Experiences of youth with medical complexity and their families during the transition to adulthood: a meta-ethnography

Search strategy

The search strategy was developed in consultation with a health sciences librarian. The first reviewer (LL) comprehensively searched the following databases from inception to February 2019: Medline, CINAHL, Embase, PsycINFO, Social Sciences Index, and Sociological Abstracts. These databases were selected because articles on YMC are most likely to be published in health and social science journals due to the heavy involvement of YMC with health and social services. The search strategy was informed by Cohen et al.’s [1] definition of medical complexity, which is based on needs and functional status and not on diagnostic criteria. Search terms included keywords related to youth, transition or transfer, and medical complexity. Because we used a needs-based definition of medical complexity, we included key words known to be associated with this population (e.g. complex, medically fragile), as well as terms describing medical technologies commonly used within this group (e.g. gastrostomy, tracheostomy). See Supplementary Material 2 for a sample search strategy. After the initial search, we decided to exclude records published before 2002, to coincide with a key consensus statement that led to an increased focus on the transition of youth to adult health services [21]. The search strategy also included hand searching of key journals and the GotTransition^® website [22], as well as reference lists of included articles and literature reviews identified through the search.

Study selection

Three reviewers (LL, MB, PS) performed pilot screening of 20 records, after which inclusion and exclusion criteria were refined. Due to the inconsistency in terms used to define this population in the literature, we decided to include youth populations who fit either Cohen et al.’s [1] previously described definition of medical complexity, or Major et al.’s [23] definition of complex care needs. Major et al. [23] describe care needs as complex based on the breadth (i.e. multiplicity and diversity) and depth (i.e. intensity or frequency) of needs. Two reviewers (LL and MB) conducted title and abstract screening using the following inclusion criteria: (1) study focuses on the transition to adulthood or transfer from children’s to adult services, (2) youth have a chronic condition that fits either Cohen et al.’s [1] definition of medical complexity, or Major et al.’s [23] definition of complex care needs, (3) youth or family perspectives are included, and (4) work is published in the English language. Next, LL screened full-text manuscripts that were included after title and abstract screening, with two additional inclusion criteria: (1) primary qualitative or mixed methods study with direct quotes from youth/family participants and (2) at least 50% of youth have complex care needs. A significant challenge of doing a comprehensive literature search to identify studies that include YMC is that “medical complexity” cannot be reliably identified based on diagnoses, yet research studies are often condition-specific. Therefore, studies may include both youth with and without medical complexity. To mitigate the possibility of underrepresenting the voices of YMC and their families, we decided to only include studies in which 50% or more youth have complex care needs. See Table 2 for the final inclusion/exclusion criteria.

Reading and relating the studies

Study characteristics and key contextual data were extracted by LL using a modified version of the Joanna Briggs Institute (JBI) Qualitative Data Extraction Tool [24]. The full texts of the abstract, findings, discussion, and conclusion sections of each article, along with relevant tables and figures, were extracted into Dedoose for coding of conceptual data. To understand how the studies related to each other, three reviewers (LL, MB, PS) compared each study’s aims, context, participant characteristics, key concepts, and overarching explanations. Studies whose aims did not align closely with the aim of the meta-ethnography or were too narrow in context were excluded at this stage. Consensus was reached by these reviewers about how the studies were related and which studies to include in the synthesis. All studies were then read again by LL to determine whether they were reciprocal accounts (studies that described similar interpretations) or refutational accounts (studies that reported contradicting or disconfirming interpretations).

All studies that were included for synthesis were independently critically appraised by two reviewers (LL/MB or LL/PS) using the JBI Checklist for Qualitative Research [25] and discrepancies were resolved through consensus. While quality appraisal is common in meta-ethnography, there is a lack of consensus about how it should be used [26]. We chose not to exclude studies based on quality, but instead, to reflect on which studies contributed more or less to our interpretations. We assumed that studies with more conceptual depth would naturally have greater influence on the final synthesis output. To test this assumption, we checked whether our findings would still be supported if we had only included studies that were high quality or contained substantial conceptual depth.

Translating and synthesizing the studies

Translation involves comparing and reducing the studies to key concepts, while preserving the context and relationships of each particular study [16]. Two reviewers (LL/MB or LL/PS) independently performed open coding of the abstract, findings, discussion, and conclusion sections of each article, as well as any relevant tables and figures. The data were interpreted in terms of first- and second-order constructs (see Table 1 for definitions). For reciprocal accounts, we translated concepts from one study into another, and then translated the next study into the synthesis of the previous two studies, and so on. We chose to begin by translating conceptually rich, methodologically strong studies, and translated more descriptive and methodologically weak studies later. To preserve the conceptual meanings in each study, each concept was given a working definition using language from that study. As we identified similar concepts across studies, the concept descriptions were updated to reflect all the studies to which they applied, thereby creating a “reciprocal translation.” We also looked for possible refutational accounts and after examining the contextual factors of these studies to understand why differences occurred, we determined that there were no truly refutational accounts (see Findings). However, had true refutational accounts been identified, we would have conducted a refutational analysis as an additional stage in the meta-ethnography.

The analysis team (LL, MB, PS) met eight times to discuss ongoing translation and synthesis. An audit trail of all decisions was kept and reviewers created analytic memos that were discussed during these meetings. To ensure that we were not simply “re-categorizing” the findings, we considered a wide range of interpretations, including both obvious relationships between concepts, as well as more hidden ones. We did this by discussing and questioning each reviewer’s evolving interpretations of key concepts and relationships. The final list of first- and second-order constructs was synthesized into third-order constructs by examining how the second-order constructs were related. Each third-order construct reflects the relationship between two or more second-order constructs. Next, each member of the analysis team independently diagrammed the relationships between third-order constructs, and subsequently met to discuss, merge, and reach consensus on a final overarching model, or line of argument. Lastly, to ensure that the context and relationships were preserved within and across studies, LL compared the main concepts and overarching explanation of each primary study to the third-order constructs and line of argument developed by the analysis team.

Methodological quality and study characteristics

The 10 included studies employed the following designs: descriptive qualitative design [28], [29], [30], [31], interpretive qualitative design [32], case study [33], ethnography [34], grounded theory [35], phenomenology [36], and realist evaluation using mixed methods [37]. Overall, most of the included studies demonstrated congruity between research methodology, objectives, methods, and the interpretation of findings. The most common methodological limitations included: not including a statement that locates the researcher(s) culturally or theoretically (n = 9), not addressing the influence of the researcher(s) on the research (n = 7), and not stating the philosophical perspective or theoretical underpinnings of the study (n = 7). It is possible that these limitations reflect study authors’ or publishers’ decisions to omit this information and do not necessarily mean that reflexivity and philosophical perspectives were not considered. Another important limitation of this body of literature was the descriptive nature of four of the included studies, as meta-ethnography relies on rich conceptual data. However, we confirmed our initial assumption that studies with more interpretive depth had greater influence on the synthesis. Four studies [32], [34], [35], [37] were identified as both conceptually rich and methodologically strong; as such, we labeled these as “key studies” in the synthesis. Details about the quality assessment of each included study can be found in Supplementary Material 4.

The included studies were conducted in the following countries: Canada [28], [31], [32], [33], [36], the United States [34], [35], the United Kingdom [29], [30], and Ireland [37]. Study participants included youth and their parents/family caregivers. There were no studies that included siblings or other family members. Among studies that reported youth and parent gender, 57% of youth were male and 43% were female [28], [29], [30], [33], [36], while 22% of parents were male and 78% were female [28], [30], [31], [32], [36]. Study characteristics can be found in Table 3.

Relating the studies

Because there were a small number of included studies and many similar concepts, it was feasible and practical to create reciprocal translations across all studies. We did not identify any refutational accounts because differences in youth and family experiences were predictably explained by contextual variation. For example, while most studies reported negative youth and family experiences related to coordination, information sharing, and planning [28], [29], [30], [32], [33], [35], [37], participants in three studies also reported positive experiences as a result of proactive planning and coordination by providers [28], [33], [37]. These positive and negative experiences explain two sides of the same phenomenon, and are therefore complementary rather than refutational.

First-order constructs

We identified five first-order constructs (Table 4) representing the experiences of YMC and their families that were each supported by “key studies.” Nine studies described system- and service-related experiences resulting from the termination of children’s services and the need to navigate the new adult system [28], [29], [30], [31], [32], [33], [34], [35], [37]. Seven studies described relational experiences of youth and their families with their providers [28], [29], [30], [32], [34], [35], [37]. Nine studies reported emotional experiences related to the uncertainty and immense burden associated with transition [28], [29], [30], [31], [32], [34], [35], [36], [37]. Nine studies reported experiences related to adulthood, resulting from changes in developmental maturity or shifts in societal expectations [28], [29], [30], [31], [32], [33], [34], [36], [37]. Lastly, eight studies described health- and disability-related experiences in which the youth’s disability, health status, or medical complexity were perceived to influence transition [28], [29], [30], [32], [33], [34], [36], [37].

Second- and third-order constructs

We identified 16 second-order constructs representing study authors’ interpretations, which were synthesized into six third-order constructs (Table 5). The next section describes each of these third order constructs, with examples of supporting second-order constructs. All constructs were backed by “key studies.” A list of all studies contributing to each first- and second-order construct can be found in Supplementary Material 5.

Line of argument

A line of argument (Figure 2) explains the relationships between the third-order constructs. YMC and their families experience the transition to adulthood and transfer to adult services as interactions between changing relationships, goals and expectations, and actions to facilitate transition. Youth and families “make sense” of transition through emotional processing of these interactions. For example, ending relationships with providers creates a sense of uncertainty, which then compels families to take action, establish new relationships, and modify their expectations. Being respected and valued contributes to youth and family empowerment and positive transition experiences. Youth and their families feel supported when their goals, relationships, and actions are concordant and create a positive feedback loop that allows transition to progress. On the other hand, when these components are discordant, youth and their families experience uncertainty and inner turmoil. The transition experience is bounded by the nature and process of transition, which is both multidimensional and determined, in part, by system-mandated critical points in the process. Lastly, transition is influenced by contextual factors at the youth, family, social, and system levels.

Figure 2:

Making sense of transition: a line of argument conceptual model.

Approximate placement: after line of argument/before discussion.

Emotional needs

Transition is emotionally burdensome on youth and their families due to feelings of uncertainty and abandonment (construct 5) that result from the abrupt termination of services and relationships with providers (construct 2) [33], [35]. In an integrative review on the transition experiences of YMC, Joly [15] reports similar findings of youth and their families experiencing loss, abandonment, and uncertainty. Furthermore, while youth and families demonstrated resilience, perseverance, and skills in navigation and advocacy, parents also described increasing difficulty in managing ongoing care needs coupled with less energy as both the youth and parent got older [31], [32], [36]. To address the strong emotional component of transition, health care providers should assess and support the emotional needs of youth and their families, and begin planning for transfer well in advance to avoid eliciting feelings of sudden abandonment.

Transition is a journey

Schultz [35] p. 361] describes transition as being “sparked by a crisis”, and this finding is reflected across the included studies, which report age or funding as the primary agent for change or loss of services (construct 1). Joly [15] p. e93] compares the transition process to “being pushed off a cliff, falling into the abyss, and landing at the bottom”. This sense of crisis is likely due to health care providers’ focus on the one-off event of transfer, rather than a more gradual and holistic transition process. Health care providers should take the time to discuss with youth and their families how they envision their future (construct 3), instead of focusing solely on the transfer of services. Furthermore, we argue that transition should be conceptualized as a journey, rather than an event or outcome. Stewart et al. [38] describe this journey in three phases: the “preparation”, the “journey”, and the “landings” in the adult world. As a journey, transition should be defined by the youth and family’s preparation, readiness and adaptation to change, and eventual reintegration into a new phase of life.

The need for collaboration

While the experiences of youth and their families were mostly negative, there were reports of both positive and negative experiences related to transition planning in three of the included studies [28], [33], [37]. These positive experiences were associated with proactive planning by families and their providers, as well as availability and access to information and services, while negative experiences were associated with the opposite. These contrasting experiences illustrate the powerful impact that providers can make through actions such as communicating, collaborating, and providing information (construct 4). This highlights the importance of collaboration among stakeholders from both children’s and adult systems to prepare and support YMC and their families before, during, and after transition. This need for collaboration applies not only to YMC and their families, but to other youth with special health care needs as well [8].

Collaboration should extend beyond the domain of health services, as transition involves multiple sectors (construct 1). In a review of North American and European systems of care, Kraus de Camargo [39] describes the tendency of health and education services and policies to focus on their own issues, while social services focus more on the lives of citizens. Interestingly, studies in this meta-ethnography that examined transitions in multiple sectors [29], [30], [31], [33], [34], [37] reported similar findings to studies that examined only health services [28], [32], [35]. For example, issues that were common to health, education, and social services included: inadequate access to funding and services (e.g. adult health services, tuition support, options for independent living) [32], [33]; poor communication and collaboration [28], [37]; and lack of expertise among adult providers about complex pediatric conditions or programs and grants to support post-secondary students with “intensive” care needs [32], [33]. These similarities suggest that the three systems face common issues and could benefit from an integrated approach to transition. Other authors have made similar calls for active collaboration from all involved stakeholders [8], [40]. A multi-sectoral transition strategy should focus on building capacity among adult providers to meet the needs of YMC and their families entering the adult system, and supporting the social inclusion of YMC and their families [8].

What about youth development?

Lastly, the findings highlight that the bureaucratic processes related to health care transition (construct 1) are largely inflexible to youth development. In other words, the transfer from children’s to adult services typically happens at a predetermined chronological age, rather than based on youth developmental stage and needs. The included studies overwhelmingly describe youth and family perceptions of limitations of the adult health system. This is likely due to the predominant focus of adult health services on the sizable population of older adults, rather than on young adults with chronic conditions. For example, Cook [33] p. 6] describes how the “maintenance style” culture of extended care facilities limited young adults in achieving their developmental goals for independence. In Canada, pediatric health services often stop when the youth reaches a predetermined age, such as the age of majority. However, there are also many exceptions to this rule, further contributing to the messiness of transition. Furthermore, YMC may not be developmentally ready to transition to adulthood and adult services at this age, or may require a prolonged transition period. By prematurely placing these vulnerable youths into an adult system that is not designed or ready for them, they risk having their voices drowned out due to their relatively small numbers. Recognizing the need for a prolonged transition period, the American Academy of Pediatrics has recommended starting the transition process as early as age 12, and completing the process as late as age 26 [41]. As such, policy makers should consider flexible policies for age cutoffs related to pediatric health services and funding. This recommendation is not meant to simply delay the same issues; rather, it is meant to allow for additional time to plan for transition and offer increased flexibility for families.

Implications for practice, policy, and research

The conceptual model that we have proposed can be used as a framework to guide practice, policy, and research that aims to address the multidimensional and dynamic nature of transition. Service providers may want to focus on the inner constructs of the model (e.g. goals, relationships, actions), as they have limited control over the outer constructs (e.g. system factors and the nature and process of transition). On the other hand, policy-makers have more sway over these outer constructs, and less influence over individual interactions. Regardless of the area of focus, all users should consider the model in its entirety. For example, a health care provider may want to take actions to proactively address the changing relationships brought on by transition, but will need to consider these actions in relation to the youth and family’s goals and expectations, as well as the nature, process, and context of transition. Researchers can also use this model to examine relationships between specific constructs.

Based on our synthesis of youth and family experiences with transition, we suggest that providers: (1) begin planning for transfer early and link youth to adult health providers well before transfer; (2) ensure access to accurate and consistent information about services, supports and resources; and (3) communicate and collaborate with all involved stakeholders across agencies, sectors, and children’s and adult systems. Providers should also consider: (4) offering family-centred support services to address youth and family emotional needs and (5) discussing with youth and their families how they can be supported in realizing their visions for the future. Many of these suggestions are already embedded in published guidelines and position papers on health care transition [41], [42], and our findings echo previous reports of inconsistent implementation of evidence into practice [8]. However, there does appear to be some evidence of increased uptake over time. For example, while earlier studies mostly described poor youth and family experiences with service transfer [29], [30], [32], [35], more recent studies reported a mix of positive and negative experiences [28], [37], suggesting that providers are aware of current recommendations but are not consistently applying them. Transition planning is time-consuming and it is likely that the lack of time, resources, and incentives prevents providers from engaging in transition planning activities. Therefore, at the system level, policy makers should consider developing or improving existing policies to optimally support providers in implementing recommended practices, such as allocating dedicated resources for transition, rather than adding on to current provider workloads [37]. Policies should also focus on using incentives to drive system change, such as offering financial incentives or holding organizations accountable to quality standards for transition planning and continuity of care. Lastly, policy makers should consider adopting flexible age cutoffs for children’s services and funding – tailored to individual youth development and family needs – as well as supporting multi-sectoral strategies aimed at the social inclusion of YMC and their families and building capacity for adult providers to meet the needs of this population.

Reflexivity

The review team consisted of interdisciplinary team members with clinical backgrounds in nursing (LL, MB, NC, JP, PS) and medicine (JWG). Three reviewers (LL, MB, JWG) have clinical and research expertise in working with CYMC and two reviewers (LL, JWG) have research interests in the transition to adulthood. Additionally, three reviewers (NC, JP, PS) are senior qualitative researchers. In particular, the analysis team (LL, MB, PS) has combined expertise in qualitative data analysis, CYMC, and transition to adulthood. The various backgrounds, disciplines, and perspectives brought by each team member influenced the interpretations, implications, and conclusions developed in this meta-ethnography.

Strengths

We consulted a health sciences librarian in developing the search strategy, which contributed to the methodological strength of this synthesis. Two reviewers (LL, MB) who had extensive nursing experience working with CYMC screened all records. This was another strength, as we were able to draw on clinical judgment to identify youth who fit the inclusion criteria. Furthermore, because the number of included studies was relatively small and manageable, we were able to become highly familiar with each study’s concepts, context, and conclusions, which allowed us to stay grounded in the data. Lastly, we followed a systematic approach and established reporting guidance in conducting and reporting this meta-ethnography.

Limitations

A limitation of this meta-ethnography is the inconsistency in how medical complexity has been defined in the literature. As a result, comprehensively searching for research related to this population proved to be challenging, despite having the guidance of a health sciences librarian. Additionally, our search did not include grey literature, which may have limited the identification of studies or dissertations relevant to the aim of this meta-ethnography. Furthermore, although we purposefully drew on clinical expertise to decrease the likelihood of incorrectly excluding studies, there remains the possibility that relevant studies on transition for youth with specific chronic health conditions were missed in either the search or screening phases.

Another limitation was the mixing of data types and perspectives that were external to the aim of our meta-ethnography. We wished to focus only on the interpretive accounts of youth and families; however, one study included mixed methods data [37] and three studies included service providers as participants (in addition to youth and families) [30], [34], [37]. To address these issues, we only included direct quotes from youth/family when analyzing first-order constructs, and we checked that second-order constructs from these studies were supported by first-order constructs. However, we recognize that the study authors’ interpretations (second-order constructs), and therefore our interpretations (third-order constructs), would inevitably be influenced by all data within a study. While the mixing of data types and perspectives posed a challenge to our analysis, we found that this “extra” information provided context and deeper insight into the experiences of YMC and their families.

Another limitation of this meta-ethnography is the descriptive nature of four of the studies, which may have limited the interpretive depth of the findings. For example, the construct of contextual factors is used to categorize, rather than explain the essence of the synthesized concepts. This was due to the majority of included studies describing factors as having a static relationship with transition (i.e. barriers and facilitators). Lastly, we propose that cultural factors may have a major influence on transition experiences, as adulthood is a social construct with a strong cultural component. However, only one study reported culture-based experiences with transition [34]. Therefore, this may be an important component that is currently missing from the model.