We worked with two parents with lived and living experience of neurodevelopmental disorders on this review to ensure our work and findings were relevant to NDD families. Both parents were recruited through team members and have experience partnering with researchers on projects. One parent provided feedback on the protocol, search strategy, screening forms, data extraction forms, and on the descriptive presentation of the data (e.g., Table 1). The other parent worked with us on an infographic for families based on our findings.

Studies were eligible for inclusion if they involved children (0–25 years of age) with neurodevelopmental disorders (NDDs) or their caregivers of any age, including underserved groups such as fathers, immigrants, refugees, low-income, Indigenous, or racialized families, caregivers with disabilities, language minorities, those in rural/remote communities, or those identifying as 2SLGBTQIA + (Population). Eligible studies used photovoice methodology as defined by Wang and Burris (24) to engage these participants in research (Intervention). We included peer-reviewed qualitative, quantitative, mixed methods, or methods-focused studies published in English or French between 1994 (the date of the first photovoice study (25) and January 2025, with no geographical restrictions (Comparator not applicable). Outcomes of interest included the health research domains in which photovoice was applied, the demographics and diversity of children with NDDs and their caregivers, feasibility considerations (contextual factors, adaptations, practicality), and recommendations or future research directions (Outcomes).

Studies were excluded if they: (1) focused on adults (26 years of age or older) living with a NDD, or (2) focused on parents or caregivers of adults (26 years of age or older) living with a NDD, or (3) focused on clinicians/staff, or (4) were case reports, editorials, conference proceedings, abstracts, reviews, protocols, books, book chapters, commentaries, theses/dissertations, or research briefs. As per the DSM-5 definition of neurodevelopmental disorders (1), NDDs of interest were: intellectual disabilities, communication disorders, autism spectrum disorder, attention deficit hyperactivity disorder, specified learning disorders, motor disorders, Tic Disorders, and other neurodevelopmental disorders (other specified neurodevelopmental disorder, unspecified neurodevelopmental disorder). Our definition of children was informed by the Lancet Commission on Adolescent Health and Well-being and included those aged 0–25 years (56). Caregivers referred to unpaid adult family members (e.g., parents) caring for the person living with a NDD (57).

The search strategy was developed by a professional librarian (CJN) in Medline, and peer reviewed by a second librarian (RC) using the PRESS checklist as a template (58). We searched five bibliographic databases to identify potentially relevant publications for this review: Ovid MEDLINE(R) and Epub Ahead of Print, In-Process, In-Data-Review & Other Non-Indexed Citations and Daily; Ovid Embase; Ovid APA PsycInfo; CINAHL with Full Text (EBSCOhost); and Academic Search Complete (EBSCOhost). The only database limits used for the search was a limit to English- or French-language publications. The search was last conducted on January 08, 2025. Complete search histories for each database are available online (https://doi.org/10.34990/FK2/7J5JNK).

Search results were imported into Covidence (59) to manage review and selection of articles including rationale for exclusion. Duplicate records were removed by the Covidence software and manually. The title and abstract screening form was first pilot tested. Three reviewers (MG, PYY, SEN) screened three articles, discussed ratings, and the screening form was revised based on discussions. To ensure reliability across reviewers, a calibration exercise was completed where all three reviewers screened 10 articles. We used percent agreement as the measure of interrater reliability as all reviewers had training and experience with the screening process (60). Interrater agreement was set at a threshold of 80%. The calibration exercise for the title and abstract screening phase was completed twice at which point percent agreement was 96% across reviewers. Two dyads of reviewers (MG and PYY, MG and SEN) then independently screened the title and abstract of selected studies for relevance. Disagreements were resolved through discussion and in consultation with the senior author on the team (RLW). The same process was used for full text screening. The same two dyads of reviewers (MG and PYY, MG and SEN) completed the calibration exercise twice and achieved inter-rater agreement of 96%. The same reviewers then screened selected studies using a full text screening form that outlined the inclusion criteria. Two reviewers agreed on articles to be included, excluded, and disagreements were resolved through discussion and in consultation with RLW.

Two reviewers (MG and PYY) critically appraised each included study independently. Any disagreements that arose between the reviewers were resolved through discussion or with a third reviewer (RLW). The Mixed Methods Appraisal Tool (MMAT) version 2018 (61) was used for this purpose given this version was specifically developed for systematic reviews that integrate qualitative, quantitative, and mixed methods evidence (62, 63). This tool provides methodological quality criteria for five different study designs including qualitative, quantitative, mixed methods studies. For each category of study design, the MMAT provides a checklist of five criteria with response options of “Yes,” “No,” or “Can’t tell.” For qualitative studies, methodological quality criteria include: fit between research approach and research question, adequacy of qualitative data collection methods, whether findings are derived from the data, whether results interpretation is substantiated by data, and coherence between qualitative data sources, collection, analysis, and interpretation (61). Mixed methods studies are appraised using the appropriate criteria for their qualitative, quantitative, and mixed methods components. In line with the recommendations for using this version of the MMAT (61), there were no scores calculated for overall study quality as using a global score is less informative than evaluating quality based on specific criteria. Strengths of the MMAT tool include ease of use (has limited number of core criteria per study design), focus on methodological quality and trustworthiness, and good content validity (61).

Three reviewers (MG, PYY, SEN) first piloted the data collection form. Reviewers extracted data from two of the selected articles, discussed issues that arose while extracting data, and the form was revised based on discussions. One reviewer (MG) then extracted data from selected studies, a second reviewer (PYY) verified the extracted data, and discrepancies were resolved through discussion and in consultation with the senior author on the team (RLW). The following data were extracted: author, year of publication, location (country), research design, recruitment setting, sample population (e.g., youth, caregiver), population subgroup (e.g., immigrant family), sample size, child age, child gender, child disability, child functioning ability, assistive devices used by child (if any), caregiver age, caregiver gender, study research question, health area photovoice was being used to understand, how photovoice was implemented (adaptations made), challenges encountered, author-reported usefulness, and recommendations made to advance the use of photovoice including future research directions.

Study characteristics were synthesized descriptively and presented in tabular form. We used a convergent integrated synthesis approach to synthesize qualitative and quantitative data. In line with this approach, quantitative data were coded (textual descriptions) separately as were the qualitative data (55). The transformation of quantitative data into codes, allowed for the integration with qualitative findings into a single dataset. The combined dataset was then analyzed using a qualitative content analysis approach (64). We chose this analytic approach as it allowed us to identify categories and findings while also documenting frequency of utterances within each category to support interpretation. Our content analysis approach involved three steps: (1) grouping codes into categories, (2) documenting number of utterances per category and studies reporting adaptations or recommendations respectively, and (3) integrating categories into six main findings: health research areas that have used photovoice and diversity of study participants, contextual considerations when using photovoice with this population, adaptations made when using photovoice with this population, practicality of photovoice as reported by authors, recommendations reported by authors, and future research questions suggested by authors. For the adaptations data, we positioned our results within the photovoice steps recommended by Wang and Burris (24) to facilitate contextualizing our findings. The data were coded by one team member (MG), reviewed for congruence of coding by a second team member (RLW), and discrepancies were resolved by discussion until consensus was reached. To enhance methodological rigour, we used careful line-by-line analysis of the extracted data, prolonged engagement with the data, and ensured that two researchers engaged with the data and a third researcher (PYY) was available to resolve discrepancies.

Two changes from the published protocol were made in response to the nature of the final set of included studies and the data extracted. First, the original protocol proposed exploring facilitators to photovoice use. However, during data extraction, it became evident that many “facilitators” were adaptations, a focus already addressed by another review question. To avoid conceptual overlap and redundancy, we reframed this question to examine “practicality” as a dimension of feasibility (49, 50). This shift allowed us to more meaningfully contextualize implementation-related findings (e.g., adaptations). Second, the published protocol proposed using a descriptive synthesis approach. However, given the inclusion of both qualitative and mixed-methods studies, we adopted a convergent integrated synthesis approach and used qualitative content analysis to analyze the combined dataset (as noted in previous section).

The search yielded a total of 1,806 articles, leaving 1,139 articles after duplicates were removed. Based on title and abstract screening, 184 articles were selected for full text screening. Of these, 18 articles met the criteria for inclusion in this review. See Figure 1 for details about the search and selection processes. For information about excluded studies during full text screening, see Supplementary Material Table 2.

The MMAT results are shown in Supplementary Material Table 3. Of the 16 qualitative studies, 12 met all MMAT criteria for good quality (65–76), one study lacked quotes and visual data to justify the reported themes (77), and the remaining three studies lacked detailed information about their analytic approach (78–80). The two mixed-methods studies demonstrated overall good quality (81, 82). They satisfied all five MMAT qualitative criteria. When the quantitative components of these studies were assessed, it was determined that their samples were not representative of the target population due to the use of purposive (82) and convenience sampling (81). These studies met four of the five MMAT mixed methods criteria, with neither study providing a rationale for using a mixed methods design. Overall, the studies included in this review were determined to be of good quality. The majority met all relevant criteria with remaining studies falling short in one or two areas, indicating that the research was generally robust.

Qualitative Studies (n = 16). The qualitative studies (see Table 1) were conducted in Canada, the U.S.A., Australia, Vietnam, the U.K., and the Republic of Ireland. In these studies, participants were recruited through community clinics (66, 67), education settings (65, 68, 73, 75), community organizations (70, 71, 72, 74, 76, 77, 78, 81), or from various soucers (e.g., networks, schools, online platforms) (69, 80). Diagnoses and conditions identified included autism spectrum disorder, intellectual disabilities and/or developmental disabilities, cerebral palsy, and fetal alcohol spectrum disorder. Data collection methods used included semi-structured interviews, photovoice, focus groups, and field observations.

Mixed Methods Studies (n = 2). Both mixed-method studies (see Table 1) were conducted in the

USA (81, 82). One study recruited participants from a community organization serving children with autism (81) while the other study's recruitment setting was not reported (82). Cerebral palsy and autism were the disabilities of interest in these studies. Data were collected via photovoice, interviews (semi-structured and in-depth), parent focus groups, standardized scales, accelerometry data, activity logs, and questionnaires developed by researchers. For additional information about the included studies, see Supplementary Material Table 4 and Table 5.

We identified six main findings and describe them below.

The selected studies clearly described using photovoice across six research areas. One research area was methodological and reflected studies whose purpose was to describe how photovoice had been used to meaningfully engage our population of interest. The remaining five research areas explored the experiences of children with NDDs or their caregivers in life situations (83): in education settings (65, 80), physical activity participation (81), social life (72, 79), transitioning into adulthood (70), and experiences of health and wellness (68). These research areas focus on individual (e.g., perspectives), interpersonal (e.g., friendship), and organizational (e.g., school) contexts that influence an individual (84, 85). Lacking is photovoice research with NDD families exploring community-level (e.g., design, spaces) and public policy level (e.g., federal, provincial) influences on their lives.

With regards to diversity of study participants, there was limited diversity of participants in the included studies. The majority of the youth were white and male. Where caregiver gender and ethnicity were reported, caregivers were predominantly white and female. Further, no studies were identified that focused on immigrant, refugee, low-income, or racialized families, or caregivers who identify as having a disability, being a language minority, or being 2SLGBTQIA + . The lack of diversity in health research is well-documented (86–89). Our finding regarding the sex and ethnicity of study participants is also consistent with research that indicates that some NDDs (e.g., autism) predominantly affect males (90) and that females are more likely than males to be caregivers (57). Nevertheless, the lack of diverse representation of study participants found in our review highlights the need for research to better understand how photovoice can be adapted to engage these subgroups of the population and make research participation more accessible.

In terms of neurodevelopmental disorders represented in the included studies, the majority of the studies were conducted with children with autism spectrum disorder (ASD) and/or intellectual disabilities (ID). This finding is consistent with the focus of recent reviews that have explored the use of photovoice with people (any age) with ID (44) or individuals (26 years of age or younger) with ASD (43). The remaining studies included in this review focused on cerebral palsy (n = 1), fetal alcohol spectrum disorder (n = 1), and in one study, children had comorbid NDDs. This highlights the scarcity of and need for photovoice research with other NDDs.

In this review we were interested in examining author-reported practicality of photovoice-that is, whether study authors reported that photovoice can be used with children with NDDs and their caregivers despite contextual constraints (e.g., available resources). In all but two of the included studies (where no information was provided in this regard), authors highlighted that photovoice can be used and described the methodology as valuable in helping participants express their experiences, facilitating meaningful engagement, and promoting empowerment. This finding is consistent with the literature on the feasibility of photovoice for adults with disabilities (47, 91) and other underserved groups (38, 92–94). Further, in eight of the included studies, study participants were asked for feedback about their experience with research activities including taking photos. This feedback was positive (e.g., participants enjoyed the photovoice process) and supports using photovoice to engage this population in research. Synthesizing the literature on NDD families' experiences of using photovoice was beyond the scope of this review as was synthesizing results of studies that use standardized instruments to measure practicality. Research in these two areas would strengthen the validity of our results.

Crucial contextual considerations need to be taken into account when implementing photovoice with children (0–25 years) with NDD and their caregivers. Consistent with our findings, previous literature highlights the challenges of implementing photovoice due to participants' physical and cognitive functioning which leads to difficulty taking photos and talking about the photos (95, 96). However, our findings underscore the need for flexibility, patience, and time to mitigate these difficulties. Another contextual consideration we identified is the provision of additional support required in the form of modification and human assistance in the use of equipment or ascertaining meaning from the photos, consistent with previous literature (91, 97, 98). Fear of criticism when sharing photos in a small group setting or of breaking the rules of photo-taking was another contextual consideration identified. Fear of public scrutiny when working in groups or when sharing photos has been previously documented (91, 97). The contextual considerations we have identified in this review highlight the specific needs of children and youth with NDDs (0–25 years) that may be different from other age groups and conditions.

The adaptations documented in this review reflect flexibility, creativity, and the importance of using a strength-based approach when using photovoice with NDD families. For instance, to express their experiences, participants were asked to take photos using disposable cameras, their own devices, or to choose from a wide range of creative possibilities (e.g., web images, magazine images, drawings). Time provided to take or compile photos was also flexible. The adaptations made in the included studies point to using the nine steps proposed by Wang and Burris (24) as a flexible guide. This is consistent with research that highlights the importance of using photovoice flexibly to meet the needs of project participants (39, 99, 100). A notable finding regarding adaptations is that when discussing photos with participants, only two of the included studies used the SHOWeD method. To discuss photos, study authors adapted interviews or meetings and used creative methods to ask participants about the photos (e.g., used simple verbal prompts, child could describe what they wanted about the photo). This finding corroborates previous work that has highlighted the need for using methods that are creative and person-centred to meet the communication preferences and needs of each child (101–106).

Beyond highlighting the importance of exploring how photovoice can be integrated into practice settings, authors of selected studies provided recommendations for future photovoice research projects. These recommendations point to key messages. First, our findings emphasize the importance of building rapport and developing partnerships with participants and advocates to facilitate active photovoice involvement in studies. Various strategies were suggested including working one-on-one with participants and working closely with advocates (e.g., parents, teachers) to involve them in the research and get them to support children/youth throughout the research process. Meaningful connections can harness trusting relationships, and trust is critically important in the involvement of individuals with disabilities in research (107). In addition, facilitating meaningful research engagement through inclusive strategies (e.g., providing choice and ownership of research process, communicating respect) is crucial and requires thoughtful consideration. Effective, inclusive strategies to this end are needed and necessitate research attention (108–111). Third, interviews can be useful when researchers take the time to find out as much as possible about participants (e.g., needs,), incorporate multiple visits into the study's design, and use specific strategies such as allowing time for repetition and comprehension, rephrasing questions, and pointing to the photos to re-direct participants to the interview topic. This aligns with previous recommendations for interviewing children with disabilities (104, 112, 113) and with literature highlighting interviews as useful for engaging children with disabilities in research (110, 114, 115). Fourth, triangulation and including the child/youth views can prove useful when interpreting photographs. Various strategies were suggested to this end including using multiple data sources and perspectives and ensuring analysis involves a two-step approach (ensuring children are co-creators of meaning during step one and using the co-creation of meaning to inform the analysis conducted by researchers during step two). This recommendation supports the literature advocating for the use of multiple data collection methods and perspectives to assist with the interpretation process (101, 116–120). Strategies for mitigating potential biases introduced by intermediaries should be given careful thought. Photovoice projects should involve children/youth living with neurodevelopmental disorders (NDD) in data interpretation. Finally, these recommendations as well as the feasibility considerations we have documented point to the need to sustain mindful presence when conducting photovoice research with NDD families (121). This means moving through the research with careful forethought, attentive pace, receptive attention, as well as respect, openness, flexibility, empathy, and relational engagement (121). Adopting these qualities can facilitate building rapport, creating a safe and comfortable research space, adapting methods based on the child's strengths and needs, and anticipating and mitigating potential challenges. Being present and attentive to context can also help researchers remain attuned to the various ways children communicate as well as their physical and emotional wellbeing throughout the research process.

The results of this review indicate that research is needed about how photovoice can be used with children who have significant verbal or intellectual limitations. Young children, especially those with limited language or cognitive skills, have the potential to be engaged using photovoice given the fewer demands on literacy and language skills than required in other research methods (122). Research in this area can fill gaps that were previously identified in the literature about how young children with disabilities can partner or participate in research (110). Drawing from existing research that has examined the systemic exclusion of young children with disabilities and individuals with significant intellectual and cognitive disabilities may shed light on the ethical and practical considerations that can be applied when using photovoice to increase the inclusion of these populations in research.

Future photovoice research should also focus on the experiences of NDD families from diverse backgrounds and settings. Disability research can run the risk of being conducted in environments that are not typical of every-day life, limiting opportunities for generalizability and meaningful impact. Photovoice can open possibilities for conducting research that is embedded in individuals' environments, empowering participants to capture ideas that are personally meaningful. This can be a vital shift for people who are from geographic, cultural or linguistic communities that are often excluded in research. Future research directions highlighted in this review also point to exploring how photovoice can be used to meaningfully engage families affected by NDDs in health research, service design, and policy. We also believe that it is essential for NDD families to be a part of communities of practice to define research about NDDs and decide how and when to use photovoice (123). Finally, three of the studies included in this review conducted photovoice via virtual platforms. Implementing online photovoice with NDD families is in its nascent stages and has yet to be fully explored.

First, the majority of studies included in this review focused on autism and intellectual disabilities limiting the generalizability of our findings to other NDDs. Second, although the scope of this review was broad and designed to capture photovoice studies with NDD families of diverse backgrounds, studies focusing on diverse families were limited. Third, this review was limited to studies published in English or French. While this may have excluded relevant studies in other languages, the decision was based on the language proficiency of the review team and resource constraints (availability of funds for translation purposes). This may introduce some bias in the results and should be considered when interpreting the findings. Finally, although various dimensions of feasibility have been proposed, we focused on practicality, contextual considerations, and adaptation. Thus, we have synthesized the literature on these three feasibility considerations. Reviews addressing other dimensions of feasibility (e.g., methodology's acceptability, appropriateness) are warranted to provide a more comprehensive picture of the feasibility of photovoice with this population.