This is a descriptive, observational and cross-sectional study, which included children without known eye and/or systemic diseases. The protocol followed the tenets of the Declaration of Helsinki. The study was reviewed and approved by the ethical committee of the Universidad Europea de Madrid (Spain), and the study participants have signed an informed consent form.

The sample was selected by means of a convenience sampling undertaken in three schools (Carmelo Teresiano school in Madrid, Dulce Chacón Public School, in Fuenlabrada, and Hermanos Torá public school in Humanes) and a hospital (Rey Juan Carlos University Hospital of Mostoles) all of them located in the Community of Madrid in a percentage of children of each age which is representative of the frequency by which it is detected among the Spanish census population. The census distribution was obtained from http://www.ine.es/dyngs/INEbase/es/categoria.htm?c=Estadistica_P&cid=1254735572981

The children included in the study conducted at schools ranked between pre-primary education students and sixth grade students, which are included by age in the range of 3–12 years. The children of the hospital facilities included in the study were those who attended a new first routine ophthalmology examination and who did not show any obvious eye pathology or symptoms of refractive error. Children with systemic or eye pathologies were excluded from the study. The existence of the pathology was not determined by means of complementary explorations performed additionally to the measuring of visual acuity, but on the basis of the information provided by the parents or guardians in the anamnesis that had been previously done. A total of 1548 subjects between 3 and 12 years old were recruited, all of them of Caucasian origin. The study subjects were included during the third school quarter of the 2016–2017 academic year and the first quarter of 2017–2018.

The visual acuity measurement was performed with the Bueno-Matilla vision unit (UBM). This is a CE certified medical device for Class I visual examination, registered with the British Medicines and Healthcare products Regulatory Agency under reference number 5943 (https://optonet.es/). The unit consists of a computer application which incorporates, along with the VA test, other different optometric tests frequently used in the Opticians’ offices for the exploration of the eye system.

The test was displayed on a 24'' monitor, which was checked and calibrated on a daily basis for the distance selected, which was three meters. In all cases, the calibration distance of the test was respected and the lighting conditions of the room were kept within a pre-determined range. The illuminance conditions of the room were measured with a Testo 540 luxometer (Testo SA, Cabrils, Barcelona, Spain) and they should fall within the range of 400 lux and 600 lux. According to Tidbury et al.,9 if this range is maintained during the measuring of visual acuity, the effect caused by the different illuminance is limited to 0012 LogMAR.

For the collection of demographic data and the monocular and binocular decimal VA assessments, a space was enabled in each school, where students were taken for the screening, separated by group and by school levels. The testing protocol for each patient was performed by the same optometrist, previously trained to avoid interobserver errors. In schools, there could be several children waiting in the same measuring room; consequently, in all cases, and to prevent them from memorizing the letters, the tests were presented randomly. The data collection was performed in the following order: evaluation of VA of the right eye, evaluation of VA of the left eye, and evaluation of binocular VA.

The VA measurement was always performed without optical compensation of the patients, or with their usual optical compensation, in the case of those patients who use glasses, contact lenses or are subject to an orthokeratology treatment. Subjective refraction was not performed to determine the refractive status of the patient, as it was not part of the study protocol.

The measurement was taken with the patient seated and trying to keep the monitor always at the child's eyeline level at all times.

The paediatric optotype test used was that of symmetrical letters (Fig. 1) that was being displayed by isolated VA lines containing 5 letters each. For the measurement of VA in younger individuals, over 36 months and under 72 months, the recommendations published by Cotter et al. in 201510 were followed. For younger patients or those who were unable to vocalize their responses, pairing cards were used, consisting of letters printed on a sheet, wherein the child could point out the letter which is being shown to it on the screen. For the measurement of monocular VA, occlusion glasses for children were used, which causes less rejection than the eye patch.

Figure 1.

Test of symmetrical letter optotypes of the Bueno Matilla Unit used for the determination of the visual acuity of the subjects evaluated during the study, wherein the same letters considered symmetrical (H, T, V, U, O, X) are repeated, varying in size with logarithmic reduction.

(0.06MB).

The measurement was done by isolating the rows of letters used to measure visual acuity with contour separation, following the procedure recommended by Bueno-Matilla, designers of the test.

Each complete line of the test read corresponded to 0.1 VA units and each supplementary letter equalled to 0.02 units. So, for example, the correct visualization of all the letters of the 0.5 line plus 2 additional letters of the 0,6 line would give a score of 0.54.

Before commencing the study, it has been made a preliminary study to ascertain the intersession repeatability of the test used, including a comparison of the results obtained for left and right eye decimal VA with that sample. During that pilot test, since the Bueno-Matilla unit allows the use of different letter-based tests, the eventual existence of significant differences between the results obtained through the symmetrical letters test and the Sloan Letter Chart and the Landolt C tests was also analysed.

Data obtained from study variables such as age and monocular and binocular VA were sent anonymized to the statistical department of the university where the statistical study was conducted. The statistical analysis consisted of:

To analyse the agreement between the tests, both in the pilot test and in the total sample, normality was assessed in the first place in order to apply parametric or non-parametric tests. Subsequently, we calculated the effect size, the intraclass correlation coefficient, the coefficients of variation and the Bland–Altman limits of agreement.

Also, in addition to the above-mentioned analysis of agreement, and in order to verify the effect of age, an ANOVA was implemented to ascertain the eye by participants‘age interaction. Likewise, we also calculated the descriptive statistical values as a boxplot diagram and the 5th, 10th, 25th, 50th, 75th, 90th and 95th percentiles for monocular and binocular VA.

All the analyses have been made with a 5% significance level (p < 005), using the IBM SPSS Statistics v.24 software (IBM Corporation, Armonk, NY, EEUU). To determine the normality and the effect size in the pilot survey, we have used the JASP v.0.13.1 software (Jasp Team, University of Amsterdam, Amsterdam, The Netherlands).

In order to homogenize the sample with respect to the census of the Spanish population, a series of individuals of 4, 7, 8, 9 and 10 years had to be randomly removed, thus leaving the sample in 1300 subjects distributed according to the census population frequency (Table 1). 51.1% of participants were male patients.

• A

  Comparison between the decimal VA (DecVA) of the right eye and the left eye

The comparison between the DecVA of the right eye and the left eye was analysed in the preliminary study performed with 36 subjects. To analyse whether there were differences between the measurements, the Shapiro-Wilk test of the variable was performed in the first place, resulting in acceptance of normal distribution (p = 0067). Therefore, the t-Student test was then executed, but it did not show any significant differences (p = 0486). Subsequently, effect size (ES) by Cohen´s d, intraclass correlation coefficient (ICC), the coefficient of variation (CV) of both measurements and the limits of agreement (Bland–Altman) were calculated, obtaining the following values: −0,11 [95% CI: −0,44;0,21] for the value of d; 0.70 for the ICC; and 0.17 and 0,20 as CV respectively. The Bland–Altman concordance analysis is attached as Fig. 2.

• B

  Test repeatability

Figure 2.

The Bland Altman plot for right eye (RE) versus left eye (LE) in the preliminary study. (DecVA: Decimal visual acuity).

(0.29MB).

The repeatability of the test used (measurement 1 versus measurement 2) was analysed in the same preliminary study. To analyse whether there were differences between the measurements, the Shapiro–Wilk test of the variable was performed in the first place, resulting in the rejection of normal distribution (p < 0,01). Therefore, the Wilcoxon signed-rank test was then executed, but it did not show any significant differences (p = 0,406). Subsequently, ES by rank biserial correlation, ICC, CV of both measurements and the limits of agreement were calculated, obtaining the following values: 0.92 [95% CI: 0,67;0,98] for the rank biserial correlation; 0.97 for the ICC and 0.16 for both CV measurements. The Bland–Altman concordance analysis is attached as Fig. 3.

• C

  Agreement between VA test

Figure 3.

The Bland–Altman plot for Measurement 1 (M1) versus Measurement 2 (M2) in the preliminary study. (DecVA: decimal visual acuity).

(0.18MB).

The VA data obtained with the symmetrical letters test were compared to the Sloan Letter chart and the Landolt C tests for the same 36 subjects. To analyse whether there were differences between the measurements, the Shapiro–Wilk test of the variable was performed in the first place, resulting in acceptance of normal distribution (p > 0,05 for both comparisons). Therefore, the t-Student test was then executed, but it did not show any significant differences (p > 0,05) in VA levels between the different tests. The coefficients of agreement between the symmetrical letters test and the Sloan Letter Chart were: −0,16 [95%CI: −0,49;0,16] for the value of d, 0.71 for the ICC, and 0.16 and 0.14 as CV respectively. For the comparison between the symmetrical letters test and the C Landolt test, the coefficient of agreement results were: −0,34 [95%CI: −0,67; −0,006] for the value of d; 0.56 for the ICC and 0,16 and 0,14 as CV respectively. The Bland–Altman concordance analysis is attached as Fig. 4.

Figure 4.

The Bland–Altman plot for Symmetrical letters test versus the Sloan Letter Chart (top) and Symmetrical letters test versus the C Landolt test (down) in the preliminary study. (DecVA: decimal visual acuity).

(0.26MB).

Percentile curves have been obtained by analysing the data collected from the sample for distant vision. In order to show one single monocular VA percentile curve, we analysed during the pilot test whether the VA values were interchangeable. As the ICC obtained was low, the same agreement tests were repeated for the total sample.

The t-Student test applied after the verification of normality did not show any significant differences (−0.02 ± 0.21; 95% reliability [−0.04: −0.01], p > 0.05). Effect size (Cohen’s d), ICC and the coefficient of variation were, for the total sample: −0.21 [95%CI: −0.27; −0.13] for the value of d; 0.89 for the ICC and 0.34 for both eyes as CV. The Bland–Altman analysis of agreement is shown in Fig. 5. An ICC value of 0.89 is deemed sufficient to consider that the results obtained for RE and LE are interchangeable.

Figure 5.

The Bland Altman plot for right eye (RE) versus left eye (LE) for the full sample. (DecVA: Decimal Visual Acuity).

(0.15MB).

Then, an ANOVA was made to verify the influence of age. The interaction eye by participants’ age was not significant (p = 0.142), and consequently we understand that there are no significant differences in the VA of both eyes for the ages analysed.

Thus, a single monocular VA percentile curve was calculated, corresponding to the average values obtained for both eyes.

Fig. 6 shows the box diagram for the monocular DecVA calculated at 3-month intervals, and Fig. 7 shows the monocular DecVA percentile curves with their numerical values.

Figure 6.

Box diagram for the decimal VA corresponding to the average of both eyes from 3 to 12 years, calculated at 3-month intervals, for the full sample. The box represents 50% of the data and the limits of the boxes correspond to Q1 and Q3, quartiles. Whiskers represent the data between quartiles and 1.5 times the interquartile range (IQR). (DecVA: Decimal Visual Acuity).

(0.13MB).

Figure 7.

The figure shows the smoothed percentile curves of monocular visual acuity for the average of both eyes, expressed on a decimal scale with their numerical values. In black and in bold, the curve corresponding to the 50th percentile.

(0.82MB).

Fig. 8 shows the box diagram for the binocular DecVA at 3-month intervals, and Fig. 9 shows the binocular DecVA percentile curves with their numeric values.

Figure 8.

Box diagram for binocular decimal VA from 3 to 12 years, calculated at 3-month intervals, for the full sample. The box represents 50% of the data and the limits of the boxes correspond to Q1 and Q3, quartiles. Whiskers represent the data between quartiles and 1.5 times the interquartile range (IQR). (DecVA: Decimal Visual Acuity).

(0.36MB).

Figure 9.

The figure shows the smoothed percentile curves of binocular visual acuity, expressed on a decimal scale with their numerical values. In black and in bold, the curve corresponding to the 50th percentile.

(0.85MB).

Percentile curves show that DecVA increases from the age of 3 years onwards, reaching a value of 1 for the 50th percentile at the age of 5 years and 3 months in monocular and slightly before that age (5 years) in binocular. The 50th percentile reaches a DecVA of 1.20 from the age of 7 years and 3 months in monocular, and before that age (6 years and 9 months) in binocular. From these ages onwards, DecVA undergoes no significant changes until the age of 12 years in monocular, and slightly increases to values around 1.5 in binocular.