1. Introduction

There is growing evidence that children both in the UK [1] and internationally [2, 3], are struggling to develop motor skills that were the norm for equivalently aged children from previous generations. There appears to be a particularly worrying downtrend in one specific group of motoric abilities known as fundamental movement skills (e.g. running, jumping, hopping, throwing, kicking and balance), which provide the building blocks for developing more complex movement patterns [4]. This is particularly concerning as motor difficulties in childhood are known to have a detrimental effect on education, health and wellbeing outcomes [5–7], which can ultimately impact social mobility [8, 9].

In the UK, there are multiple issues with the pathway to receive diagnosis and support for motor difficulties. Children need a diagnosis of Developmental Coordination Disorder or ICD equivalent (of which fundamental movement skills form part of the assessment criteria) to access support for these skills, but the healthcare system is complex and inequalities mean that parents/guardians of low socioeconomic position and ethnic minorities are less likely to access their family doctor for support [10]. Additionally, children with difficulties won’t be eligible for a diagnosis of Developmental Coordination Disorder if they have late onset of ‘symptoms’, if their difficulties don’t ‘significantly and persistently’ interfere with activities of daily living (e.g. dressing oneself) and academic achievement or productivity at school, or if they have an intellectual disability, visual impairment or neurological condition affecting movement [11]. Finally, there will be children that do not meet the criteria for this diagnosis but still have substantive functional motor difficulties and are therefore unable to access services. Considering all of these factors together, it is likely that many children with motor difficulties are not being identified and supported in a timely manner. This aligns with research which has shown international parental dissatisfaction for the support available for children with motor difficulties [12–15].

The impending ‘cost of living’ crisis is likely to have a disproportionate effect on disadvantaged families, further worsening inequality in these sections of society [16, 17]. It is therefore imperative that children with additional needs can be identified in a timely manner to ensure they do not fall further behind their peers. One way to achieve this could be to universally screen children for difficulties in education settings. Unfortunately, research has shown that pre-existing clinical tools for diagnosing Developmental Coordination Disorder, such as the Movement Assessment Battery for Children (MABC) [18], are not feasible for use in schools at scale as it is time consuming (1 hour per child), costly to purchase (~£1,000), and cannot be implemented by teachers, only trained clinicians [19, 20]https://www.zotero.org/google-docs/?KdchXa. In addition, less time consuming methodologies (such as proxy-report questionnaires) can be highly subjective and require knowledge about what constitutes a good level of motor development [21]. Research suggests that teachers often have low levels of understanding of fundamental movement skills [19] and thus it is likely these methods wouldn’t accurately identify children who would benefit from additional support.

To tackle these issues, a school-based screening tool (FUNMOVES) was developed with Primary schools in Bradford [22]. FUNMOVES allows two members of teaching staff to assess the fundamental movement skills of a whole class (~30 pupils) in under an hour [22]. The content and structural validity of FUNMOVES has been demonstrated when it is delivered by trained teachers [22]. An important next step in establishing the feasibility of school-based screening is to test whether teachers are reliable in their assessments and check that these tools are valid at detecting children who would likely receive a diagnosis of Developmental Coordination Disorder if they were referred into the health system for more comprehensive assessment.

This research aimed to evaluate inter-rater reliability (teacher vs researcher scoring), as well as concurrent and predictive validity of FUNMOVES against the MABC-2 [18]. The MABC-2 is endorsed by the European Academy for Childhood Disability as the ‘gold standard’ for use as part of the diagnostic process for Developmental Coordination Disorder [23].

2. Materials and methods

2.5 Analysis

All analyses were conducted in R and R Studio (versions 4.2.1). Alpha level was 0.05 for all analyses.

2.5.1 Inter-rater reliability.

Agreement between teachers and researchers (inter-rater reliability), was measured using Kappa statistics (both unweighted and linear weighted values were calculated).

2.5.2 Concurrent and predictive validity.

Agreement between tests (concurrent validity) and predictive validity were measured using sensitivity, specificity, positive predictive value, negative predictive value, and accuracy. Definitions for each of these and how they were calculated can be found in Table 1.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Concurrent and predictive validity analyses.

https://doi.org/10.1371/journal.pone.0297412.t001

Percentiles for FUNMOVES were calculated based on performance compared to the rest of the year group in the school. MABC percentiles were calculated using the standardised normative data for total score. For FUNMOVES a total score is derived from a sum of all activities. Information on calculating a total score for the MABC-2 can be found in the user manual [18]. FUNMOVES total score was also compared to the average percentile for MABC-2 subscales which contain fundamental movement skills (Aiming & Catching and Balance), removing manual dexterity, to generate a more directly comparable measure.

3. Results

3.1 Participants

No opt-out consent forms were returned to the school, so all children (present on days of testing) took part. A total of 217 children across school years 1–6 (ages 5–11 years) from a Primary School in Bradford participated in this research. Of this 217, a subsample of 119 children (one class per year group [1–6]) were scored by both researchers and teachers to evaluate inter-rater reliability. For concurrent and predictive validity, the school was asked to identify children that they believed had motor difficulties, out of the 217 recruited. Teachers identified 30 children, five from each year group. Of this subsample, 22 completed both the MABC and FUNMOVES, due to school absences (children who had full parental consent but were not present on days of testing), and thus were included in analyses. For demographic information about these samples, see Table 2. Nine teachers and nine teaching assistants scored the children using FUNMOVES.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Demographic information about the sample.

https://doi.org/10.1371/journal.pone.0297412.t002

3.2 Inter-rater reliability

Inter-rater reliability was ‘almost perfect’ for all individual activity scores [28]. The total score had ‘substantial’ levels of agreement [28], however when incorporating linear weighting this value improved substantially (see Table 3).

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Inter-rater reliability sensitivity of scoring.

https://doi.org/10.1371/journal.pone.0297412.t003

Difference scores were also calculated to evaluate how far from researcher scoring teachers were. For a full breakdown by activity see Table 3.

3.3 Concurrent and predictive validity

The proportion of true positives, true negatives, false positives and false negatives can be found in Table 4.

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. True positives, true negatives, false positives and false negatives.

https://doi.org/10.1371/journal.pone.0297412.t004

All six children with false negatives scored below the 15^th percentile on the Manual Dexterity subscale (mean = 5.08, SD = 4.32, 95%CI = 1.62–8.54). None of these children scored below the 15^th percentile on the Aiming and Catching subscale (mean = 36, SD = 22.27, 95%CI = 18.18–53.82). Five scored below the 15^th percentile on the Balance subscale (mean = 7.52, SD = 5.37, 95%CI = 2.81–12.23). With Aiming and Catching and Balance subscale percentiles averaged (calculating a fundamental movement skill only total score; mean = 22.33, SD = 10.67, 95%CI = 13.79–30.87), only one false negative remained. There were no false positives (i.e. children who would be accidentally referred, and thus would waste resources) in any combination of scores.

To further assess the utility of FUNMOVES as a screening tool for both overall motor ability (MABC-2 Total Score) and fundamental movement skill ability (MABC-2 Aiming & Catching and Balance), sensitivity, specificity, positive predictive value, negative predictive value and accuracy were calculated (see Table 5).

Download:

• PNG
  larger image
• TIFF
  original image

Table 5. Concurrent and predictive validity for FUNMOVES.

https://doi.org/10.1371/journal.pone.0297412.t005

4. Discussion

This research aimed to, for the first time, assess the inter-rater reliability, concurrent validity and predictive validity of FUNMOVES. The results show that teachers were able to score all activities within FUNMOVES in alignment with researchers after an hour of group training (unweighted Kappa > 0.85). This is a positive step forwards as many common motor skill assessment tools require trained professionals to deliver them, precluding their use by teaching staff [29]. To effectively screen children for fundamental movement skill difficulties, schools will need to be self-sufficient (with appropriate support in place) after training, and this would be made possible through the use of FUNMOVES.

It is important that FUNMOVES and the MABC-2 identify the same children as needing additional support. With healthcare services already stretched [26], it is vital that any children that are referred to these services do indeed need specialist healthcare support. Analyses revealed that the probability of FUNMOVES correctly classified all children that do not have motor difficulties (i.e. did not fall below the 15^th percentile on the MABC-2; specificity = 1, 95%CI = 0.63–1). Additionally, FUNMOVES had a positive predictive value of 1 (95%CI = 0.68–1)such that the percentage of children who score below the 15th percentile that also scored below this threshold on the MABC-2 total score was high. Conversely, negative predictive value (0.57, 95%CI = 0.42–0.71) and sensitivity (0.57, 95%CI = 0.29–0.82) were moderate, meaning that some children who score above the 15th percentile on FUNMOVES may be classified as ‘below average’ by the MABC-2 total score. Additionally, accuracy of FUNMOVES for correctly classifying children was 73%. It is likely that part of the reason for this is that the MABC-2 total score also incorporates manual dexterity (fine motor) skills (e.g., placing pegs, drawing a trail, and threading beads). FUNMOVES focuses on fundamental movement (gross motor) skills, which are a separate, distinct group of motor abilities. This is reflected in the fact that all of the false negatives scored below the 15th percentile on the manual dexterity subscale.

Research has suggested that teachers are well placed to identify children that have difficulties with fine motor skills, such as handwriting [30], which is perhaps expected given that this is often the method used to assess a child’s academic ability in school. Contrastingly, research has demonstrated a poor level of understanding amongst teachers about fundamental movement skills [19]. When looking at the two subscales of the MABC-2 which directly link to fundamental movement skills, the classification analyses were all strong (sensitivity = 0.89, specificity = 1, positive predictive value = 1, negative predictive value = 0.93 and accuracy = 95%). These preliminary results demonstrate that FUNMOVES may be an effective method of screening children for fundamental movement skill difficulties.

It is, however, important to consider that the sample size for these aspects of validity were small (n = 22). This limits the conclusions that can be drawn. Similarly, teachers were only asked to refer children they thought had motor skill difficulties for researcher-led assessment on the MABC-2, which could impact the generalisability of the results (along with convenience sampling)–however children within this sample did score within ‘normal’ ranges, showing a range of abilities. In addition to this, the children that were assessed on FUNMOVES received percentile scores based on their performance compared to the rest of the children in their year group. This does not necessarily represent their performance compared to a broader, nationally representative sample. It would be beneficial to formulate a normative database to understand fundamental movement skill ability levels beyond Bradford. Once a representative normative database has been formed, it would be beneficial to conduct this research on a larger scale. Nevertheless, the current study shows that such a programme of work is well justified.

FUNMOVES raw scores were converted into Percentile Ranks relative to the age of the children in the study. This was done to align with the MABC-2, and enable easy comparison of children struggling with motor development using the <15^th percentile cut off. It is, however important to note that there are a number of problems with use of these standardisation metrics which need to be considered. Firstly, it is likely that standardised samples do not include percentile ranks for each raw score that is achievable on the test, and secondly, extreme scores can distort the assignment of percentile ranks [31]. When developing a normative database for FUNMOVES it will be important to consider more continuous solutions to avoid these pitfalls to ensure that the correct children are identified as having difficulties [31]. The results did show though that FUNMOVES is a strong predictor of fundamental movement skill difficulties, as assessed by the MABC-2. By embedding FUNMOVES into school Physical Education lessons, and using its insights alongside teacher knowledge about their pupils’ classroom skills (e.g. handwriting and using scissors), there is the potential to reduce inequalities and improve access to support (provided there is sufficient communication and collaboration between education and health to facilitate quicker referrals based on FUNMOVES scores). It is, however, important to note that the MABC-2 isn’t the only metric used to detect Developmental Coordination Disorder [23], so children who score poorly on FUNMOVES may not always be eligible for a diagnosis. It is crucial that children waiting for diagnoses (considering the lengthy waiting time for assessment) and those who have difficulties but do not meet the criteria for diagnosis, that would also benefit from additional support to reach their full potential, are able to access it.

Previous research has shown that school-based assessment of motor skill difficulties (through placing Occupational Therapists within the classroom environment) can be effective at reducing time to diagnosis [32]. This research has shown that FUNMOVES may prove to be an effective way to triage access to services using a universal approach. Moreover, if paired with school-based intervention, FUNMOVES could help children receive more timely support as part of a needs-based approach, without the need for a diagnosis (thus supporting those who do not meet clinical thresholds for Developmental Coordination Disorder but still have substantive functional difficulties). This shift towards ‘participation’ in alignment with the international classification of functioning, disability and health framework developed by the World Health Organisation [33] will likely have broad-ranging benefits for population level education, health, wellbeing and life chances [5–7].

5. Conclusions

Screening children’s fundamental movement skills in schools using FUNMOVES has the potential to expedite access to further assessment and intervention through facilitating increased communication and collaboration between healthcare, families and education. FUNMOVES therefore offers a unique opportunity to use a robustly developed assessment with strong validity and reliability to universally screen ability levels in schools.