Background
Increasing levels of poor mental health in children and young people alongside inadequate responses from service providers have been globally concerning for the last two decades (Green et al.,
2005; Merikangas et al.,
2009; Pitchforth et al.,
2019; WHO,
2017). Mental health difficulties are common in young people, with current evidence suggesting one in six 5–16-year-olds have a probable mental health disorder in England, (NHS Digital,
2017,
2020,
2022,
2023). This demand is not currently being matched by expansion of service provision (Lenon,
2021). Early intervention models address some of the treatment disparity by investing upstream to reduce costs and delay or prevent the need for clinical appointments and inpatients stays (Knapp et al.,
2014; Wade et al.,
2007; Williams,
2013) alongside improving outcomes (Correll et al.,
2018). Key to early intervention is the ability to identify at-risk young people in the community (Weist et al.,
2007), with schools cited as important environments to access and support youth populations (Fazel et al.,
2014; Sanchez et al.,
2018).
Universal mental health screening (UMHS) in schools is the systematic assessment of the mental health and wellbeing status of a school population (Humphrey & Wigelsworth,
2016). Respondents to UMHS can be teachers, parents, or pupils themselves. A universal approach is argued to reduce the risk of pupils with emerging or unmet mental health needs being overlooked (Dvorsky et al.,
2013; Siceloff et al.,
2017) and acts as a population-based strategy for mental health illness prevention in the community (Humphrey & Wigelsworth,
2016; Purtle et al.,
2020). Recent evidence from Sekhar et al. (
2021) found that UMHS for adolescents significantly increased the likelihood of early detection and treatment uptake for major depression. Despite potential advantages of these approaches, school participation remains low (Brown,
2018; Bruhn et al.,
2014; Burns & Rapee,
2021).
The scalable nature of UMHS has led to increased availability of screening tools and systems aimed at detecting at-risk pupils in schools. Such tools vary widely in their design and characteristics which may differentially impact on stakeholder groups (Harrison et al.,
2013) particularly in terms of their acceptability and practical applications. For any healthcare intervention, including UMHS in schools, acceptability is a fundamental factor in ensuring successful adoption (Diepeveen et al.,
2013; Sekhon et al.,
2022; Wilson & Jungner,
1968). As such low acceptability is a central barrier to widespread implementation of UMHS (King,
2021) with schools and districts reluctant to adopt such programmes without assurance that they will be supported by key stakeholders such as teachers, parents and pupils (Soneson et al.,
2018; Stepanchak et al.,
2022; Williams,
2013).
Implementing UMHS in schools requires overcoming complex barriers and often hinging on aligning multiple stakeholder perceptions around perceived effectiveness and implementation costs. This includes instilling stakeholder trust in participating without potential negative consequences. Such as the perception of stigmatisation of pupils, which have been found in surveys targeting teachers (Graham et al.,
2011; Soneson et al.,
2018; Williams,
2013) and parents (Fox et al.,
2013). Teachers and parents may also hold concerns regarding the accuracy of screening (Soneson et al.,
2018) and the consequences of labelling students (Childs-Fegredo et al.,
2021). Pupils themselves have also reported concerns around anonymity and confidentiality, especially relating to the sharing of assessment data (Demkowicz et al.,
2020; Stepanchak et al.,
2022).
Despite its importance, systematic reviews on the acceptability of UMHS in schools have been infrequent. Anderson et al. (
2019) and Soneson et al. (
2020) conducted reviews on the cost, effectiveness and feasibility of UMHS in schools, with limited and inconclusive findings. A third systematic review by Brann et al. (
2020) assessed usability including acceptability as a component but focused exclusively on studies from the United States. Brann et al. (
2020) found some evidence for teacher acceptability however this was assessed only in a minority of studies and concluded that evidence on acceptability remains a significant gap in the literature. King (
2021) echoed these concerns, highlighting that limited evidence on acceptability remains a significant barrier to the full endorsement of such approaches. Without strong evidence on acceptability from key stakeholders, opportunities to fully implement and assess such approaches therefore remain unsupported.
Acceptability has been regarded by some as an abstract and difficult to define construct (Sekhon et al.,
2017). Wolf (
1978) defined acceptability within a narrow behavioural focus as a combination of social significance, acceptance, and satisfaction. In contrast Glover and Albers (
2007) described acceptability via the concept of “usability”, emphasising its feasibility, accessibility and utility to users. Proctor et al. (
2011) took an implementation science perspective that focused on seven implementation outcomes including “appropriateness” and “adoption” but viewing acceptability as a distinct subscale. Sekhon et al. (
2017), in their Theoretical Framework of Acceptability (TFA) take a more holistic multi-dimensional approach assessing both anticipated and experienced cognitive & emotional responses to interventions, while incorporating elements of usability. The current review most closely aligns to Sekhon et al. (
2017), focusing on different stakeholders’ perceptions of screening, incorporating affective attitudes alongside concerns and perceived costs of participation. Several acceptability tools have been used in the past to standardise acceptability data, for example the Usage Rating Profile (URP-A) (Briesch et al.,
2013) and Assessment Rating Profile-Revised (ARP-A) (Eckert et al.,
1999). However, many investigations into acceptability remain unstandardised with participants asked directly, often via survey, interview or focus group as to whether they find a program “acceptable” or not (Fox et al.,
2013; Woodrow et al.,
2022).
Given these concerns, acceptability is an important element to understand and will be required for the effective adoption of UMHS in schools. The current review aims to understand more about stakeholder acceptability for UMHS in schools to improve uptake and allow more students to benefit from early mental health interventions.
Methods
The review was carried out adhering to PRISMA guidelines and a protocol was registered with PROSPERO (ID no. CRD42022312218).
Inclusion and Exclusion Criteria
The review included all study types that captured acceptability data as part of a screening intervention or investigation into the acceptance of screening approaches. Student theses, other grey literature and unpublished reviews of all kinds were excluded to maintain a focus on peer-reviewed publications. Included studies had to contain a primary/elementary or secondary/high school population, either subjected to universal screening or providing acceptability data on proposed universal screening. Pre-school or post-16 populations were only included if part of a broader primary/elementary or secondary/high school population to maintain a focus on main statutory school ages. Informants could be pupils, parents, teachers or related school staff associated with schools. Screening could be for general or specific mental health disorder, socio-emotional behaviour or a combination of these.
Search Strategy
The electronic databases, MEDLINE, Embase, PyschINFO, Education Research Complete, ASSIA, and Web of Knowledge, were searched for relevant articles up until July 2023. Due to the limited availability of studies specifically addressing social acceptability data on UMHS, searches were expanded to include related concepts such as social validity, stakeholder experiences and usability.
Search Terms Included Combinations of Keywords
Mental health (e.g. “mental health disorders”, “anxiety disorders”), Emotional Wellbeing (e.g. “emotional health”, “well-being”), Screening/Assessment (e.g. “screen”, “universal screen*”), Stakeholders (e.g. “teach*”, “school*”), Attitude/Views (e.g. “acceptability”, “attitudes”).
To manage the commonality of these broader search terms (e.g. experience, attitudes) a proximity search strategy was designed to use within database search parameters. Full search terms and search strings for databases are included in the supplementary materials.
Screening and Selection
Two independent reviewers (CP, JK) selected studies in three stages: (1) all titles were assessed for removal of duplicates (2) initial screening of title and abstracts was based on an understanding of the inclusion and exclusion criteria (3) full texts of remaining articles were assessed using a detailed inclusion and exclusion criteria tool for inclusion in the review. Any disagreements in full text screening were resolved via a third reviewer from the research team (HT). Duplicates were systematically removed in Endnote in line with recommendations from Bramer et al. (
2016). After removal of duplicates a total of 9035 unique records were screened by both reviewers using Rayyan systematic reviewing software (Ouzzani et al.,
2016). At the title and abstract stage most exclusions were due to studies either not addressing universal screening in schools, not focused on relevant stakeholders or not assessing acceptability. Our exclusion rate likely reflects the specific focus of the review and the relatively recent development of UMHS in schools. Two blinded reviewers achieved a high agreement rate of 99.63% and a moderate level (Landis & Koch,
1977) of inter-rater reliability (IRR) of 0.520 (
p < 0.001) at the title and abstract stage. This increased to a substantial level IRR of 0.749 (
p < 0.001) at the full text screening stage, when more information from articles were available to reviewers.
Extraction tables were refined, piloted, and agreed via consultation with the research team and one reviewer carried out data extraction (CP) which was then checked by the second reviewer (JK). Data extracted were as follows: first author; year of publication; country of study origin; study design; acceptability informant type; school level; characteristics of screening sample; characteristics of acceptability sample; identification model (pupil/parent/teacher or prospective), screening condition (Mental health disorder, socioemotional behaviour or both); acceptability concept; acceptability outcome/findings/type; study aims; annual time of screening; estimated time taken to complete screening per pupil (estimated from number of items; short = less than 5 min, medium = 5–10 min or long = 11 min and over) and screening format (paper-based or digital). (extraction table available in supplementary information).
Quality Assessment
Included studies were appraised for quality using the Mixed Methods Appraisal Tool v2018 (MMAT) which has previously been employed as a critical appraisal tool for systematic reviews including mixed-study designs (Hong et al.,
2018). The tool covers five categories of study design: qualitative research; randomised controlled trials; non-randomised studies; quantitative descriptive studies and mixed methods studies and can be used for primary research based on experiment, observation, or simulation. The MMAT provides a set of criteria tailored to each study design to assess quality (e.g. “randomisation” for RCTs or “sufficient interpretation” for qualitative studies). A star rating/percentage score relating to the number of criteria met for the study design can be calculated to categorises studies into low (= <40%), medium (=60%) or high (= >80%) quality. Hong et al. (
2018) note that this is an approximation and the criteria a study fails on is also important to know therefore we provide the full table of MMAT criteria in the supplementary information. Due to resource constraints a random sample of 20% of studies (5 articles) were independently assessed by a second reviewer to ensure consistency in applying MMAT criteria. Across 35 individual criteria ratings the agreement between reviewers was 82.9%. Discrepancies were resolved through discussion aligning interpretations of quality criteria.
Method of Synthesis
Given the heterogeneity of included study in terms of design, intervention type, and outcome measures employed, a meta-analysis was deemed unsuitable. Instead, we performed a narrative synthesis alongside use of the Theoretical Framework for Acceptability (TFA) (Sekhon et al.,
2017,
2022) to analyse data. This was guided by a framework for systematic reviews by Popay et al. (
2006) employing four stages: (1) developing a theory of how interventions work, why and for whom; (2) developing a preliminary synthesis of findings of included studies; (3) exploring relationships in the data; and (4) assessing the robustness of the synthesis.
TFA components and scoring was applied to outcome data something informed by previous research on acceptability in childhood screening (Carlton et al.,
2021). The TFA has seven components; affective attitude; burden; ethicality; perceived effectiveness; intervention coherence; self-efficacy and opportunity costs, that align well to the area acceptability.
Components can be scored individually or averaged to calculate an overall acceptability score. The TFA also includes a temporal assessment as to whether acceptability was measured prospectively, concurrently or retrospectively to an intervention (screening).
We mapped acceptability findings to their corresponding TFA component and then converted scores to a 1–5 scale, higher scores indicating greater acceptability. Percentages were converted using a 20th percentile method and 1–6 Likert scales were proportionally mapped to the 1–5 scale (1–1.5 = 1, 1.6–2.5 = 2) to maintain original distribution and relative position. Aggregated TFA component scores were used to calculate mean and standard deviation summarising trends within the data. As aggregated study level scores were unsuitable for meta-analysis or statistical significance testing, we instead report means and standard deviations alongside our narrative syntheses. Three studies (all for school staff) could not be converted due to insufficient reporting which could not be aligned with TFA components leaving 21 studies that provided mappable quantitative data. Additional entries reflect cases where multiple screening characteristics or stakeholder data were extracted from the same study.
Qualitative data was also mapped to TFA components and thematic counts calculated to identify common themes in the data. These combined frameworks allowed us to conduct a structured synthesis whilst accommodating the complexity and limitations of assessing acceptability within universal mental health screening contexts.
Discussion
This review identified 28 studies reporting a total of 11,933 participants acceptability data on universal mental health screening in schools, with most originating from the United States or the United Kingdom. Heterogeneity of study design and lack of randomised control trials have been noted in similar reviews on UMHS in schools (Anderson et al.,
2019; Brann et al.,
2020; King,
2021; Soneson et al.,
2020). Such variability complicates synthesis and highlights a currently weak evidence base for acceptability research. The quality of included studies reflected this limitation with just under half of all included studies (
n = 11) meeting 40% (low) or less of MMATs quality criteria, another 9 meeting 60% (medium) and the remaining 7 studies at 80% (high) or higher. Quantitative descriptive studies tended to be weak on sample representativeness and risk of non-response bias, whilst mixed methods studies were limited by inadequate coherence and poor integration of findings. In contrast, qualitative designs performed better on quality criteria, perhaps reflecting that qualitative designs are more suited to meeting MMAT quality criteria for acceptability research.
Building on the findings of Brann et al. (
2020), who reviewed studies from the US and focused exclusively on socioemotional behavioural screening, our review provides a broad understanding with greater focus on acceptability from multiple countries and differing screening contexts. Our review findings support and expand the Brann et al. (
2020) review showing that school staff were the most assessed stakeholders reporting the highest level of acceptability, with parents and pupils reporting slightly lower but still moderate to high levels of acceptability. Similarly to Brann et al. (
2020), we conclude that acceptability is generally not a specific focus of research. Only 8 of the reviews included studies using validated measures to assess acceptability and most studies focused instead on intended use or appropriateness. Our findings suggest that screening of socio-emotional behavioural elements appears to be perceived as somewhat more acceptable to pupils and school staff than screening focused solely on mental health disorders.
Importantly, acceptability was often higher when data was collected concurrently or retrospectively to screening suggesting the possibility that experience and therefore familiarity with screening improves acceptability. However, it may also reflect sampling bias where individuals already engaged in screening are more likely to report positive views. This is of particular importance when considering the main active participants of screening, in this case pupils, something we explore in more depth in our subsequent section on pupil findings. Despite these insights, significant gaps in the literature remain with many reporting that approximately 20% of respondents do not positively endorse the acceptability of UMHS in schools. Understanding those who do not endorse UMHS in schools is crucial, especially as this group my include pupils who stand to most benefit from screening. Pupil demographic differences such as ethnicity and gender are insufficiently explored and something likely to impact engagement as seen on previous research on adolescents help-seeking for mental health (Planey et al.,
2019; Rickwood et al.,
2007). Understanding more about pupil engagement and acceptability of screening is critical to ensuring the proposed scalable nature of universal screening (Sekhar et al.,
2021) is equitable and does not exacerbate health inequalities.
School Staff
School staff reported a moderate to high level of acceptability overall and across all acceptability components. Qualitative data supported these findings, with school staff recognising key benefits of screening around universality (Dvorsky et al.,
2013) and early intervention (Weist et al.,
2007). However, qualitative data also suggested that acceptability may depend on support from health professionals. These findings map well with previous research indicating that teachers feel inadequate and under-resourced in managing pupils’ mental health concerns (Reinke et al.,
2011; Rothì et al.,
2008). Such concerns are not confined to teachers but have also been reported by other school mental health professionals (Burns & Rapee,
2021). This highlights the importance of training and support to build confidence in those involved in implementing screening. We also found intervention coherence was high for this stakeholder group which perhaps suggests concerns are more related to these more practical elements of resource management rather than a lack of understanding regarding programmes. In addition, teachers also held concerns around the potential for labelling and stigmatisation, an issue also that is also reflected in previous research (Graham et al.,
2011).
Parents
Parents were markedly less often consulted in research studies than school staff regarding the acceptability of UMHS in schools resulting in less reliable conclusions being possible. For this stakeholder group our findings generally suggest a moderate to high level of positive attitude towards screening but includes concerns around ethicality and the burden placed on schools. The limited qualitative data broadly supported these findings, however similar to school staff, parents held concerns around how poorly resourced screening approaches might impact support provided by schools. This echoed previous research on parent concerns around confidentiality and a perceived lack of effectiveness of under-resourced school-based mental health services (Ohan et al.,
2015). A greater emphasis on parental acceptability is required to better understand and address this groups acceptability needs, especially as parents play an integral role in child and adolescent help-seeking (Villatoro et al.,
2018).
Pupils
Pupils were another understudied group reflected in the small number of pupil respondents within studies of which none included primary school pupils. This again limits the reliability of conclusions and the ability to conduct any more detailed analysis given the small number of studies involved. Despite these limitations we found some evidence that pupils overall hold positive views in relation to screening. Yet, pupils also reported the lowest scores for ethicality (a sense of practices fitting with an individual’s values system) and perceived effect (the extent that an individual perceives an intervention to achieve its purpose) suggesting they hold concerns around potential negative consequences of screening and have less faith that screening may result in positive outcomes. This mapped well to qualitative findings, and we found that whilst pupils reported screening as a beneficial, effective and a good way to engage with support, they held concerns which centred around ethical aspects of screening such as confidentiality and data sharing. In several studies pupils stated they were more comfortable sharing data with health professionals than teachers themselves mapping to previous help-seeking research which suggests young people have preferences in relation to who they share their mental health data with seeking potential help-givers they can trust (Corry & Leavey,
2017; Del Mauro & Jackson Williams,
2013). More research is needed to explore these concerns around ethicality, including how identification and data sharing may impact pupil acceptability. Understanding nuances around why pupils appear positive towards the aims of screening but are less positive about how effective screening might be is central to ensuring pupils feel comfortable and willing to participate.
Implications for future research
Pupils are a critically under-studied group and from the limited data that is available, pupils report lower levels of acceptability than parents and teachers following participation with screening. Connected to this, available research only accounts for the views of those who engage and participate in screening, leaving us without insight into the motivations of those who do not choose to participate. This is particularly pressing considering pupils are the primary targets of screening, with their engagement being essential to meeting the inclusive and equitable goals of UMHS in schools (Humphrey & Wigelsworth,
2016).
Pupils who do not regard screening as acceptable may instead choose to either disengage completely or provide inauthentic responses to assessment measures therefore it is vital to address the frequency of dropouts and reasons for pupil disengagement or non-participation. To date no study included in reviews have reported on the demographics of pupils who decline screening. Only O’Dea et al. (
2019) provided some follow-up analysis of dropouts finding that for 33 pupils who declined screening the most frequent reason for non-participation was a lack of interest (45%). Collecting drop-out data is inherently challenging yet it is vital to identifying barriers to inclusion. Prospective and longitudinal research may in this case be more feasible and practical to fully understand who engages with screening. This is important as factors such as ethnicity have been found to predict post-screening engagement for mental health service referrals (Guo et al.,
2017) and pupils may provide inauthentic answers to avoid detection (Demkowicz et al.,
2020). Addressing these gaps is key to improving the responsible scalability, inclusivity and equity of universal screening programs.
For this review we posited > 4 (out of five) to represent a high level of acceptability. This is a somewhat arbitrary threshold, but some researchers may consider it sufficient in both a logical and pragmatic sense. As the evidence base for UMHS in schools continues to grow it would be useful if an agreed threshold for acceptability emerged alongside a more standardised definition and approach to measuring acceptability. This would also help schools to benchmark and monitor their own levels of acceptability when conducting UMHS in schools.
Implications for schools
Universal mental health screening in schools is a relatively recent development and therefore research around its effectiveness, feasibility and importantly its acceptability is still emerging. Whilst the evidence base is currently small regarding acceptability, what there is suggests mainly positive affective attitudes towards its use and aims. School staff in particular appear positive across most aspects of acceptability. As a foundation this should be encouraging for schools who might be considering engaging with this approach, yet they should also be aware that we are currently some way off a gold standard approach with different universal screening programmes and measurement tools entailing disparate elements which may impact on acceptability.
Kern et al. (
2017) highlights the importance of stakeholder involvement in the design and implementation of screening processes. Such surveys, particularly if anonymous, can help provide valuable insight into barriers and concerns that impact participation and acceptance. This proactive approach not only helps schools better understand demographics and concerns within their school but also fosters a sense of shared ownership of the screening process. In this review there was some evidence across all stakeholder groups that health professionals such as school nurses or counsellors were preferred to manage screening data. We also found some quantitative evidence that pupils and teachers found screening with a socioemotional behavioural element more acceptable than more MHD-focused screening. This aligns with previous research that well-being focused screening maybe less associated with stigma associated with mental health disorders (Dowdy et al.,
2012). Consequently, schools may find strength-focused or dual factor (an integration of MHD and SEB) approaches to UMHS are more positively received by stakeholders. Such approaches align naturally with the everyday language of schools and compliment common school-wide strategies of promoting positive development (Suldo & Shaffer,
2008), which may reduce some of the concerns regarding stigma. To secure trust with these groups schools should consider how best to protect pupil data and provide transparency around how data is processed and who it is shared with. Recommendations by Kern et al. (
2017) provide schools with good advice on effectively planning, communicating and consultation on screening that can improve stakeholder “buy-in” and reduce concerns. These recommendations can help schools improve their UMHS practices and in turn increase uptake and equity in pupils’ access to mental health support.
Strengths and Weaknesses
To our knowledge this is the first systematic review on the acceptability of UMHS in schools to include multiple screening designs, diverse stakeholders and range of countries in its scope. Understanding this could be a complex investigation we consulted with previous reviewers on UMHS and understood that assessment of acceptability is difficult as it is rarely included as a component of screening programmes (Anderson et al.,
2019; Brann et al.,
2020). To counter this we focused on a broad definition of acceptability using a comprehensive search strategy and applied a systematic appraisal to our data using the TFA. This in turn enabled the incorporation of results from bespoke measures across acceptability components to provide some granularity to findings alongside temporality of acceptability data. As this was a novel approach, we mapped the quantitative findings to a narrative synthesis to improve their reliability and cross-referencing with qualitative findings.
Assessing outcome data that is so highly heterogeneous is challenging and data were found to not be suitable for meta-analysis or statistical testing which would have provided greater reliability of findings, as such our findings remain descriptive. Another limitation of this review was the inability to extract quantitative data from 3 studies involving school staff therefore reducing the overall scope for analysis. The TFA provided a valuable framework for this review but there are limitations to its utility, and it is important to acknowledge the risks of subjectivity in application especially in mapping components to data. Another limitation of this study was that due to lack of resources only 20% inter rater checks could be conducted for quality assurance, albeit that a high-level of agreement on MMAT scoring was achieved. Despite these limitations we were able to provide some systematic assessment of the literature to progress research in this increasingly topical area.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.