From an evidence-based practice viewpoint, the most important finding of this study is that self-reported asthenopia of moderate extent using the CITT Study Group’s criteria [
41,
50] appears to be extremely common (45%) in Australian middle school children. As expected, students with obvious binocular impairment reported the highest number of symptoms, followed closely by those with significant hyperopia (≥ + 2.00D). However, more surprisingly, the RESC-defined ‘emmetropic‘group (refractive errors up to + 2.00DS) demonstrated a mean symptom survey score of approximately 16, that is, a score one standard deviation above that found for those without strabismus and having normal binocular vision in the CITT clinical population using the same survey tool [
41]. Notably, myopes reported only marginally less severe symptomatology than did emmetropes. On the other hand, despite previous suggestions that astigmatism, particularly hyperopic astigmatism, might be a driver of asthenopia [
25‐
28], we found no systematic variation relating to astigmatic imagery. Thus, all refractive groups reported a comparatively high prevalence of symptoms on the CISS survey that was normed on individuals without binocular problems, suggesting that moderate to high refractive error alone cannot remain the sole criterion for managing visually related issues in children. Rather, as we [
60] have previously noted that 39.2% of young elementary school children (aged 8.3 ± 2.4 years) showed binocular vision anomalies, binocular vision assessment should be considered a necessary addition to standard vision screenings.
Prevalence of symptoms
The number of students in our unselected non-clinical population who reported symptoms is particularly high compared with the pooled prevalence of 19.7% experiencing asthenopia sourced from the five studies considered by Vilela et al. [
16] in their meta-analysis (one study from Australia in 2006, two from Sweden in 2006 and 2007, and two from India in 2011 and 2013) but using different survey tools. Against studies using the same CISS tool, the frequency of symptoms experienced by our students (score of 16.3 ± 10.9) is on average nearly twice that found in the CITT studies [
41] on children without binocular anomalies (CISS 8.1 ± 6.2), and is also higher than the scores found by Marran [
42] on students without hyperopic or astigmatic refractive error or reduced visual acuity(CISS 10.3 ± 8.2).
The most likely differences in prevalence and frequency of symptoms may be, in part, due to sampling differences with respect to individual difference among the population, age range, and the protocols used. In particular, most previous uses of this tool have used clinical populations and either excluded those with binocular vision anomalies in the control group [
41,
50] or excluded only those with amblyopia and strabismus [
19], or had no exclusion criteria for the control groups [
17,
18,
20]. A further population difference for the CITT studies is Borsting et al’s exclusion of children with attention deficit hyperactivity disorder (ADHD) or learning difficulties [
40,
41]. As it is our intention to develop a protocol that uses simple tools during an in-school vision screening in order to identify those students reporting asthenopic symptoms and hence require referral for further, more intensive clinical examination, for whatever reason, we deliberately chose to have no exclusion criteria. Furthermore, as the school involved in our study enrols children typical of mainstream schools in Australia, at this stage of developing our protocol we had no reason to separate out students known to have ADHD or learning difficulties. Pertinent to this, Barnhardt et al. [
61] found children with parent-reported ADHD did score higher on the CISS (22.2 ± 11.0 versus 13.8 ± 9.2), particularly on the ‘performance’ questions but not the ‘ocular’ questions. The consequence of our study having some students with ADHD or learning difficulties would be that our mean CISS score would be expected to be higher. Notably however, any causal relationships between increased CISS scoring on performance-related questions, binocular dysfunction and academic aspects such as poor reading, overall academic performance and the existence of ADHD have not been established [
43]. The reliability of the CISS itself should also be considered when comparing studies. Borsting et al. [
41] carried out an appraisal across 2 weeks of the reliability of the CISS on students aged 9 to 18 years who had CI, and found the test to be reliable with the difference in CISS scores being 0.98 ± 5.7.
In general, increasing symptom prevalence has been found with increasing age [
61,
62]. Rouse et al. [
62] found an increase in CISS scores of 2.6 when comparing the scores of adults to children, presumably due to greater periods of prolonged near work for adults [
61] and increased duration of handheld electronic devices for leisure [
44]. Our students were on average nearly 2 years older (13.17 ± 1.4 years) than those in the CITT group’s [
41] earlier study (mean 11.4 ± 2.2 years), and this age difference may account for a portion of the increase in symptoms found in our students. In a later study, the CITT group had a larger group of children, and found an age effect existed with increasing CISS scores in older students, which has been corroborated by others [
61]. Similarly, the children in Marron’s sample were on average 2 years younger than our students and exhibited a lower mean CISS [
42]. An Australian study by Ip et al. [
18] undertaken on children just 6 years old, far younger than those of the current study, revealed a low prevalence of symptoms (12.4%). In Sweden, Abdi et al. [
17] found that 26.7% of children in grades 4 and 8 in Sweden (aged 10 to 16 years) reported near symptoms. In the study by Tiwari in India [
20], the children were 2 years younger than our children and 32.2% reported asthenopic symptoms.
A further contributor to sampling differences could be the differing distribution of refractive errors across participants in each study in the meta analyses. We and others [
30] found symptoms to be greater in hyperopes, albeit, our numbers with moderate to high hyperopia were relatively small. The relative distributions of refractive errors for both CITT studies [
41,
50] was reported in their later study and indicated that respectively 38.9/30.4% were myopic > − 0.50DS (twice as many as the current study), 3.7/0.0% were hyperopic > + 1.00DS (compared with 10.3% in the current study) and 57.4/69.6% were emmetropic. The Swedish study [
17] and Australian study by Ip et al. [
18] used reasonably representative whole-of-school sampling. On the other hand, the Indian study [
20] examined children who until recently had been child-workers in the gem industry but were placed into further education as a result of government initiatives, with children from a similar socioeconomic background used as controls.
Further biases may come from the years between data collection, as recent evidence from studies in adults suggests that the rapid increase in the use of computers and handheld devices has led to an increase in symptomology [
44,
63]. It is pertinent to note that the data from the current study was collected at a time not dissimilar to the other studies already mentioned. Another bias that has been raised, concerns the ethnicity of the students in the current study. We have found no evidence in the literature linking racial or ethnic differences and the prevalence of asthenopia. Our students are mainly second-generation Australians living a typical western lifestyle and being schooled in English according to a government mandated curriculum. Currently there is only one study from the Middle East [
64]. This study used the CISS to monitor asthenopia in university students undergoing treatment for CI and found at the end of treatment a very significant improvement in CISS scores to values of 13.3 ± 7.5 and 11.3 ± 4.5 depending on which CI treatment. These post-treatment scores are well within the range of ‘normal’ CISS scores as determined by our Australian-born students of Lebanese background.
Symptom survey value
The use of formal surveys of symptomatology to replace less standardized structured history-based studies to determine potential existence of asthenopia, particularly relating to refractive error, has been limited [
17,
19,
40‐
42,
45,
65‐
69]. The advantage of a survey such as the CISS, is that the frequency with which symptoms occur is also established. A well-designed survey will usually ask important questions in more than one way as a check on reliability. Surveys also offer the advantage of the participant being left alone to consider and re-consider answers without dealing with the social stress of communicating with an unknown adult. Additionally, surveys can at times be administered by a lay person. The CISS was originally administered orally to each subject [
40,
41,
50] whereas in our case the teacher administered the CISS to the whole class as a written exercise. It is unknown whether this change in protocol for the CISS has yielded different scores compared with one-on-one administration. The student’s capacity to read and understand could be brought into question, although in the current case the students were aged 10 to 15 years, schooling was in English, plus, their scores from the Raven’s Coloured Progressive Matrices Test (1998 Edition) and teacher assessment would suggest < 1% potentially did not have the cognitive capacity to read and understand the questions. However, a potential disadvantage of a survey is that it is not possible to ask follow-up questions if an answer sounds unlikely [
70]. For example, how does the child conceptualize the time frame relating to each question: was it whilst they were actually answering the question, or, in recent times, or as far back as they could recall? For students who own spectacles, did they differentiate between comfort with/without spectacles? Clearly with respect to spectacle wear, if spectacles have been appropriately prescribed and are regularly worn, then the overall CISS scores should be lower. These possible differences in temporal-precision underpinning a student’s interpretation of symptoms have not been explored. Data regarding compliance and associated CISS scores for spectacle wear by students in the current cohort was not available. Hence, the finding here that high CISS scores remain common is all the more perturbing and demands follow-up using compliance diaries.
The potential value of a written survey of symptoms such as the CISS questionnaire is highlighted by the high scores we have attained from the amblyopes/strabismics and higher hyperopes. Although originally designed for use in cases of dysfunctional convergence, claims have been made that the CISS primarily identifies the binocular anomaly of accommodative insufficiency more so than convergence insufficiency [
42]. Thus, we see the CISS as potentially useful in informing not only the debate regarding identification of binocular anomalies, but also the debate regarding criteria for prescribing for children and identifying tolerance of uncorrected hyperopia. Our high scores argue for the use of the CISS as an evidence base in clinical decision-making relating to prolonged near work issues in children [
23,
42,
71‐
77].
Despite the greater degrees of asthenopia reported by individuals with strabismus and/or amblyopia, it could be argued that the CISS may not be necessary for children with ocular misalignment or poor visual acuity as they would be detected during an oculo-visual screening. However, four of the seven amblyopic students with reduced acuity without conspicuous strabismus could not recall a previous visual exam despite a visual acuity difference between their two eyes of between 2 and 5 lines. This finding of a high mean CISS for these students argues for the validity of including the CISS in a battery of tests conducted by laymen (e.g. by school teachers) to identify students visually at risk and worthy of further detailed follow-up. One could also argue that perhaps the high CISS scores from this category of students contributed significantly to our choice of the term ‘unexpectedly high’ in the title of this article. However, the mean CISS for non-amblyopic/strabismic students (16.3 ± 10.9) was only 0.5 lower than the mean CISS including amblyopic/strabismic students (largely due to the relatively low numbers of students with these particular visual problems). Thus, the CISS scores for students without amblyopia or strabismus remain (unexpectedly) high.
Exploring whether there is a physiological basis for the high symptom scores, as distinct from uncorrected refractive error, has not been explored in depth here, and hence remains to be investigated under more rigorous conditions than a typical screening. Although we suspect that unidentified binocular problems are a likely cause [
41,
42,
50], few studies to date have looked at the true prevalence of oculomotor or binocular vision anomalies in large cohorts of unselected school children [
78]. Previous studies indicate that based on the criterion of a single binocular vision sign, as many as 39% of children appear to have at least one aspect of impaired functional vision [
60] whereas based on multiple criteria it is likely that at least 10% of school children in Australia [
60] the USA [
41,
79] and Sweden [
17] will have a significant accommodative or convergence problem, or both. Currently, literature investigating the reading distance adopted for reading/writing at a desk or during near leisure activities [
80], especially for children [
44,
81], is scarce. In addition, the emerging worldwide use of handheld devices and associated short reading distances is concerning. For example, one study [
81] of fully-corrected myopic Chinese 6 to 13 year old children found the viewing distance was 21.3 ± 5.2 cm whilst playing handheld video games and 24.9 ± 5.8 cm whilst writing. This is especially concerning given that video games on handheld devices are often played for durations greater than 1 h [
82]. With adults, it has also been noted that the viewing distance when using a smartphone decreases significantly and the degree of asthenopia increases over a 1 h viewing period [
83]. The impact upon symptomatology of factors such as screen time, lighting, font size, distance from the device, etc., do not appear to have been explored in depth in children in the new era of digital media and merit investigation. Unfortunately, two recent studies on asthenopia during use of handheld devices both excluded those with amblyopia and/or strabismus [
44,
83].
Educational value
The pursuit of ophthalmic drivers behind asthenopic symptoms associated with near work in in children should be an important issue given that ocular discomfort has been shown to act as a disincentive to reading and academic progress [
37,
45‐
52]. The clinical guidelines from the professional association of ophthalmologists in the USA recommends prescribing for children with asthenopia [
84]. Therefore, the CISS could become a clinically valuable and useful tool to facilitate decision-making related to prescribing for refractive error, though worryingly, some ophthalmic practitioners do not subscribe to prescribing on the basis of the presence of symptoms alone [
77,
85]. Evidence exists indicating that uncorrected hyperopia during school years is more commonly linked to impaired cognitive functions [
86], impaired literacy [
51] and impaired visual motor integration [
87]. The current study showed moderate to high hyperopes to be highly symptomatic, suggesting that they may be more likely to benefit from refractive correction and thus minimize educational disadvantage. Correction of visual anomalies with spectacles in middle school children has also been found to lead to reporting of less symptoms [
65], though for not all children, as even some emmetropic children diagnosed with reading and writing difficulties still report significant asthenopic symptoms [
45]. Either way, our study suggests that a more comprehensive visual assessment than simply refractive error is required if best quality of life and academic outcomes are sought.