Paradigm and methodology

This study employs a grounded theory (GT) methodology. Developed by sociologists Glaser and Strauss,24 GT is widely used in health and social sciences to generate theoretical accounts of social phenomena. Described as the ‘the most comprehensive qualitative research methodology available’,25 ^(p1) GT is appropriate when research aims to explain a process where the concerns of those involved are central to its understanding and cannot be predetermined.24 26

GT methodology is distinguished by its core strategies: recursive study design, theoretical sampling and constant comparative system of analysis.27 Data collection and analysis occur iteratively, results of one cycle informing the next. Throughout this process, data sets are constantly compared with each other and against developing conceptualisations to generate successively more abstract concepts. Sampling proceeds on theoretical grounds whereby hypotheses generated from data guide further data sampling for theory refinement. Theoretical sampling continues to ‘saturation’, the point at which further data collection provides no new information for theory development.24 Theories thus ‘grounded’ in data are ‘mid-range’, lying between hypothetical statements about narrowly defined phenomena and highly abstract, all-encompassing ‘grand’ theories.28

GT methodology has ‘evolved’ since its initial description. There are now several variants.29 While core features are retained, different versions reflect their proponents’ world views and adopt differing positions related to timing in the use of literature, the researchers’ role and specific analytical techniques. This study employs Corbin and Strauss’s29 approach, within a pragmatic paradigm developed by Rorty.30 Consistent with the tenet of pragmatic epistemology, that there are no fixed points from which reality can be observed, researchers and participants are understood as interactively constructing knowledge during the research process.29 31 Resultant theory acknowledges ‘temporal, cultural and structural context’31 ^(p524) and conditions that give rise to the phenomenon. The theory is understood as provisional, its value dependent on its relevance to real-world problems and capacity to support change.

Study team

The study team comprises an early career researcher (Author 1) and three clinician-researchers with complementary expertise in the topic, recovery and methodology. The study is coordinated by author 1 (‘researcher’), a PhD candidate and registered psychologist with 14 years’ experience working clinically within specialist youth mental health services. Authors 3 and 4 are experienced academics with backgrounds in clinical and educational/developmental psychology. Author 2, also trained as a psychologist, is a health services researcher with substantial experience in qualitative research including GT.

Study setting

This study is being conducted in Queensland, Australia, with participants recruited from specialised Child and Youth Mental Health Services (CYMHS) funded by the Queensland Government. Delivered by medical, nursing and allied health professionals in multidisciplinary teams, CYMHS provide specialist assessment and treatment for young people (<18 years old) experiencing substantial distress and disruption in daily functioning, typically meeting diagnostic criteria for severe or complex mental disorders. Consistent with national and state ‘recovery-oriented’ policy, CYMHS operate on the premise that young people can and do recover from mental disorders.32

Participants and recruitment

Participants are adolescents attending CYMHS in four study sites selected to represent geographic (regional/rural/inner-city), socioeconomic and cultural variation. Eligibility criteria (box 1) are intentionally broad to promote recruitment of a sample with varied experiences.

While sample size in GT is determined on the basis of credibility in context of the purpose of inquiry,33 practicalities and time require an estimation. Based on team members’ experiences and Cresswell’s28 recommendations for achievement of saturation, we aim to recruit 30 participants. Recruitment commenced July 2015 continuing through January 2017. At the time of writing, 19 participants have been recruited and data collection is continuing.

Recruitment follows procedures outlined in figure 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Study recruitment process

Data collection

GT is a flexible methodology29 that does not delimit the types of data acquired, techniques of collection or the way they are used. Indeed use of different kinds of data is recommended, providing alternate vantage points on the phenomenon.24 In this study, primary data are participants’ first-person accounts, supplemented by two self-report quantitative measures of mental health. Data are collected in semi-structured interviews conducted at two timepoints, first at recruitment then 3–6 months later, depending on participants’ availability and engagement with CYMHS. Interviews are conducted by the researcher in private rooms at CYMHS, or other suitable facilities. The researcher uses a conversational approach following leads in participants’ accounts and participants’ language to encourage articulation of experiences and views.

Data collection is supported by flexible use of a topic guide (box 2). The topic guide was informed by a review of literature related to young people’s experiences of ‘recovery’,23 34 interviewing young people,35 36 interviewing for GT26 37 and the study team’s clinical and research experiences with young people. Two visual aids, employed at the participant’s discretion, are used to support description and critical reflection on experiences. First, early in both interviews, participants are invited to complete a timeline38 of their ‘journey’ from awareness of difficulties to first interview, then from initial interview to the second. As timelines are constructed, participants are prompted to identify and explore personally significant events and feelings associated with these events. Second, a set of cards called ‘Ups and Downs’39 is available to support expression of feelings and description of experiences. Depicting a figure experiencing diverse emotions while floating at sea, the cards are used in conjunction with the timeline or separately.

The topic guide and use of timelines were piloted with two adolescents (known to study team members) who had experienced difficulties related to their mental health. When debriefed and invited to critique the interview, they reported finding the interview enjoyable and an opportunity to learn about themselves. No changes to the topic guide or interview process were recommended.

To supplement personal accounts of change in mental health (if any), at each interview, participants are also invited to complete two self-report measures: the Strengths and Difficulties Questionnaire (SDQ)40 and Brief Multidimensional Students’ Life Satisfaction Scale (BMSLSS).41 The SDQ, selected because it is routinely used within CYMHS, measures psychological, behavioural and social difficulties, prosocial function and impact. Completion involves rating the extent to which 25 attributes are ‘true’ for the participant on a three-point scale (‘not’, ‘somewhat’ or ‘certainly’). Designed for 4–16 year olds but validated for use to age 19 years, the SDQ has a reported Cronbach’s α of 0.82 for a score of total difficulties, demonstrating satisfactory internal reliability. With consent, participants’ SDQ scores collected by clinicians at treatment commencement are obtained from CYMHS to measure changes from treatment commencement to timepoint 1. Selected because it is commonly used within services to evaluate change to quality of life, the BMSLSS41 generates a single ‘satisfaction’ score on a continuous scale from 1 to 5 within each of five dimensions (family, friendship, school/work, self and living environment). Scores are summed to calculate overall life satisfaction. The BMSLSS demonstrates satisfactory psychometric qualities with Cronbach’s α=0.76–0.85 in normative samples42 and α=0.77 for a clinical sample of 11–17 year olds.43

Audio recorded with permission, interviews last around 90 minutes. Following each interview, participants are offered a AU$25 gift voucher to acknowledge their contribution.

Data management and analysis

Dependent on type, data are managed and analysed using SPSS V.22 or NVivo V.10.

Following each interview, quantitative data are entered into SPSS.44 Scores are calculated according to instrument guidelines. Within-case scores between treatment commencement and each interview timepoint are compared to identify clinical and reliable change. Results are described (eg. ‘no change in social function’, ‘change from abnormal to normative range in emotional symptoms’) then appended as a notated piece of data to the participant’s interview transcript relating to the relevant timepoint.

Interviews are transcribed verbatim. Along with timelines and notations about which picture cards are used, transcripts are uploaded to NVivo V.1045 for storage and management. Prior to formal coding, each transcript is carefully read to identify chronology of events and important aspects related to them. A chronological summary of each case is made to track each participant’s journey during the analytic process.

Combined data are then coded and analysed. As described by Corbin and Strauss,29 analysis involves deconstructing, conceptualising and creatively reconstructing data in new ways, enabling development of new theories. Analysis in this study involves six iterative stages: (1) within-case analyses at timepoint 1; (2) cross case comparisons at timepoint 1; (3) within-case analyses at timepoint 2; (4) cross case comparisons at timepoint 2; (5) within-case analyses across timepoints incorporating chronology; (6) cross case comparisons across timepoints. Analysis at each stage involves systematic coding and constant comparative methods. Each transcript is segmented into units of meaning which are labelled descriptively. Concepts thus formed are examined in relation to each other, with similar concepts categorised according to more abstract themes and labelled to reflect their similarity. Subsequent interviews are coded using existing and newly generated codes as applicable. Comparisons between concepts and categories are made within and between cases, codes either added to existing categories or new categories developed. As interviews progress, dimensions and properties of each category are established to form distinct, richly described ‘saturated’ categories. Theory building then commences, moving from inductive interpretation of data to reasoned analysis of how the categories relate.

Theory development

Strauss and Corbin’s storyline approach will be used to construct and convey ‘a descriptive narrative about the central phenomenon’.46 ^(p116) The first step in theory building involves linking progressively developing categories chronologically and conceptually to form a storyline. Supporting theory formation, propositional statements are made regarding how categories relate, within which contexts, under what conditions and with what results (in context C, process A occurs resulting in experience B). Saturated categories are further examined to develop an understanding of the most salient issues and to identify a ‘core’ category or central storyline to which all other categories relate. Finally, the theory is situated contextually and boundaries are established. Theory development is regarded as complete when the conceptual framework forms a systematic theory that reasonably represents, in a manner appropriate to use, the phenomenon of interest.

Rigour in analysis

Because qualitative research is inherently subjective,47 the researchers’ wise judgement and diligence are central to the integrity of the product.48 Credibility relies on presentation of analysis such that readers are persuaded of ‘the plausibility of interpretations’.49 ^(p204) Rigour in GT involves ensuring adequacy and trustworthiness of the theory. In this study, rigour will be promoted in three interlinked ways: audit trail, critical dialogue among authors and checking understanding with participants as the theory develops. First, to enable others to follow the process of analysis, procedural and analytical decisions are documented, including theoretical sampling decisions and progression, and challenges to and resolution of hypothetical formulations.50 Second, the analytical process is supported by dialogue between co-researchers who meet monthly during data collection and analysis phases of the research. Co-researchers act as critical friends, challenging the researcher to justify coding and analysis with reference to data, and test alternative explanations.50 Corbin and Strauss51 suggest that ‘opening up’ analyses for peer scrutiny helps guard against bias by promoting reflexivity. Third, to ensure that the theory remains grounded in data, the topic guide is refined to support exploration of developing categories, and tentative understanding of processes are checked with participants. Disconfirming data and exceptions are sought. Finally, ‘reflexivity’, self-awareness and metacognition regarding the researcher’s influence on the research process and its outcomes52 are promoted by keeping a record of the researcher’s thoughts, reactions and feelings during data collection and analysis. This self-reflection assists in identifying assumptions and biases, and it provides opportunities to revise research questions and maintain openness to other possible interpretations of the data.

Quality of the theory

There is consensus among proponents of GT that quality of the theory is first and foremost dependent on it being grounded in the data.24 26 37 Dependent on approach and purpose, various other properties have been recommended as characterising a good theory. In this study, quality of the theory will be assessed on the basis of its fit, relevance, workability and usefulness.24 26 ‘Fit’ is evidenced when categories and concepts relate to the data. Assessment of fit will be enabled by presenting data excerpts to illustrate concepts. Relevance relates to resonance of the theory26 - the extent to which it ‘rings true’ to participants or others sharing their circumstances, eliciting an emotional response.53 A workable theory explains and interprets what is happening in the process under investigation, accounting for any variation.24 To enable assessment of workability, the storyline will be presented in diagrammatic and narrative form. Explanation of the storyline will include description of contexts and conditions within which events, actions or interactions occur. It will also describe the manner in which they vary, under which circumstances and with what consequences. Pragmatically, the ‘usefulness’ of the theory should be assessed in relation to its applicability, its suitability as a basis for making recommendations within everyday settings and for generating discussion and further research.26 Assessment of the theory’s usefulness will be aided by outlining study limitations and recommendations for future research. Recommendations for policy and practice will also be made.

Ethical considerations associated with this research

Researchers are obliged to balance participants’ rights against risks in research participation. Children and adolescents have the right to express their views on matters of relevance to them.54 One way of doing this is through research participation. To promote their right to consider participation in this study, (1) young people have direct access to age-appropriate research information via displays in service reception areas, (2) the researcher engages proactively with clinicians to identify potential participants and (3) clinicians are encouraged to consider rights to research participation, thus not making decisions for the young person, for example, regarding their interest in the study.

Established protocols describe means of minimising risks in research participation for children and people experiencing mental disorders.55 To maintain rights to privacy, potential participants are first approached by clinicians involved with their care. While clinicians are encouraged to invite all eligible service users to consider meeting the researcher, to minimise risk, clinicians consider possible impacts of research involvement and assess potential participants’ emotional capacity for participation before formally referring. On referral, research purposes and procedures are explained in age-appropriate language. The voluntary nature of the study and participants’ right to withdraw without consequence are emphasised. Confidentiality and its limitations are discussed with particular focus on duty of care and management of privacy (secure storage of information) and participants’ anonymity in presentation of results. Before inviting participation, potential participants are asked to describe in their own words what they are being asked to do and for what purpose. Participants provide consent in their own right, and co-consent is sought from parents/carers or another responsible adult familiar with the young person. During data collection, various strategies are employed to minimise power differentials. The researcher maintains focus on participants as experts in themselves and uses developmentally appropriate techniques during interviews, including use of participants’ expressions relating to their experiences. Participants may, at their discretion, invite a support person to accompany them during interview, on the understanding that the researcher will not ask questions of the support person.

It is possible that participants experience emotional discomfort while sharing personal stories. Should participants appear distressed, they are encouraged to take breaks or terminate the interview. If participants disclose risk of harm to self or others, duty of care in the best interest of the young person takes precedence in line with the relevant professional code of ethics,56 and issues of risk are referred, in consultation with the young person, to their clinician or carer. Participants expressing concerns about their mental well-being are referred to their treating team. A document detailing local, online and telephone mental health support options is also provided.

Dissemination of results

The research will be reported as a PhD thesis and in various other formats, disseminated online through the University library, via workshops, conference presentations and published articles, summary reports to participating research sites and within clinical peer/student supervision forums. Participants will be offered a summary of research findings or links to publications. Research and Consumer/Carer Advocacy groups within participating districts will be offered a copy of relevant publications.

When reporting study processes and findings, the 32-item consolidated criteria for reporting qualitative research57 checklist will be used to address three domains relating to research team and reflexivity, study design, and data analysis and reporting.