Boyle, 2013; Spendlove, 2018; Varcoe et al., 2013; Yeh et al., 2014). In six studies triangulation was conducted by comparing the data from different methods (i.e., interviews, field notes, document reviews, focus groups) (Harte et al., 2016; Lange et al., 2016; Lindsay et al., 2012; Mondy et al., 2016; Newnham et al., 2017b; O’Boyle, 2013), and in four studies data triangulation referred to the comparison of findings between different study participant groups (Gagnon et al., 2013; Østergaard, 2015; Varcoe et al., 2013; Yeh et al., 2014). Which data sources or methods were triangulated was not defined in five studies (Flacking & Dykes, 2013; Kennedy et al., 2010, 2013; Logsdon & Smith-Morris, 2017; Spendlove, 2018).

Other approaches to analysis used included thematic analysis (15 studies), a process of constant comparison usually in reference to grounded theory (six studies) (Cheyney et al., 2014; Finlay & Sandall, 2009; Flacking & Dykes, 2013; Logsdon & Smith-Morris, 2017; Spendlove, 2018; Taylor et al., 2009), content analysis (three studies) (Chen et al., 2015; Grassley et al., 2015; Mondy et al., 2016), an inductive process following the framework for the analysis of ethnographic data by Roper and Shapira (three studies) (Allen et al., 2015; Higginbottom et al., 2013, 2015), an inductive process without reference to a specific framework (two studies)(Dove & Muir-Cochrane, 2014; Østergaard, 2015), narrative analysis (two studies) (Kennedy et al., 2010, 2013), content analysis (two studies), and “thick description” as a way of providing cultural context and meaning (one study) (Newnham et al., 2017b). In six studies the process to analysis was not described (Black, 2011; Danielsen, 2017; Lee et al., 2009; O’Boyle, 2014; Thwala et al., 2011; Véras & Traverso-Yépez, 2011).

A total of 26 studies included some level of theoretical development, and used theory to interpret findings and/or contribute to theory (Table 1).

Discussion

This review identified only one study that described the ethnographic approach used in detail, and covered the key features of ethnography. In most of the studies the methodology of ethnography was not clearly described, and why ethnography was selected as the most appropriate approach was not consistently articulated. Only five studies described the epistemological position/theoretical framework of their chosen ethnographic approach (Allen et al., 2015; Newnham et al., 2015, 2017a; O’Boyle, 2013, 2014; Spendlove, 2018). Eleven studies stated that their study was ethnographic or informed by ethnography without further explanation (Arnold et al., 2015; Black, 2011; Danielsen, 2017; Finlay & Sandall, 2009; Lange et al., 2016; Miltenburg et al., 2018; Olson & Couchie, 2013; Østergaard, 2015; Thwala et al., 2011; Torres, 2014, 2015; Varcoe et al., 2013).

There was overall little evidence that an in-depth understanding of the problem or culture was developed over time, consisting of multiple phases of data collection and analysis. While most studies described an inductive approach to data collection (33/39), less than half (18/39) described the ethnographic process of data collection and analysis as conducted concurrently in an iterative and cyclical manner (Francis, 2013; Reeves et al., 2013). In many of the studies, the process of data collection and analysis appeared more akin to inductive qualitative research than traditional ethnography. Furthermore, only two studies (Hunter, 2010; Yeh et al., 2014) used the ethnographic data collection method of informal conversations with key informants (Adler & Adler, 1994; Francis, 2013; Hammersley & Atkinson, 2007; Morgan et al., 2016). Instead, insider perspectives were derived through more formal traditional interviews. Another key method used in ethnography that was less featured was document analysis (Francis, 2013).

Consistent with the prominence of observation to ethnography, the majority of studies included observation as a method. However, overall little detail was provided around how observation were conducted, by whom, how they were recorded, and the number of hours of observation. When stipulated the observation period was relatively short. Traditionally, ethnography is associated with prolonged periods of observation (Atkinson & Hammersley, 1994). While certain approaches to ethnography, specifically focused ethnography, are more time-limited (Francis, 2013), only seven of the studies were described as such (Allen et al., 2015; Aubrey et al., 2017; Gagnon et al., 2013; Goodwin et al., 2018; Higginbottom et al., 2013, 2015; Hugill et al., 2013). The majority of time-limited studies included in this review were described as “ethnographic” (i.e., not as “focused ethnographies”).

The process of reflexivity and consideration of the researcher on the insider-outsider continuum were also not reported by many studies. Nine studies stipulated the position of the researcher on the insider-outsider continuum (Allen et al., 2015; Flacking & Dykes, 2013; Hugill et al., 2013; Miltenburg et al., 2018; Mondy et al., 2016; O’Boyle, 2013, 2014; Spendlove, 2018; Yeh et al., 2014), and 11 made reference to reflexivity, mostly superficially. There was overall little evidence that the inevitable impact of the researcher’s identity on the research process was considered. Furthermore, while most (26/39) studies had some theoretical development, only three studies made reference to the underpinning theoretical framework of ethnography as seeking to understand and capture social meanings (Flacking & Dykes, 2013; Hunter, 2010; Hunter & Segrott, 2010). While the majority of studies used multiple methods and/or sources of data, only 15 studies made reference to data triangulation, mostly superficially.

While there was often omission of some ethnographic features such as reflexivity and data triangulation, the underuse of other features are easier to understand. Given that features such as reflexivity and data triangulation are relatively easily implemented and improve study quality and rigor without adding huge costs or ethical challenges (Bunce et al., 2014; Denzin, 1970; Morgan et al., 2016; Murphy & Dingwall, 2007; Reeves et al., 2013), omission leaves gaps and unanswered questions for the reader. On the other hand, the benefit of conducting interview versus key informant consultations, and time-limited observation periods over prolonged immersion in the field are easier to appreciate. The limited use of key informant conversation can likely be explained by requirements from human research ethics committees, who may not place the same value on the collection of unstructured data through more informal ways as ethnographers do (Hammersley & Atkinson, 2007; Liamputtong & Ezzy, 2005). It is likely that ethics approval is more forthcoming for more structured interviews, although there is evidence of barriers to ethics approvals at times with qualitative research (Newnham et al., 2013). Evidence of prolonged immersion in the field with long periods of observation is likely missing as it is time-consuming, costly, and often not practical in healthcare settings (Bunce et al., 2014; Savage, 2000). Given research funding realities and a range of practical considerations, periods of observation are often time-limited (Bunce et al., 2014; Francis, 2013; Savage, 2000).

The use of a time-limited approach to ethnography is increasingly accepted, referred to as focused ethnography (Francis, 2013). Many of the studies included here, that were not defined as focused, appear to correspondent more with focused ethnography than traditional ethnography. Furthermore, a number of these studies do not include key features of ethnography (focused or not), and were perhaps best described as observational studies. Fifteen studies provided no justification or explanation for using ethnography, except for the use of observation (Allen et al., 2015; Arnold et al., 2015; Danielsen, 2017; Dove & Muir-Cochrane, 2014; Finlay & Sandall, 2009; Flacking & Dykes, 2013; Lange et al., 2016; Lindsay et al., 2012; Miltenburg et al., 2018; Mondy et al., 2016; O’Boyle, 2014; Olson & Couchie, 2013; Torres, 2014, 2015; Varcoe et al., 2013).

Within the context of the maternity unit, ethnographic studies can capture practices that may not be visible when using traditional non-observational research (Newnham et al., 2021). For example, how women deal with maintaining their privacy in a constrained hospital environment cannot necessarily be examined to its full extent with a pre-determined quantitative method such as a survey. There are also issues with ethnographic observation techniques in maternity settings due to the private nature of the care (i.e., women may reject an observer being present during intimate conversations with carers or during examinations/birth). There may be a risk that the researcher becomes over-involved in clinical or social and emotional care of participants, which may jeopardize the rigor of a study (Newnham et al., 2021), although this may be appropriate when needing to step in to prevent harm. When midwives conduct research in maternity settings, an ethnographic “middle ground” approach, where the researcher takes on the role of both insider and outsider, may be required (Burns et al., 2012). Deep reflexivity and planning of appropriate ethnographic techniques are essential for success within these settings.

The value of observational studies for gaining insight into healthcare practices and behavior (within an ethnographic framework or not), is increasingly recognized (Bunce et al., 2014; Mays & Pope, 1995; Morgan et al., 2016). Nonetheless, despite well-documented advantages of observation over other forms of qualitative data collection, particularly its ability to record what people do and how they act, rather than just what they say, observation methods continue to be underused in healthcare research (Mays & Pope, 1995; Morgan et al., 2016). Like others before us, we suggest that observational data can be positioned as the central method of a research design, as a more pragmatic alternative to traditional ethnography suitable for health settings, including maternity care (Mays & Pope, 1995; Morgan et al., 2016).

While we suspect that some healthcare researchers using observational data as a primary method perceive ethnography as the most appropriate methodology, we suggest that methodologies such as case study research (Morgan et al., 2016; Yin, 2014) or realist evaluation (Pawson & Tilley, 1997, 2004) may sometimes be more suitable alternatives. Like ethnography, both these methodologies are well suited to the study of complex healthcare settings and use multiple data sources and methods and an inductive and cyclical approach to data collection and analysis to gain deeper insight into social contexts (Lalor et al., 2013; Morgan et al., 2016; Pawson & Tilley, 1997, 2004; Rosenberg & Yates, 2007; Yin, 2014).

Strengths and Limitations

Strengths of this systematic review lie in its methodology, which followed the PRISMA reporting guideline, and the inclusion of 10 years of maternity-related research using ethnography. A limitation of this paper is that we developed and used our own, unvalidated data extraction platform for ethnographic studies due to the lack of available tools. In addition, data extraction was limited to what was reported, and reporting may have been limited by factors such as journal guidelines and word count restrictions. A further limitation is that the keywords used to identify studies may not have captured all relevant studies within this time period, and publications that were not in English were excluded. Our findings were limited to maternity care where ethnographic methods might not be conventionally used, and studies published since 2018 were not included, thus this review may not reflect new developments in ethnography in maternity care. Despite these limitations, we believe that the 39 studies identified was sufficient to achieve our aim and identify the extent to which the key principles of ethnographies were used or reported on maternity care. Nurses, midwives and other healthcare providers, when considering the use of ethnography to address maternity, or any other health services, to advance practice or knowledge, could benefit from the findings of this study.

Conclusion

This review highlights the inconsistencies and often poor reporting of ethnography methodology in studies of maternity services. In particular, the justification for using the methodology, the relationship of the researcher to ethnography, its underpinning frameworks, and how data were triangulated. There is a need to develop reporting guidelines to guide researchers undertaking and reporting on ethnographic research.

Click here for additional data file.

Supplemental material, sj-pdf-1-gqn-10.1177_23333936211028187 for The Use of Ethnography in Maternity Care by Dominiek Coates and Christine Catling in Global Qualitative Nursing Research

Click here for additional data file.

Supplemental material, sj-pdf-2-gqn-10.1177_23333936211028187 for The Use of Ethnography in Maternity Care by Dominiek Coates and Christine Catling in Global Qualitative Nursing Research