Measures of identity in adolescents/young adults with long-term physical health conditions: a systematic review.

The study protocol is available through the Open Science Framework, and there were no deviations from the protocol: https://osf.io/y856n . This review is reported using recommendations from Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA; Page et al., 2021 ), and the review of measures follows an adapted version of COSMIN ( Mokkink et al., 2018 ). This review specifically focused on measures that examined personal identity, defined as an individual’s sense of self, rather than other forms of identity such as social, gender, ethnic, or professional identity. To be included for review, papers were required to have been empirical and to have adopted quantitative or mixed methods. Papers were required to have been (1) published in peer-reviewed journals, (2) written in the English language, (3) include at least one measure of identity formation, and (4) include participants aged 16–24 years with a diagnosed or self-diagnosed LTC-P. Papers were included if they had participants in this age range, regardless of whether or not the results were reported separately for the 16–24-year-old age group. Papers were excluded from the review if they did not collect primary data, included populations in which the primary focus was on (1) mental health conditions (e.g., depression, anxiety, obsessive compulsive disorder, etc.), (2) neurodevelopmental conditions (e.g., autism, attention-deficit hyperactivity disorder, Tourette syndrome), or (3) cognitive impairment (e.g., Down’s syndrome, traumatic brain injury). Searches were conducted on October 24, 2024, in the following databases: APA PsychNET, PubMed, Web of Science, SCOPUS, and CINAHL. No restrictions were put on the search strategy in terms of date. Additionally, forward and backward citation searches for all eligible papers were conducted using Google Scholar. With guidance from a specialist subject librarian, a search strategy ( Table 1 ) was devised that aimed to capture three main concepts: (a) personal identity, (b) LTC-P, and (c) measures of personal identity development. To fully capture the recent expansion of the age boundary of adolescence ( Sawyer et al., 2018 ), no age-related concept was included in the search terms. Restricting the search to words associated with AYA may have excluded papers that conceptualize this group as 18 years and below. Instead, this inclusion criterion was screened in either the Title/Abstract stage, or the full-text review. The main concepts of personal identity, LTC-P, and measures of personal identity development were limited to searching in the Title/Abstract fields only. The full-search strategy for each database can be found in the online supplementary material 1 . Search terms. Identity: Identity OR “Identity Formation” OR “Identity Development” OR “Identity Exploration” OR “Illness Identity” AND Long term conditions—physical: “Long Term Condition” OR “Chronic condition” OR LTC OR “Recurrent Condition” OR “Chronic Disease” OR “Chronic Pain” OR “Persistent Pain” OR Diabetes OR Crohn’s OR IBD or “Ulcerative Colitis” OR “Inflammatory Bowel Disease” OR CRPS OR “Complex Regional Pain Syndrome” OR Fibromyalgia OR Arthritis OR Endometriosis OR Migraine OR “Chronic daily headache” OR Osteoarthritis OR “Spina Bifida” OR Asthma OR Allergy OR “Atrial fibrillation” OR “Chronic Constipation” OR “Chronic Fatigue Syndrome” OR CFS OR ME OR “Myalgic encephalomyelitis” OR “Chronic Kidney Disease” OR “Congestive Heart Failure” OR Epilepsy OR Hypertension OR HIV OR “Irritable Bowel Disease” OR “Multiple Sclerosis” OR Anaemia OR Angiooedema OR “Sickle Cell” OR “Autoimmune Disorder*” OR “Autoimmune Disease” OR Lupus OR “Sjögrens syndrome” OR “Bronchopulmonary dysplasia” OR Cancer OR “Cardiac arrhythmias” OR “Coeliac Disease” OR “Coronary heart disease” OR “Cystic Fibrosis” OR “Endocrine disorder*” OR “Widespread pain” OR Gout OR “Heart Failure” OR “Lung Fibrosis” OR “Ehlers Danlos syndrome” OR haemophilia OR Rheumatoid OR Neurofibromatosis OR “Charcot-Marie-Tooth disease” OR Scoliosis OR scleroderma OR vasculitis OR “juvenile idiopathic arthritis” OR “rheumatoid arthritis” OR “congenital heart disease” OR thalassemia OR “chronic liver disease” OR cancer OR dermatitis OR eczema OR psoriasis AND Papers retrieved from searches were uploaded into Covidence (2024) , an online systematic review tool, and duplicate papers were removed. In the first instance, all papers were screened by the first author (T.R.), with a second reviewer (H.J.) independently screening at least 50%. Any disagreements were resolved by a third reviewer (A.J.). Any papers for which eligibility was not clear from the title/abstract alone were included at this stage. After this initial screening process, included papers were assessed for eligibility via retrieving and reviewing the full text. Full-text papers were independently reviewed by two reviewers (drawn from T.R., A.J., and H.J.), with disagreements again being resolved by a third reviewer (L.C.). Data extraction was then performed independently by two reviewers of the study team (drawn from T.R., L.C., and F.T.), using a piloted data extraction form that was hosted within Covidence (2024) . Data extracted included information regarding: the paper (date, authors, journal name, country of origin), study information (definition of identity used, participant age range, gender identity, sex, sexual orientation, ethnicity, race, socioeconomic status (SES), religion, LTC-P, study design, recruitment site), measure information (name, purpose, target population, translations, manual included, theories/concepts captured), and any COSMIN psychometric evaluation outcome properties available (PROM development, content validity, structural validity, internal consistency, cross-cultural validity/measurement invariance, reliability, measurement error, criterion validity, hypotheses testing for construct validity, and responsiveness). When data were missing or unclear, the first author of the paper was contacted, with subsequent authors approached if no response was obtained after two attempts. A two-step approach to assess the quality of each identity measure was conducted. First, all measures were rated using the Cohen et al. (2008) criteria for evaluating evidence-based assessment, identifying measures as either “well-established,” “approaching well-established,” or “promising assessment.” To achieve a rating of well-established, measures must have been presented in at least two peer-reviewed papers by different researchers or research teams, with sufficient detail available about the measure to allow for critical evaluation or replication (e.g., a manual provided or available upon request) and detailed statistics indicating good validity and reliability in at least one peer-reviewed paper. To achieve a rating of approaching well-established assessment, measures must have been presented in at least two peer-reviewed paper by different or the same researcher or research team, with sufficient detail available about the measure to allow for critical evaluation or replication (e.g., a manual provided or available upon request) with validity and reliability information presented in either vague terms (e.g., no statistics presented) or moderate values. Finally, measures rated as promising assessment were presented in at least one peer-reviewed article, with sufficient detail to allow critical evaluation and replication (e.g., a manual provided or available upon request), and reliability information presented in either vague terms (e.g., no statistics presented) or moderate values. In Step 2, measures rated “well-established” and “approaching well-established” using the Cohen criteria were assessed for quality using an adapted COSMIN checklist version 1.0 ( Mokkink et al., 2018 ). Data were independently extracted from each paper containing each measure by three reviewers (T.R., F.T., and L.C.). These data were then rated independently by two reviewers (T.R. and L.C.) on structural validity, internal consistency, cross-cultural validity, measurement invariance, reliability, construct validity, and criterion validity using the COSMIN Risk of Bias checklist, v1.0. In the first step, the quality of the methodological approach of each paper in assessing a particular COSMIN measurement property was rated using a four-point rating system, ranging from “very good,” “adequate,” “doubtful,” or “inadequate.” The overall quality rating of the paper was then determined by taking the lowest rating of any COSMIN measurement property the paper addressed, with this rating used to grade the quality of the evidence for each measure (e.g., a measure with “inadequate” evidence will decrease trust in the overall conclusions of the measurement property). If information regarding a particular measurement property in the checklist was missing, then the rating was left blank. In the second step, each paper’s evidence for a given measurement property was rated as sufficient (+), insufficient (–), or indeterminate (?) (see online supplementary material 2 for criteria). Any disagreements between the two reviewers were discussed, and a consensus was made. Next, to determine an overall rating for each measurement property of a measure, we summarized the ratings from all relevant papers. Following COSMIN guidance, a summarized rating was considered sufficient, insufficient, or indeterminate if at least 75% of the individual study ratings met that criterion. In the case of inconsistent ratings, overall ratings were based on the majority rating. In the case of an equal split between ratings, the decision was made on the ratings of the most recent publications assessing that measurement property of the measure. In the final step, a modified Grading of Recommendations Assessment, Development, and Evaluation (GRADE; Mokkink et al., 2018 ) approach was used to downgrade evidence if there were any concerns about the quality of evidence presented, with evidence starting at high quality, and subsequently downgraded by one or two levels per factor. The factors comprised risk of bias (as assessed in the first step described above), inconsistency (of results across studies), indirectness (evidence coming from different populations than those of interest in current review), and imprecision (total sample size of studies included in review). While the COSMIN checklist was used to assess other measurement properties, the criteria for evaluating content validity were found to be overly rigid for the purposes of this review, which has been noted previously ( Rothmund et al., 2023 ; Smith et al., 2021 ). Applying COSMIN standards would have led to “inadequate” ratings for many measures, which we believe would not accurately reflect the efforts undertaken by the paper authors and lead to the dismissal of potentially valuable measures being used in the future. Instead, we chose to summarize how content validity was addressed in each paper in a narrative review, looking at item generation, content validity, and face validity, allowing readers to make their own informed judgments on potential use in future research of measures. This was performed by one reviewer (T.R.). In the final step, each measure was categorized using an adapted COSMIN guideline: Category A (at least low-quality evidence for “sufficient” internal consistency), Category B (not meeting criteria for A or C), or Category C (high-quality evidence for an “insufficient” property). Category A measures are considered suitable for recommendation, Category B require further research, and Category C are not recommended. As a full COSMIN content validity assessment was not performed, Categories A and B should be interpreted with caution, and further research on content validity, including concept elicitation and stakeholder involvement, is advised. A meta-analysis was not undertaken, as the aim of this review was to evaluate the development and psychometric quality of identity measures, rather than to synthesize outcome effects.

The electronic searches returned 15,926 records, with 8,091 duplicates removed. Of the 7,835 papers that remained, 7,751 papers did not meet the inclusion criteria after title/abstract screening, leaving 84 for full-text eligibility assessment. After full-text screening, 47 papers were excluded. The main reasons for exclusion included: (1) papers did not contain a measure of identity development ( n  = 19), (2) the age range of participants was not reported ( n  = 10), or (3) papers did not contain participants aged 16–24 years ( n  = 9). Figure 1 contains the study PRISMA flow chart. After screening, 37 papers measured identity and were included in the review. Within these papers, 16 unique measures were identified. PRISMA flow chart. PRISMA = Preferred Reporting Items for Systematic Reviews and Meta-Analysis. Cohen’s Kappa for the title/abstract screening was 0.51, signifying moderate agreement between reviewers. Most discrepancies arose from papers employing the Illness Perception Questionnaire ( Weinman et al., 1996 ), of which the “Identity” subscale does not correspond to the construct of personal identity specified for this review. After discussion, the team agreed to exclude these papers from further analysis. Cohen’s Kappa for full-text screening was 0.84, indicating strong agreement. Each paper was assigned a unique code, which is presented in Table 2 alongside a summary of the included papers. Papers were published between 1987 and 2024, with the majority after 2014 ( n  = 31). There were 16 LTC-P observed in the papers. Type 1 diabetes was the most investigated LTC-P ( n  = 10), followed by congenital heart disease ( n  = 9), HIV ( n  = 5), and MS ( n  = 5). Overview of studies included for review. MT: 20–67 (36.63, 13.44) a NMTL 23–63 (37.94, 10.64) a Ethnicity: Latino(a) 38.7 Race: Black 55.9 l White 19.4; Other 18.3 NMD Men 69.5; Women 30.5 T1D Men 69.5; Women 30.5 Refractory epilepsy: tertiary referral center in the Netherlands (Epilepsy Centre Kempenhaeghe) CHD: Belgian branch of Assessment of Patterns of Patient-Reported Outcomes in Adults with Congenital Heart disease—International Study (APPROACH-IS) Sex: CHD: Men 54.3 MSDs: Men 17.4% IBD: Crohn’s & Colitis Foundation Belgium CHD: database of congenital cardiology of the University Hospitals Leuven Epilepsy: tertiary referral center in the Netherlands (Epilepsy Centre Kempenhaeghe) MSDs: database of rheumatology of the University Hospitals Leuven Indicates not separated in reporting. Across all papers, 9,486 participants were included with an age range of 13–83 years. However, only five papers provided separate data for participants in the age range 16–24 years, representing 383 participants from the overall sample. Seven papers were longitudinal, with the remaining employing a cross-sectional design. Papers were mainly based in Belgium ( n  = 12) and the United States ( n  = 10). Table 3 presents the definitions of identity presented in the included papers. In reviewing the definitions of identity across the included papers, 12 did not provide an explicit definition of identity ( 01,02,03,05,06,10,12,15,24,29,32,34 ), with one using the term identity interchangeably with self-concept ( 15 ). The remaining papers offered different definitions of identity, as summarized below. Eleven papers explicitly referenced Erikson’s (1968) theory of identity development ( 13,14,19,21,26,28,30,31,36,37 ), highlighting identity development as a process involving crisis, exploration, and commitment. Five papers ( 04,07,08,09,16 ) conceptualized identity as a sense of continuity, stability, and uniqueness across time and life changes, most commonly citing Bosma and Kunnen (2001) . Six papers ( 07,11,21,22,27,34 ) adopted a lifespan approach to explaining identity, and its dependence on social and contextual factors. Eight papers focused specifically on adolescence and emerging adulthood as critical periods for identity exploration across multiple life domains, emphasizing identity as domain-specific and evolving through exploration ( 14,17,18,19,20,21,23,25 ). Additionally, two papers framed identity as multidimensional, encompassing various elements of the self ( 11,35 ). Definitions of identity provided in included papers. a “An important developmental task for late adolescents is personal identity formation. Based on Erikson [1] and Marcia [2], Luyckx and colleagues have proposed an identity model that includes both the formation and the evaluation of identity commitments.” Only included papers that explicitly provided a definition of identity are shown. Table 4 presents an overview of measures included in this review. A wide variety of theoretical concepts and frameworks informed the development of measure items. Five measures (Ego Identity Incomplete Sentence Blank [EI-ISB], Extended Objective Measure of Ego Identity Status [EOM-EIS], Ego Identity Process Questionnaire [EIPQ], EOM-EIS-2, and Objective Measure of Ego Identity Status [OM-EIS]) drew on Ego Identity Status theory ( Erikson, 1968 ; Marcia, 1966 ), while two (Illness Identity Questionnaire [IIQ] and Accepting Diabetes and Personal Treatment [ADAPT]) were informed by Illness Identity Theory ( Charmaz, 1995 ; Oris et al., 2018 ). The centrality of stigmatizing attributes to identity was explored in two measures (Collective Self-Esteem Scale [CSS] and HIV Centrality), based on work by Quinn and Chaudoir (2009) . One measure (DIDS) reflected the five dimensions of identity development presented by Luyckx et al. (2008) , and another (Identity Reconstruction Assessment Scales [IRAS]) drew from qualitative research on identity reconstruction ( Yoshida, 1993 ). Additional measures were informed by extensive review of stigma literature and clinical expertise (HIV Stigma Scale [HSS]), the Motivated Identity Construction Theory (Identity Motives Scale [IMS]; Vignoles et al., 2006 ), and Social Identity Theory (CSS; Tajfel & Turner, 1979 ). Others incorporated broader theoretical explorations of identity and self-concept (HIV-Positive Identity Questionnaire [HPIQ]), identity change across four domains (Identity change: amount and affective change), and mixed-method research on HIV-related identity changes (Impact on Self Concept Scale [ISCS]; Updegraff et al., 2002 ). Characteristics of identity development measures (original development studies). Age: 14–25 years Sample: Type 1 diabetes Age: 17–20 years Sample: Normative Age: 19–51 years Sample: Normative Age: 13–25 years Sample: Type 1 diabetes Age: 21–63 years Sample: HIV 4 (two adapted from the original identity subscale and two based on previous work with people living with concealable stigmatized identities) Age: 18+ years Sample: Normative; Male Age: 18+ years Sample: Normative Age: 18–77 years Sample: HIV Age: 19–82 years Sample: HIV Age: 40 years (mean) Sample: HIV Age: 19–73 years Sample: Chronic illness Amount: Participants are asked to think about who they are since diagnosis; then rate how much they have changed as a person (personal identity), in terms of identifying with social groups (social identity), who they are in their relationships (relational identity), and how they act or communicate about themselves (performative identity). Each item was measured on a scale from 1 ( not at all ) to 7 ( very much ). Affective: Participants asked to assess whether they viewed change in that identity aspect since their diagnosis to be negative, neutral, or positive. Items measured on a 7-point scale: range 1 to 7, where a value of 4 indicated a change that was perceived as neither more negative nor more positive, values 4 a positive change. Age: 20–74 years Sample: Multiple sclerosis Age; 18–54 years Sample: HIV; Male Participants were asked to respond to both positive and negative items by indicating how often the feel similarly to each statement, using a scale ranging from 1 ( never ) to 6 ( always ). Age: 17–24 years Sample: Normative Age: 18–45 years Sample: Normative Age: 17-51 years Sample: Normative The study samples of the original measure development papers varied from adolescents, young adults, and older adults (range 13–82 years). Eight measures were developed for use with a normative population (DIDS, IMS, EI-ISB, EOM-EIS, EIPQ, EOM-EIS-2, and OM-EIS), five were specifically developed for individuals with HIV (ISCS, HPIQ, HSS, HIV Centrality, and CSS), one was developed for use with any chronic illness (Identity change: amount and affective change), one for those diagnosed with Type 1 diabetes (ADAPT), and one for MS (IRAS). Most measures were developed in English, although several have been translated into other languages (e.g., Dutch and German). Across all papers identified for review, only three measures were used in more than one paper: 18 used the IIQ ( Oris et al., 2016 ), 4 used the Dimensions of Identity Development Scale (DIDS; Luyckx et al., 2005 ), and 4 used IMS ( Vignoles et al., 2006 ). The remaining 15 measures were used in a single study each. Using Cohen et al.s’ (2008) criteria for evaluating evidence-based assessment, only three measures were rated as “well-established,” with the remaining measures rated as “promising.” Quality was only assessed for these three, as the authors considered that analysis would only be meaningful where data were available from more than one study. Consequently, only the IIQ ( Oris et al., 2016 ), DIDS ( Luyckx et al., 2005 ), and IMS ( Vignoles et al., 2006 ) were assessed for quality using an adapted COSMIN checklist version 1.0 ( Mokkink et al., 2018 ). Table 5 shows a summary of the quality assessment, and full COSMIN evaluations can be seen in online supplementary material 3 . The content validity of all papers was synthesized using a narrative review and will be represented at the end of this section. Quality assessment using COSMIN by measure. Note . In the COSMIN evaluation, “+” indicates sufficient assessment and “?” indicates indeterminate assessments. Empty cells in the table indicate that none of the included studies provided a relevant analysis for a measurement property. IIQ = Illness Identity Questionnaire; DIDS, Dimensions of Identity Development Scale; IMS, Identity Motives Scale. Quality of evidence ratings: High, moderate, low, very low. The included papers only rated four measurement properties—structural validity, internal consistency, measurement invariance, and construct validity. Additionally, COSMIN requires the review team to create a hypothesis for testing construct validity; however, due to the diverse nature of constructs and comparator instruments being captured in this review, we chose to only assess construct validity with regard to the hypothesis chosen by the authors of the studies. Only the measurement properties reported in the included papers are presented below; properties not shown were not assessed in those papers. The IIQ received a “Very good” risk of bias rating for structural validity, internal consistency, and construct validity. Only one study ( Oris et al., 2018 ) looked at measurement invariance, receiving an “Adequate” score. The DIDS received a “Very good” rating for structural validity, internal consistency, and construct validity. The IMS received a “Very good” rating for construct validity. It received a “Doubtful” rating for internal consistency due to papers reporting six subscales that contributed to a total score, but only providing a single Cronbach’s Alpha, making it unclear whether the subscales could be reliably interpreted on their own. The IIQ received a “Sufficient” rating for structural validity, internal consistency, and construct validity. Only one study ( Oris et al., 2018 ) looked at measurement invariance, receiving an “Indeterminate” score due to no multiple group factor analysis or differential item functioning analysis performed. The DIDS received a “Sufficient” rating for structural validity, internal consistency, and construct validity. The IMS received a “Sufficient” rating for internal consistency and construct validity. The quality of evidence GRADE ratings was mostly influenced by risk of bias and indirectness. Structural validity was of “high” quality evidence across the IIQ and DIDS. The IMS did not report any structural validity assessments. Internal consistency for the IIQ and DIDS was “High,” whereas the IMS scored “Low” due to serious risk of bias (multiple studies of doubtful quality) and serious indirectness. The rationale for the risk of indirectness was that IMS studies included in the review were performed in populations aged 10–57 years, with averages ranging from 25 to 38 years; this was deemed to be indirect as the current review is interested in those aged 16–24 years. The one assessment of measurement invariance for IIQ was “Moderate” quality due to a serious risk of bias (only one study of adequate quality). Quality of evidence for construct validity in the IIQ and DIDS was “High.” The IMS scored “Moderate” on construct validity due to serious indirectness for the reasons noted previously. The content validity for all 16 measures, using the original measure development study or the study in which a measure had been heavily modified, as well as any further content validity performed in the review papers, was assessed by one reviewer (T.R.). One measure, the IMS, could not be assessed because the original referenced development papers ( Manzi et al., 2010 ; Vignoles et al., 2006 ) did not clearly explain how the content aligned with the measure manual reviewed by the assessors and did not offer any validation information. Additional content validity of the IIQ was performed in two papers that reported translations of the measure into Danish and Hebrew ( Ingersgaard et al., 2022 ; Meyer & Lamash, 2021 ). Full details of content validity are presented in Table 6 . Across the reviewed measures, expert involvement was often not reported. In the few instances where involvement was described, experts contributed through activities such as revising items based on clinical experience or providing specialized input during the development process. However, the specific roles, qualifications, and expertise of the experts involved were rarely detailed. As a result, while some engagement was reported, there appeared to be limited systematic engagement with professional expertise throughout the reviewed measures. Content validity processes across measures. Original development study: Not reported Validation of Danish version: Two experts (of Social Science and diabetes) assessed relevance and comprehensiveness regarding cultural appropriateness of IIQ and found items matched English version and theoretical constructs. Validation of Hebrew version: Face validity with five experts in child development and chronic health conditions. Reviewed questionnaire for content and understandability found no major issues. Original development study: Not reported Validation of Danish version: Relevance and intelligibility of the items, responses, and instructions assessed using cognitive interviews with 16–26-year-olds with Type 1 diabetes. Led to linguistic modification of two items. Validation of Hebrew version: Face validity with 5 adolescents and young adults with celiac disease. Reviewed questionnaire for content and understandability found no major issues. Note . IIQ = Illness Identity Questionnaire; DIDS, Dimensions of Identity Development Scale; EI-ISB = Ego Identity Incomplete Sentence Blank; EOM-EIS = Extended Objective Measure of Ego Identity Status; PIQ = Positive Identity Questionnaire; IRAS = Identity Reconstruction Assessment Scales; ISCS = Impact on Self Concept Scale; EIPQ = Ego Identity Process Questionnaire; OM-EIS = Objective Measure of Ego Identity Status. Translation studies included in review. Developed and validated with normative sample. Not peer reviewed. Similarly, lived experience involvement was frequently not reported. Where it was described, involvement typically took the form of feedback sessions with individuals who had relevant lived experiences; one measure specifically noted the use of cognitive debriefing ( Commissariat et al., 2023 ). These contributions were intended to enhance the relevance, clarity, and acceptability of the measure items. Nevertheless, comprehensive reporting on the nature and integration of lived experience feedback was limited, again suggesting systematic and meaningful engagement with lived experience perspectives was not consistently prioritized across the measures. Using an adapted COSMIN categorization, the IIQ and DIDS were placed in Category A (at least low-level quality evidence for sufficient internal consistency) and thus can be tentatively recommended for use in AYA with LTC-P. An important caveat in this recommendation is the need for content validity studies, using expert and lived experience voices, to be undertaken to ensure all measures adequately reflect the construct in the selected study population. While the IMS did demonstrate low-level evidence for sufficient internal consistency as required for Category A, the review team felt the lack of content validity was sufficient to downgrade the level to Category B, requiring further research to assess the quality of the measure for use in AYA with LTC-P.

We identified 37 papers reporting the use of measures of identity with individuals living with LTC-P aged 16–24 years. Across these studies, 16 unique self-report measures were used. Evaluation of the psychometric properties and methodological quality of the measures highlighted both strengths and limitations, offering valuable insights into how identity has been assessed in this population and how it could be measured in future research and clinical practice. When selecting a measure, users should carefully consider the relevance of each instrument to their specific aims. The IIQ may be suitable when examining illness identity, but it may not capture the broader reconstruction and negotiation of identity for AYA following a diagnosis of LTC-P. Similarly, the DIDS assesses an individual’s stage of personal identity development and does not specifically address issues related to living with an LTC-P, such as the reaction to disrupted life trajectories ( Hale et al., 2015 ; Kirk & Hinton, 2019 ) and shifts in social roles ( Harris et al., 2003 ; Packham et al., 2020 ). Neither measure was originally co-produced with experts or young people with lived experience, ensuring content validity. Although later translations of the IIQ incorporated limited content validity testing for cultural and linguistic reasons ( Ingersgaard et al., 2022 ; Meyer & Lamash, 2021 ), future use, especially in unvalidated populations, should include a systematic evaluation of content validity to ensure appropriateness. The remaining 13 measures identified in this review demonstrated promising assessment ( Cohen et al., 2008 ). However, future validation studies are recommended before their widespread use, particularly for measures developed some time ago, and/or within normative samples, as their content may not be fully applicable to contemporary or specific populations. A consistent limitation of measures included in this review was the lack of systematic assessment of content validity, particularly for measures originally developed for use in normative populations (e.g., “healthy” adolescents). Most measures were developed by applying theoretical frameworks (often developed for adult or normative populations) or by adapting items from existing measures, with only limited involvement from experts or individuals with lived experience. This gap in content validity risks measurement error and could limit clinical usefulness in these populations. Notably, the ADAPT measure ( Commissariat et al., 2023 ), developed for individuals with Type 1 diabetes, was the only tool to include cognitive debriefing and represented the most comprehensive approach to content validation identified in this review. To enhance validity and relevance, future measures should integrate input from both experts and AYA with lived experience through structured participatory methods (e.g., focus groups, Delphi studies, interviews, cognitive interviewing, and pilot testing) and report these processes transparently. Many of the measures included in this review were developed within the framework of developmental psychology, often drawing on Eriksonian theories of identity formation. While these theoretical models offer valuable insights into normative identity development across the lifespan, they may not fully capture the unique challenges faced by AYA living with LTC-P. For example, identity processes in this population may be shaped by factors such as disrupted life trajectories ( Hale et al., 2015 ; Kirk & Hinton, 2019 ) and shifts in social roles ( Harris et al., 2003 ; Packham et al., 2020 ), experiences not typically reported in measures grounded in normative developmental theory. As such, the direct application of these tools to clinical populations should be approached with caution, and further validation work is needed to determine their relevance to the lived experience of those with LTC-P. It is worth noting that some measures included in this review ( April et al., 2022 ; Golub et al., 2013 ; Stepleman et al., 2017 ) explicitly address aspects of identity reconstruction, change, and growth in individuals living with LTC-P. These constructs may have important clinical relevance and could inform the development of targeted psychosocial interventions. A further consideration identified in this review was the variation in how identity was conceptualized across the included studies. Although this review does not favor one definition of identity over another, it emphasizes the importance of conceptual clarity. While variation in theoretical framing is not inherently problematic, different conceptualizations can offer valuable perspectives depending on research aims, it is essential that authors clearly articulate their definition and/or theoretical model/s of identity. However, many papers failed to provide an explicit conceptualization of identity, making it difficult to assess the appropriateness of the measures used or compare results across papers. The lack of clarity around the construct aiming to be measured introduces ambiguity and undermines the validity and utility of the research. Clear communication of how identity is defined and understood is therefore critical, particularly when assessing populations whose identity development may differ from normative patterns. Additionally, across the included studies, reporting of sociodemographic characteristics was often limited or inconsistent. Data on religion were entirely absent, and information on race and ethnicity was sparse and, on occasion, conflated into a single category. Reporting of sexual orientation was similarly limited (only four papers reported this), and data related to sex and gender were frequently reported interchangeably, despite representing distinct constructs. In several studies, sex and/or gender were reported only for a single category (e.g., percentage female), which meant it was not possible to determine if any transgender, intersex, or non-binary AYA were included in the sample. Socioeconomic status was most commonly reported on via educational attainment; however, this is of limited relevance for the AYA cohort considered in this review, as many AYA are likely to still be in education, to experience educational delays, or be in transition between educational stages. These limitations restrict the extent to which experiences of identity development among marginalized groups can be examined and highlights the need for future research to adapt more comprehensive and conceptually robust approaches to collection and reporting of sociodemographic data. This review provides the first systematic synthesis of identity-related measurement tools that have been used with AYA living with LTC-P. Adhering to COSMIN methodology, this review has produced a comprehensive and methodologically rigorous evaluation of the psychometric properties of included measures. Several limitations should be acknowledged. First, the use of the COSMIN framework was applied in full to only three measures, as analysis would only be meaningful where data were available from more than one study. As a result, many of the measures identified remain in need of further psychometric evaluation. Additionally, COSMIN’s stringent criteria, while promoting high standards in measure development, have been criticized as overly conservative ( Rothmund et al., 2023 ; Smith et al., 2021 ). As many measures were developed prior to the publication of key methodological guidelines, they may have been undervalued in this review. Additionally, the reliance on COSMIN’s lowest-rating counts scoring approach may have overlooked more nuanced differences in methodological quality across studies. Amendments were made to COSMIN scoring when necessary, e.g., by accepting evidence of factor structure from other studies to support internal consistency. The content validity assessment was conducted narratively, due to COSMIN’s requirement for both lived experience and expert involvement in measure development for any rating above “low quality.” Similarly, due to the diversity in constructs assessed across measures, it was not feasible to generate consistent hypotheses for construct validity testing, leading to an “indeterminate” rating for many studies that did not explicitly propose one. These adaptations, while necessary, may have underestimated some measures’ strengths. Further limitations relate to the scope and design of the review. Despite efforts to broaden the search strategy, incomplete reporting, inability to contact authors, and exclusion of non-English papers may have introduced cultural bias and limited the global generalizability of findings. Additionally, grey literature was not included, and the divergence of identity constructs and measures prevented meta-analysis. Moreover, at the full-text screening stage, missing participant age ranges may have led to the exclusion of potentially relevant papers. Relatedly, while the review aimed to focus on the 16–24 age group, many papers included participants outside of this range. As a result, some of the measures assessed may not have been developmentally appropriate for AYA or aligned with the specific identity tasks characteristic of this life stage. Future research and measure development should be guided by the specific needs of both researchers and clinicians. To enhance the applicability and utility of existing measures, further psychometric validation is required, with appropriate modifications to ensure their relevance to clinical populations. At present, weaknesses in validity limit the extent to which identity measures can support accurate clinical assessment or detect meaningful change following intervention. Particular attention should be paid to the 16–24 age group, as this developmental stage presents unique identity-related challenges ( Sawyer et al., 2018 ) that may not be adequately captured by tools developed for older populations. There is also a clear need for the development of new measures, or the redevelopment of existing measures, grounded in the lived experiences of AYA and clinical perspectives. This approach will help ensure content validity. Additionally, measures that are co-produced with individuals with lived experience and healthcare professionals are more likely to be clinically meaningful and better suited to inform interventions ( Flake et al., 2017 ; Terwee et al., 2007 ). The development of such measures should follow established best practice guidelines ( Mokkink et al., 2018 ; Vet et al., 2011 ). Review findings demonstrated that while identity has been addressed in a range of papers involving individuals with LTC-P, notable gaps remain in the literature. Certain populations (such as chronic pain, e.g., fibromyalgia, endometriosis, and Ehlers–Danlos syndrome) were absent, suggesting a need for further research that includes these overlooked groups. Additionally, there is a particular need for the development of measures tailored to these under-represented LTC-P, which often span multiple diagnostic categories and are marked by diverse symptom presentations. In addition, future papers should aim to evaluate key measurement properties that remain underexplored, including cross-cultural validity, measurement invariance, test–retest reliability, measurement error, criterion validity, and responsiveness. Without robust evidence for these properties, identity scores cannot be interpreted with confidence in research or clinical practice, nor used reliably to evaluate whether interventions are producing meaningful change. Notably, none of the reviewed measures examined the content of identity itself, an important and underdeveloped area that could serve as a valuable direction for future research ( Galliher et al., 2017 ). Finally, given the established relevance of identity to a range of health-related outcomes, including psychological well-being, treatment adherence, and quality of life among AYA with LTC-P, it is crucial that psychometrically robust, gold-standard measures are developed for specific clinical populations. Such measures are needed to accurately assess identity-related experiences in these contexts and to better inform the development of targeted interventions. The creation of these instruments should adhere to best practice guidelines for measure development set out by the COSMIN initiative ( Mokkink et al., 2018 ; Vet et al., 2011 ) to ensure the psychometric and methodological rigor required of measures, particularly if identity is to be considered a clinical outcome.

While the tentatively recommended IIQ and DIDS may be appropriate when their conceptualizations align with research aims, remaining measures within this review should be used with caution. Researchers must assess the relevance and comprehensiveness of any measure before applying it to a given population and ensure that its content is comprehensible to the target group, taking into account condition-specific factors.