Five Approaches to Qualitative Comparison Groups in Health Research: A Scoping Review

Healthy comparison

This first type of qualitative comparison group involved comparing those who have an illness or chronic condition to healthy participants, which was found in eight studies. Analytic processes used within this type of comparison included reading all transcripts before comparing and contrasting within and between groups (Lehmann, 1998; Makela et al., 2009); mixing the order of transcripts while coding, then noting similarities and differences between the groups (Dickie et al., 2009); coding within each group before comparing across groups (Ludvigsson et al., 2015; Reedy et al., 2005; Sandelowski et al., 1992); and a framework approach to comparison (Lindsay et al., 2015).

For example, in Lehmann’s (1998) study, they explored the role that mothers play in shaping their adolescent’s future through in-depth interviews. Their sample consisted of 40 mothers (20 mothers of children with a severe disability; 20 mothers of children with no disability label but attended vocational programs). The groups were homogeneous by gender, age, and ethnicity. Applying a thematic analysis, they compared and contrasted the themes between and within the groups (Lehmann, 1998). Their results showed that the perceptions of mothers of adolescents with severe disabilities about the intensity of their roles pose a dilemma regarding developing self-determination (Lehmann, 1998). Both groups of mothers had many of the same child-rearing tasks; however, differences occurred in the amount of support that children with disabilities needed (Lehmann, 1998). Having a comparison group helped them to understand the extent of similarity between the groups and clarified the nature of child rearing for mothers of individuals with disabilities (Lehmann, 1998).

In another example, Makela et al. (2009) explored the value of genomic hybridization to diagnose the cause of intellectual disabilities. Parents of 20 children with intellectual disabilities (10 with and 10 without a causal diagnosis) were interviewed in depth about their value of such a diagnosis). For their thematic analysis, they compared within and between the two groups to assess difference in attitudes or values (Makela et al., 2009). They found no differences between the groups regarding parents’ perceptions or experiences related to the presence or absence of an etiological diagnosis for their child’s condition (Makela et al., 2009).

Another example of a healthy comparison group included Dickie et al. (2009) who explored parents’ sensory experiences of children with and without autism using in-depth interviews. Their sample included parents of 66 youth (37 with autism and a comparison group of 29 typically developing children). Both groups were recruited from different sources (Dickie et al., 2009). It is important to note that the groups varied in terms of parental age, educational level, and mental age of the youth. They mixed the order of the transcripts during the coding process before comparing and contrasting the themes within and between the groups. They reported that having a comparison group helped them to discover that sensory experiences of children with autism overlapped with typically developing children (Dickie et al., 2009).

Three studies using a healthy comparison group analyzed their data by coding within each group before comparing across groups (Ludvigsson et al., 2015; Reedy et al., 2005; Sandelowski et al., 1992). For example, Ludvigsson et al.’s (2015) comparative study highlighted that depression in old people might not be related to pathology, but rather to normal aging. They conducted semistructured interviews to compare the experiences of subsyndromal depression in 13 old people compared with eight old people with syndromal depression, and six with nondepression. For their analysis, they applied the whole coding scheme without knowledge of which group participants belonged to avoid preconceptions. They then conducted a separate analysis within each group before comparing and contrasting across groups (Ludvigsson et al., 2015). They found that subsyndromal depression differed qualitatively from syndromal depression but nondepression. Their results highlight that overlooking psychosocial aspects may pose a risk of improper diagnosis of depressive disorders (Ludvigsson et al., 2015).

Meanwhile, Reedy et al. (2005) conducted in-depth interviews to explore 10 colorectal cancer survivor’s beliefs about diet, supplements, health, and cancer in relation to a comparison group of 12 participants without cancer. They acknowledged that “using a comparison group may not be traditional in qualitative research” but that their exploratory process helped to understand whether dietary supplement use decisions were attributed only to the cancer experience or more general aging or other chronic disease-related experiences (Reedy et al., 2005). They first analyzed all of the themes separately for each group, then compared the themes by group, and then with all of the data combined (Reedy et al., 2005). Their findings showed that the experience of colorectal cancer may lead to dietary change among some survivors.

In another study, Sandelowski et al., (1990) examined infertility through multiple interviews with 53 infertile couples and a comparison group of 10 couples with no fertility impairments and three infertile couples 1 to 2 years after having their last child. They used theoretical sampling procedures with multiple open-ended and focused interviews with a constant comparison technique. It is important to note that they did not specify whether they analyzed all of the transcripts together or separately.

In a similar study by Sandelowski et al. (1992), they explored the work of pregnancy for infertile couples by interviewing 60 couples (41 infertile and a comparison group of 19 fertile) expecting a child. They used theoretical and selective sampling strategies, seeking couples who had a range of experiences with infertility and to maintain a homogeneous sample (by ethnicity, class, marital status, and age) to achieve maximum variability in fertility (Sandelowski et al., 1992). They compared and contrasted findings within and between groups and found that infertility was important but only one of several factors altering the experience of childbearing (Sandelowski, et al., 1992).

One study within using the healthy comparison group applied a framework approach to comparing differences within and between groups. In Lindsay et al.’s (2015) study, they explored how 13 youth with a physical disability experience winter compared with 12 youth without disabilities while using journals, photos, and semistructured interviews. Their comparison group involved youth without disabilities that closely matched the group with disabilities (e.g., by gender, age, and location). They used a thematic analysis where they compared the findings within and between the groups. Their findings showed that youth with disabilities have many similar challenges during winter compared with youth without disabilities, but to a greater extent. Meanwhile, youth with disabilities reported more challenges going outdoors in the winter and experiencing negative psychosocial impacts including loneliness and increased dependence compared with youth without a disability (Lindsay et al., 2015). In sum, having a qualitative comparison group that consists of healthy controls could help researchers to understand areas of similarity and difference among those with a health condition.

No intervention or treatment

Four studies in this review involved qualitative comparison groups with participants who received an intervention and those who did not. It is important to highlight that these studies involved qualitative-only designs (i.e., not mixed methods nor part of an RCT). Comparative analysis methods using this type of approach included coding all transcripts before comparing and contrasting between groups (Lindsay et al., 2016); researchers blind to which group participants belonged to before comparing and contrasting between groups (Gorawara-Bhat et al., 2017; Lysaker et al., 2015); and using matrices or frameworks to organize themes while comparing similarities and differences (Heberlein et al., 2016; Lindsay et al., 2016).

Three studies within this type of comparison group coded all transcripts as a whole before comparing and contrasting within and between groups (Gorawara-Bhat et al., 2017; Lindsay et al., 2016; Lysaker et al., 2015). For example, Lindsay et al. (2016) explored how nine youth and 11 parent’s experiences of a new inter-agency transition model for spina bifida compared to 12 youth who did not take part in the model (i.e., comparison group) by conducting semistructured interviews and a descriptive thematic analysis. They used a framework approach to help organize the themes and compare/contrast within and between groups. They found that most youth and parents felt supported by clinicians and benefited from the transition model. The comparison group reported some support from clinicians but also experienced gaps in their continuity of care (Lindsay et al., 2016).

In Lysaker et al.’s (2015) study, they explored whether metacognitively focused individual psychotherapy can affect self-experience through conducting narrative interviews. They compared 12 participants in metacognitve-oriented psychotherapy with 13 participants in supportive psychotherapy using a thematic analysis. Participants reported that psychotherapy led to improvements in self-esteem and ability to think clearer (Lysaker et al., 2015). Meanwhile, the metacognitive-oriented therapy group integrated their experiences into the larger narratives of their lives, increased sense of agency, and ability to manage pain compared with supportive therapy (comparison) group.

Gorawara-Bhat et al. (2017) explored patients’ attitudes toward recurrent prostate cancer and starting hormone therapy treatment among two groups (i.e., decision-aid, experimental group, n = 13) and standard care (i.e., comparison group, n = 13). They conducted in-depth semistructured interviews 1 week post consultation and analyzed the transcripts using thematic analysis. Researchers were blind to which groups participants belonged to during the coding process. Recruitment continued until thematic saturation was reached. Their findings showed that patients who received the decision-aid (i.e., intervention) had better comprehension of prostate-specific antigen, and an improved understanding of hormone treatment implications, external locus of control, and participation in shared decision-making and support seeking (Gorawara-Bhat et al., 2017). In contrast, the comparison group displayed worse comprehension of treatment implications, internal locus of control, and involvement in knowledge seeking and decision-making. They mentioned that having a comparison group helped them to understand how their tool worked (Gorawara-Bhat et al., 2017).

Meanwhile, Heberlein et al.’s (2016) study developed a framework of women’s prenatal care experiences by comparing 14 women in individual care with 15 women in group prenatal care (Heberlein et al., 2016). They conducted semistructured interviews using grounded theory to analyze the data. They found that participants had similar benefits in prenatal care but that the group participants received more benefits for education that were enhanced by a supportive environment. Heberlein et al. (2016) noted that their comparative design was novel and contributed to the literature by allowing them to explore a range of experiences of women while contrasting two prenatal models. In sum, having a qualitative comparison group that compares two different groups to assess the impact of an intervention sheds light on the key ingredients of the intervention.

Comparing different conditions or groups

Another type of comparison group, found in five studies in this review, included comparing two different conditions or groups. Of the studies that described their comparative analysis their approach included coding across all transcripts before comparing between groups (Askew, 2009; Moola, 2012; Whitley, 2016); and one used a framework approach to compare within and between groups (Gysels & Higginson, 2011).

For example, Askew (2009) explored why women chose hysterectomy over other treatments. They compared eight women with myomectomy with 10 women who had a hysterectomy by conducting semistructured phone interviews. Using a thematic analysis, they found that the two groups had different factors affecting their choice of surgery. They described how the comparisons helped to improve their understanding of how hysterectomy rates can be reduced (Askew, 2009).

In McAllister and Silverman’s (1999) study, they explored the process of community formation and maintenance of community roles among individuals suffering from dementia in two different institutional settings (34 participants in a nursing home; and 59 participants in an Alzheimer’s facility). Through their comparative examination using ethnographic observations, they identified facilitators and barriers to the development of community and roles that helped to increase their understanding of how the specialized features of the one facility contributed to community. Although they had differences in the setting and population examined, they found similarities in the formation of community (McAllister & Silverman, 1999). They did not describe how they compared and contrasted within and between groups in their analysis.

In another example, Moola (2012) compared the experiences of parenting youth with either cystic fibrosis or congenital heart disease. By having a comparison group, she argued that their study offered a “novel contribution to the literature” (Moola 2012, p. 212). She explored two conditions that differed greatly in terms of the trajectory and treatment—making them an intriguing line of inquiry. She found that pediatric caregiver stress differed between cystic fibrosis and congenital heart disease and that temporal dilemmas were unique sources of stress for cystic fibrosis parents only (Moola, 2012). She reported that having a comparison group facilitated a better understanding of the total life experiences by considering the whole person rather than simply focusing on the condition (Moola, 2012). Furthermore, focusing on the similarities between chronically ill children could help to enhance a continuum-based approach to understanding illness that can be less stigmatizing (Moola, 2012). She argues that a comparative approach could enable theory development while illustrating the context-specific idiosyncrasies that characterize the unique experiences of parents of children with chronic conditions. Comparison also allowed for exploration of different illness trajectories (Moola, 2012).

Meanwhile, Whitley (2016) explored the ethnoracial variation in recovery from severe mental illness. They conducted interviews with 19 Caribbean Canadian and 28 Euro-Canadians while applying a thematic analysis where they compared themes within and between groups. They found that both groups defined recovery as a gradual process and that stigma, financial strain, and hospitalization hindered recovery. They noted that religion acted as a facilitator for the Caribbean Canadian group (Whitley, 2016). Using a comparative approach helped them to understand similarities and differences within different racial groups.

One study using this type of comparison group applied a framework approach to their analysis. For example, Gysels and Higginson (2011) compared the experience of breathlessness among 48 patients, including 10 with cancer, 18 with COPD, 10 with heart failure, and 10 with motor neuron disease. They chose these four conditions because they share heavy symptom burdens, poor prognosis, high breathlessness, and palliative care needs (Gysels & Higginson, 2011). They first focused on within-group experiences in relation to the components of their framework for their analysis. Then, they compared and contrasted experiences across the different conditions. Their comparative case-based approach used an interpretive model to analyze the data (Gysels & Higginson, 2011). Their findings highlighted that all conditions shared the disabling effects of breathlessness, but differences occurred between the four conditions. Having a comparison group helped them to understand similarities and differences between the groups, why integrated palliative care is needed, appropriate therapeutic options, and collaborative efforts for these patient groups. In sum, comparing different health conditions or groups within using a qualitative approach can help showcase similarities and differences while highlighting areas where further support or intervention may be needed.

Comparing different aspects of a condition or phenomenon

Six studies within this review had qualitative comparison groups that compared different aspects of a health condition or phenomenon. Comparative analysis approaches used within this type of comparison included developing codes for both groups at the same time, then comparing within and between groups (Carrard & Kruseman, 2016; Fields et al., 2017; Hesketh et al., 2012); developing codes for each group separately, then comparing across groups (Bailey, Ben-Shlomo, Tomson, & Owen-Smith, 2016; Chambers et al., 2011); and applying a framework to explore similarities and differences (Jaye & Tilyard, 2002; Townsend et al., 2008).

For example, in Carrard and Kruseman’s (2016) study they compared the role of self-weighing as a strategy of weight control for weight-loss maintainers in comparison with a normal, stable weight group. They recruited two groups of 18 participants in each group (weight-loss maintainers; and normal stable weight) and conducted in-depth interviews. Groups were paired by gender, age, and socioeconomic status. They used a descriptive analysis and compared the themes within and between the groups. Their findings showed that most weight-loss maintainers needed regular self-weighing to be aware of their weight (Carrard & Kruseman, 2016). Changes in weight generated negative and positive affects among weight-loss maintainers. In contrast, the comparison group rarely used self-weighing and needed fewer strategies. They concluded that regular self-weighing as a component of weight-loss maintenance should be encouraged to help those trying to maintain their weight. Their comparative design highlighted areas where further support is needed.

In Field’s, et al. (2017) study, they had a qualitative comparison of barriers to antiretroviral medication adherence among perinatally (n = 18) and behaviorally (n = 12) HIV-infected youth. They first identified themes across all transcripts and then compared the common and unique barriers across the groups. They found that perinatally infected youth barriers included reactance, complicated regimens, HIV fatigue, and difficulty transitioning to autonomous care. The behaviorally infected group encountered barriers including HIV-related shame and difficulty initiating medication. Both groups had low risk perception, medication as a reminder of HIV and nondisclosure but described different contexts related to these barriers (Fields, et al., 2017). Having a comparison group helped them to develop a customizable intervention that addressed barriers and their psychosocial antecedents (Fields et al., 2017).

Meanwhile, in Hesketh’s et al. (2012) comparative study they explored how parents view their role in shaping physical activity and screen time behaviors. They conducted 16 unstructured focus groups including eight with new parents (61 parents in total) and eight with parents of preschool children (36 parents in total). They used a grounded theory approach to analyze the data and found that both groups had concerns and strategies; however, new parents had more optimism (Hesketh et al., 2012).

Two studies within this type of comparison group developed their codes for each group separately before comparing across groups. For example, Bailey et al. (2016) explored reasons for the socioeconomic disparity in live-donor kidney transplantation. They conducted interviews with a purposive sample of deceased-donor transplant recipients for 19 participants in high socioeconomic deprivation areas and a comparison group of 13 participants in low socioeconomic deprivation areas. They aimed for maximum diversity sampling in terms of age, gender, ethnicity, and primary renal disease. Bailey et al. (2016) conducted semistructured, face-to-face interviews and used a grounded theory approach for their analysis. Their findings showed themes related to the high socioeconomic disparity individuals and how they were linked to individual’s lack of confidence and skill in managing their health (Bailey et al., 2016). Having a comparison group helped give them insight into whether the same barriers to transplantation were encountered and how they were overcome.

Meanwhile, Chambers et al. (2011) explored factors affecting medication adherence in 26 stroke patients among low (n = 13) adherers and high adherers (n = 13). Through in-depth interviews and thematic analysis, they found two main themes including the importance of stability of a medication routine and beliefs about medication and treatment. High adherers remembered to take their medication and sought support (Chambers et al., 2011). Low adherers forgot their medication or sometimes intentionally did not take it and had less support. They mentioned that having a comparison group highlighted that interventions should be designed to target both intentional and nonintentional adherence to maximize medication adherence in stroke patients (Chambers et al., 2011).

Two studies using this type of comparison group applied a framework approach to analyze similarities and differences. For instance, Jaye and Tilyard (2002) conducted a qualitative comparative investigation of variation in general practitioners’ general practitioners (GPs) prescribing patterns. They had a sample of 60 general practitioners (comprised in 20 low, 20 medium, and 20 high-cost prescribers) using a qualitative comparison group to understand variations in prescribing between general practitioners. Their comparison focused on variations in the types of drugs, volumes, and cost. They structured their interview guide to allow for comparison across the groups. Their thematic analysis explored the themes within and between the groups (Jaye & Tilyard, 2002) where they also explored the similarities and differences. They found notable differences between low, medium, and high-cost prescribers.

Townsend et al. (2008) explored the reasons for frequent consultation among 23 people with multiple morbidity but contrasting consulting rates (high versus low consulters) using in-depth interviews. They used a grounded theory approach and also a framework in the early stages of the analysis to synthesize their themes and to compare within and between groups. They reported differences between the two groups including that frequent consulters had more disruptive symptoms resistant to self-management and needed more monitoring for unstable conditions. In sum, comparing different aspects of a condition or phenomenon sheds light on areas where further support or intervention may be needed.

Multiple perspectives of the same phenomenon comparison

This review found seven studies using qualitative comparison groups involving multiple perspectives of the same phenomenon. Of the studies that described, the type of comparative analysis used within this type of comparison group they included coding all transcripts before comparing within and between groups (Lim et al., 2013), analyzing groups separately before comparing between groups (Lindsay et al., 2016; Peterson, 2010), and a framework approach (Kyle et al., 2012; Schraeder et al., 2017; Sublette et al., 2017).

For example, Lim et al.’s (2013) study provides insight into different perspectives related to changes in health behaviors by comparing ethnicity (21 Chinese American; 11 Korean-American; 10 Mexican American). Doing so is critical for developing culturally tailored behavioral interventions to improve underserved breast cancer survivor’s quality of life. They conducted separate focus groups for youth and parents and used a thematic analysis where they coded all transcripts before comparing within and between the groups. Lim et al. (2013) reported several differences between the groups. For instance, ethnic and cultural differences were observed within amount of food consumed, use of alternative medicine, and need for cancer support groups. Their findings highlight the importance of understanding perceptions and beliefs regarding changes in health behaviors by underserved breast cancer survivors (Lim et al., 2013).

Two studies analyzed the groups separately before comparing between groups (Lindsay et al., 2016; Peterson, 2010). Lindsay et al. (2016) conducted semistructured interviews to explore the experiences of four young men with Duchenne muscular dystrophy, five parents, and seven clinicians who support them as they transition to adult care. They applied an open coding thematic analysis while comparing and contrasting the themes within and between the groups. They found several similarities and differences between the groups. Using a comparative approach helped them to highlight what is working well with transitioning youth to adult care needing further support (Lindsay et al., 2016).

Meanwhile, Peterson (2010) explored parent and adolescent perceptions about alcohol, tobacco, and drug use by conducting six focus groups (involving 38 adolescents) and one parent group (involving 11 parents). They used an open coding approach to identify the themes within and between the groups. Parents and youth had contrasting views. For example, youth want more prevention programs and positive adult role models while parents commented on the presence of peer pressure.

Three studies used a framework approach to analyze the differences within and between the groups (Kyle et al., 2012; Schraeder et al., 2017; Sublette et al., 2017). For example, Kyle’s et al. (2012) comparative case studies explored enablers and barriers to integrating community teams with urgent (n = 10) and emergency care (n = 12) involving health professions. They used a thematic analysis framework approach to analyze the data. Their study helped them to realize that they need integrated community nursing teams at multiple points in the urgent care system to provide an alternative to inappropriate emergency department admissions (Kyle et al., 2012). They noted that a strength of using a comparative case study approach is that complex phenomena can be understood by drawing on multiple perspectives (Kyle et al., 2012).

Schraeder et al. (2017) explored 10 parent and 10 youth perspectives on childhood mental health problems. Interviews were conducted separately for youth and parents. Using a constructivist grounded theory approach, they created matrices to analyze categories and make comparisons within and between the participant groups (Schraeder et al., 2017). Having a multiple perspectives comparison group assisted with highlighting how youth and parents varied in their experiences and perspectives and highlighted where further supports and interventions are needed (Schraeder et al., 2017).

Another example included Sublette et al.’s (2017) study where they conducted semistructured interviews involving 20 patients with hepatitis C and 20 of their health care professional’s perceptions regarding the facilitators and barriers to hepatitis C treatment adherence and completion. They used a comparative analysis, matrix-based approach to develop the thematic categories and analyze the data. They found that patients and health care professionals experienced communication difficulties that negatively impacted patients’ treatment experience. Sublette et al. (2017) reported that having a multiple perspectives comparison group was “instrumental in improving treatment outcomes for patients” (Sublette et al., 2017, p. 1309). Furthermore, a comparative approach helped them to highlight gaps in understanding between patients and professionals while challenging assumptions of patient preferences and allowing for better alignment of patient-centered care (Sublette et al., 2017).

In another study, Kyriacou et al. (2009) compared views of six patients, 12 parents, and 12 clinicians regarding emotions in anorexia using focus groups. By applying a thematic analysis, they found many similarities among patients and clinicians. Their qualitative comparison assisted them with developing an emotion and social cognition module for inpatient treatment (Kyriacou et al., 2009). It is important to note that they did not specify how they compared themes within and between the groups.