Visual function tests assess the patient's ability to perceive and integrate light stimuli into visual information,1 such as visual acuity (VA), contrast sensitivity, motion, colour perception, or peripheral vision tests.2 VA is the most used visual function test and is defined as “the ability to discriminate two stimuli separated in space at high contrast relative to background and is represented as the reciprocal of the minimal angle of resolution (MAR) at a given distance”.2

This measurement is very useful to accurately obtain the maximum visual resolution capacity, to adapt an optical correction if necessary or to evaluate the progression of ophthalmic pathologies over time, as in age-related macular degeneration (AMD), cataract, or amblyopia.3

VA is clinically measured by eye charts or optotypes which are presented to patients and are designed according to standard international guidelines, the most common being the one promulgated by the Committee on Vision of the National Academy of Sciences National Research Council (NAS-NRC),4,5 widely followed in the USA and Europe.

Although the standardized printed VA test chart for the Early Treatment of Diabetic Retinopathy Study (ETDRS) is still the gold-standard test, widely accepted for logMAR VA measurement and used in most of the clinical studies,6,7 computerised clinical VA measuring systems have emerged in recent years due to the growth of technology and digital tools, including Snellen, directional, figure or shape charts.3,8 Computerized VA charts could potentially offer versatility and precision, allowing customised settings, accurate measurements and continuous updates, improving the clinical assessment of VA. However, as the consequences of incorrect recording of visual function and its use to inform clinical management are serious, digital VA must be validated before it is adopted as part of clinical practice.9

The requirements that these computerized VA charts must accomplish are also documented in the NAS-NRC guidelines and include distance of use, lighting, luminance, size, angle, contrast, calibration, accuracy, repeatability and reliability. It is also difficult to find scientific evidence of the comparison of commercially available VA charts, so the reliability of clinical outcomes with the use of different types of computerized eye charts may be questioned.

Optonet Vision Unit (OVU) is a digital web-based eye care software developed at Optonet Ltd© that can be accessed via operating systems such as Android, iOS and Fire OS or through its own app in the Apple App Store, Google Play or Amazon Appstore,10 registered with the British Medicines and Healthcare products Regulatory Agency under reference number 5943.10 The OVU tool offers visual function measurement tests, including different AV charts oriented for all types of patients. This tool has already been used in a previous study to characterize distance VA in children (3–12 years old) using three types of OVU charts (Sloan and Symmetrical letters and Landolt C) with the main objective of obtaining normal values in percentile curves in this paediatric population and showing a promising repeatability.11 However, this software has more VA charts that are currently useful in ophthalmic consultations, but their reliability and repeatability has not been demonstrated.

The main objective of this study is to conduct a multicenter study across 28 optometric clinics throughout Spain, using a Continuing Professional Development (CPD) program to assess the repeatability of the OVU logMAR VA digital charts. CPD frameworks are increasingly used as research platforms in healthcare, as they allow for the integration of real-world clinical scenarios, active participation of professionals, and systematic data collection during educational activities.12–15 This approach aligns with current trends in health professions education, which emphasize practice-based learning and data-informed reflection. In countries such as the United Kingdom (UK)16 or United States (US),17 the use of CPD has expanded to several health professionals to generate practice-based evidence across various health disciplines. Thus, the use of a CPD-based framework to test the repeatability of VA measurements in clinical settings is a useful tool and offers research value and may support wider adoption of OVU charts in routine optometric care.

This is a cross-sectional study carried out at the University of Murcia (Spain) in collaboration with 28 optometrists from 28 different clinics distributed across Spain, recruited from a CPD program via a Moodle online campus during 2022. The research was approved by the Ethics Committee of the University of Murcia and carried out in accordance with the principles of the Declaration of Helsinki.

OVU chart design

Distance computerized OVU charts were designed according to Bailey and Lovie´s principles, i.e. ETDRS formatting18 collected in NAS-NRC5: backlit charts of contrast equal to 1 (following the NAS-NRC recommendation of at least 0.85 contrast level). All VA charts were displayed on a digital screen showing the set of logarithmic scale lines of 5 letters each line to simulate standard ETDRS graphs.2,6 OVU is classified as a Class I device with CE marking. This certification ensures that the tool complies with internationally accepted quality and performance standards and confirms compliance with applicable regulatory requirements and international standards, including ISO 8596 (which specifies requirements for the design, dimensions, contrast and presentation of optotypes used in VA test) and ISO 5725 (which establishes the principles for repeatability and reproducibility studies).

Following this formatting and the NAS-NRC regulation, five VA charts with different symbol types were used: Sloan letters with 10 letters (5 × 5 format) from ETDRS chart (C, D, H, K, N, O, R, S, V, Z) with a 10 % random hit rate19 (Fig. 1A); British letters (5 × 4 format) with 10 letter options (D, E, F, H, N, P, R, U, V, Z) with a 10 % random hit rate20 (Fig. 1B); Numbers (rounded) with 4 number options (5 ,6, 8, 9)21 with a 25 % random hit rate (Fig. 1C); Landolt C as a directional chart with 5 orientations for the aperture (Fig. 1D); and the Broken Ring chart that uses Landolt rings with contour interaction (inspired on the Broken Wheel test)22 (Fig. 1E). To maintain the same criteria as the previous tests, each size range is made up of five consecutive presentations (0.02 log unit per presentation) (Fig. 1E).

Fig. 1.

OVU chart design: Sloan letters (A), British letters (B), Numbers (C), Landolt C (D) and Broken Ring (E). Symbols marked by the red line in each chart as well as symbols from Broken Ring test (E) are equivalent to 0.30 logMAR VA (20/40 Snellen equivalent).

The results of this study demonstrate the repeatability of five digital VA charts under real-world conditions, with a minimal bias (lower than 0.0108 logMAR in all VA charts) and an excellent correlation between TCM (ICC higher than 0.968, 95 % LOA of ±0.071 (±3.5 letter difference)) and between TRT (two measurements at different time points) (ICC higher than 0.957, 95 % LOA of ±0.09 (±4.5 letter difference)). Therefore, VA charts from OVU are, to date, one of the most accurate and reliable digital VA charts for distance VA measurement.26,27 Their performance has been shown to be clinically consistent regardless of the examiner at the optometric centre, or other variables such as age range or type of ametropia correction. This reinforces the robustness of the digital charts across different clinical profiles and further supports their usefulness in routine optometric care, where patient characteristics can vary widely.

Furthermore, to the best of our knowledge, this is the first study to examine the repeatability of VA charts through a multicentre design using CPD. Previous studies in optometry have applied CPD programmes to investigate subjective refraction methods and prescription used in the UK,14 and to study the knowledge of Australian optometrists about ocular stem cell therapies.13 These two recent publications together with our findings, support the use of CPD as an innovative multicentric data collection method that facilitates the generation of robust, real-world evidence. This strategy also aims to converge clinical and research activities, reinforcing the connection between routine clinical practice and scientific research in order to improve the clinical relevance and efficiency of research, while promoting wider collaborative studies and advancing methodological standards in future clinical research in optometry.

Despite a global consensus on the methodology, units of measurement, and clinical significance of VA assessment, logarithmic optotypes and the logMAR annotation system are not yet universally adopted in clinical practice due to implementation complexity, costs, and unfamiliarity compared to Snellen or decimal systems. Consequently, the ETDRS printed charts remains the gold standard for VA testing worldwide.28

The reliability and repeatability of both printed and digital AV charts have been the subject of several studies. In the case of printed VA charts, the limits of agreement (LOA) for repeatability coefficients range from 0.07 to 0.2 logMAR,28–35 with the best results reported with differences of ±0.07 (±3.5 letters),29,30 ± 0.08 (±4 letters),31 ± 0.08 (±4 letters)31 and ±0.09 (±4.5 letters).32 Thus, our data, based on a large number of participants, yield robust TRT repeatability results that are similar or even better than those reported in previous studies.28–33,35

For digital VA charts, our results using UVO showed a high repeatability of TRT for the five VA tests analyzed, with a range of mean differences very close to 0.00 and an agreement of the measurements obtained with very good LOA repeatability coefficient (minimum differences of ±0.072 (±3.6 letters) for Sloan letters and maximum differences of ±0.09 (±4.5 letters) for Landolt C and Broken Rings). These robust results for the five tests evaluated may be since the system records (TCM) for each VA measurement. In this respect, in this work we have tested the repeatability of these TCM, obtaining very reliable results (ranging from ±0.054 (±2.7 letters) for Sloan and British charts to ±0.071 (±3.5) for Numbers chart), which therefore reduces the learning effect that the participants may experience due to the randomness of the letters/symbols.

Compared to other digital applications, our results show slightly lower TRT values than those provided by the K-VA test application, which shows a repeatability coefficient of ± 0.039, inferring a deviation of ±1.9 letters26; and similar to previous studies with a 95 % LOA of ±0.09 logMAR using the EyeChart App36 and ±0.08 log using Landolt C test from VisionApp.8

However, other digital applications validate their measurements based on the mean difference between measurements, not on the LOA data to calculate the repeatability coefficient, which gives a better assessment of the agreement between measurements. In this respect, OVU shows better values than those provided by Bellsmith et al. (2022) which shows a 95 % LOA of ±0.38 (±19 letters) for the mobile phone application37 and ± 0.40 (±20 letters) for the website test37; or the Peek palpation to assess VA in near vision, with ± 0.188–0.205 (9–10 letters),38 or compared to the printed ETDRS test with ±0.361 (18 letters).39

Then, which VA chart in the OVU software is more appropriate for clinical distance VA measurements? Regarding the VA charts with letters, the results show higher VA values compared to numbers or directional symbols (Table 2). However, pairwise comparison shows a statistically significance between them, with the VA obtained with the Sloan letter chart achieving the highest value. This difference may be due to the letter design (5 × 5 size for Sloan letters and 4 × 5 size for the British ones).

The fact that letter VA charts require a certain baseline linguistic competence40 makes the number and directional VA charts needed in clinical consultation to broaden the spectrum of patients for whom a reliable VA measurement can be taken. In this case, the difference between the letter charts and the directional charts are larger and again statistically significant (p < 0.001; mean difference ranges from −0.025 ± 0.002 (95 % CI, −0.030 to −0.042) for British vs. Landolt C to −0.047 ± 0.003 (95 % CI, −0.055 to −0.040) for Sloan vs. Broken Ring) (Table 4) which, due to the large sample size used in this study (N = 962 eyes), is also not clinically relevant.

The directional charts with a random hit ratio of 20 % (Landolt C and Broken Ring), designed for uncooperative patients or those lacking verbal competence, showed similar response accuracy to the previously described charts (Tables 2 and 4) and to other published data.11 The study by Navas-Navia et al. reported normative values of distance VA in children using Sloan and Symmetrical letters, and Landolt C charts from OVU describing percentiles of monocular and binocular VA from 3 to 12 years of age, indicating also the feasibility of OVU for VA measurement in all ages.11

Finally, the pairwise comparison study showed that the OVU VA values collected with the Number chart were significantly worse than that obtained with all other charts (maximum mean difference between Number chart and the Sloan chart (−0.071 ± 0.001 logMAR (95 % CI, −0.075 to −0.067)) (Table 4). This difference may be clinically relevant, as it represents almost a line of variation. It should be considered when interpreting results in clinical settings, especially when monitoring changes or making decisions based on small VA. These numerical charts are not very common in clinical practice, although are usually recommended for assessing VA in patients with verbal awareness but with letter recognition difficulties.

One limitation of these previous studies that may have affected the standardization of computerized VA testing to date is the inclusion of only one8,39 or few chart types,11 failing to compile a battery of validated VA charts for all kind of requirements.3 One limitation of this study is the absence of a direct comparison with the gold-standard ETDRS chart. As a result, the external validity of the VA measurement obtained with this digital application remains unknown. However, one of the strengths of this study is the inclusion of five representative types of distance VA charts (letters, numbers and directional symbols) from the same computerized tool, allowing clinicians and researchers to choose which VA chart is more appropriate for each patient and being able to compare the results, significantly enhancing clinical applicability. In addition, the collection of results through a CPD framework facilitated the participation of practising optometrists from a variety of real-world clinical settings. This approach reinforces the consistency of the results and enhances the applicability of these digital tests in everyday clinical practice.

The advantages of using computerized VA charts include providing controlled measurements3 and avoiding the disadvantages associated with the deterioration of traditional ETDRS charts over time, such as changing contrast levels that can lead to underestimations of VA.41 Additionally, symbol randomization prevents patient learning during measurements and facilitates updates of VA tests for different patient types.42 Digital systems, as OVU system offer high image quality, achieved through 4 K displays, enhanced brightness, and uniform contrast.43 These systems are also more cost-effective, enable symbol randomization to prevent memorization, and can be calibrated for various distances and contrast levels.

Finally in this study, the sample of participants was large and recruited from twenty-eight different clinical centres (n = 481 participants). Only VA worse than 0.8 logMAR was an exclusion criterion because of the difficulty of performing VA testing in these patients.41

Of the remaining participants, few had pathological conditions that resulted in low VA with their optical correction (Table 1), and the mean average of VA was not significantly lower compared to non-pathological eyes (0.1205 ± 0.2156 logMAR vs. −0.0815 ± 0.1031 logMAR, respectively). We have not found any difference in the repeatability of VA measurement between participants with lower or higher VA, indicating that the reliability of OVU is also precise for the lower VA analyzed. Nevertheless, the exclusion of participants with VA worse than 0.8 logMAR may limit the generalisability of the findings in terms of standardisation. As the current results apply only to individuals with relatively good visual acuity, further studies are needed to confirm the repeatability of VA measurements in populations with visual impairment or advanced stages of retinal diseases.

In summary, our results highlight the potential for expanding the use of the OVU system for VA monitoring in patients with ocular pathology and for clinical studies. In addition, leveraging online CPD programmes12 to foster community-based collaborative research can improve statistical and research understanding. This innovative and practical approach helps bridge the gap between clinical practice and research.