Effect of light barrier and type of refractive error on the performance of a portable wavefront aberrometer-based auto refractometer

Abstract

Purpose:

Accurate and accessible autorefractors are essential for refractive error assessment, particularly in resource-limited settings. This study evaluates the accuracy and reliability of Instaref R20 with a light barrier (Instaref R20 LB) compared to a tabletop autorefractor (AR) and subjective refraction (SR) in an adult population. A light barrier (wearable polaroid filter) was used to reduce ambient illumination and stabilize pupil size, thereby improving measurement reliability under bright conditions.

Methods:

A prospective study was conducted at a tertiary eye institute in Mumbai, enrolling 200 adults (mean age: 44.33 ± 13.97 years). All the participants underwent a comprehensive eye examination including objective refraction, subjective acceptance, and anterior and posterior segment evaluation. Non-cycloplegic refraction was assessed using Instaref R20, Instaref R20 LB, AR, and SR. Agreement between methods was analyzed using Bland–Altman plots, assessing power vector components (M, J0, J45). Statistical comparisons were performed using the Wilcoxon signed-rank test.

Results:

Instaref R20 LB demonstrated strong agreement with SR (M: 0.10D, P = 0.13; J0: -0.05D, P = 0.05; J45: 0.01D, P = 0.44), while Instaref R20 showed slight variance (M: 0.17D, P < 0.001). The agreement within ± 0.50D of SR was 64.64% (M), 85.35% (J0), and 89.39% (J45). Under bright illumination, Instaref R20 failed to capture readings in six cases, whereas Instaref R20 LB obtained stable measurements in all participants, demonstrating superior reliability in bright environments. Subgroup analysis showed higher accuracy in myopic eyes, while AR exhibited significant deviations in hyperopes.

Conclusion:

The Instaref R20 LB improves measurement accuracy and reliability, particularly under bright illumination. Its strong agreement with SR and portable design make it a viable tool for refractive error assessment in resource-constrained settings.

In today’s global health landscape, innovative and cost-effective solutions to address refractive errors are critical, particularly in low-resource settings. Uncorrected refractive errors (UREs) currently affect an alarming 153 million people worldwide,[1,2] and in India alone, approximately 10.2% of adults are estimated to be affected by URE.[3] These statistics underscore the urgent need for a flexible and scalable vision screening strategy to diagnose and promptly treat URE effectively. Innovative solutions, such as mobile eye clinics and telemedicine programs, have emerged to address the issue by bringing vision care directly to underserved populations. However, the lack of trained healthcare professionals to operate these programs often limits their accessibility and widespread implementation.[4,5]

Autorefractors have traditionally been employed to optimize operational efficiency and reduce chair time in clinical setups.[6,7,8] However, these programs are often limited in remote areas due to the bulky, nonportable, and expensive equipment traditionally used for eye screening. To reduce the need for external help and lower the costs of conventional screening methods, there is a growing need for sustainable, accessible solutions. This has led to the development of affordable, easy-to-use handheld autorefractometers, which provide accurate refraction measurements and can reach diverse and remote populations more effectively.[9,10,11,12]

One such portable handheld autorefractometer based on Shack–Hartmann aberrometer technology is the Instaref R20 (Remidio Innovative Solution Pvt. Ltd). Its lightweight, portable, and user-friendly design makes it highly convenient, allowing nonspecialist healthcare practitioners and even layperson screeners to easily operate it for vision assessments. It can be used in large-scale studies involving refraction, especially those conducted outside the clinical or laboratory setting. Furthermore, studies have revealed it to be as reliable as a table-mounted autorefractor and subjective refractive testing for both pediatric[13] and geriatric populations.[14]

A limitation identified with the Instaref R20 was its inability to freeze readings under bright illumination, especially in the presence of small pupil sizes. To address this challenge, a novel solution was developed: wearable linear polaroid brown filter glasses with 15% light transmissibility, referred to as “light barrier glasses.” These glasses effectively control light exposure, allowing for more accurate measurements in challenging lighting conditions [Fig. 1]. This innovative approach allows for the acquisition of precise measurements even in situations with high levels of ambient light. This study aims to evaluate the performance and utility of the Instaref R20 with a light barrier abbreviated as (Instaref R20 LB) in comparison to a tabletop autorefractor (AR) (URK-800, Biomedix) and subjective refraction (SR). Additionally, it seeks to determine whether the Instaref R20 LB performs comparably to the Instaref R20 without light barrier under controlled room illumination.

Figure 1.

The left side of the image shows the use of the InstaRef R20 device with light barrier glasses, while the right side features the Polaroid light barrier glasses with an adjustable headband for a secure fit

Methods

A prospective study was conducted at a tertiary eye institute in Mumbai to evaluate the functionality of the newly developed light barrier accessory used with the Instaref R20. The study compared the performance of the Instaref R20 LB against the Instaref R20 without any additional accessory and the conventional tabletop auto refractometer. Additionally, the subjective acceptance of the participants was measured to assess the overall effectiveness and accuracy of the light barrier accessory. The study was approved by the Institutional Ethics Committee (SSF/IRB/0010/2025) and adhered to the tenets of the declarations of Helsinki. Written informed consent was obtained from all the adults who participated in the study. The study included subjects older than 18 years with a best corrected visual acuity (BCVA) of 20/20 in each eye and refractive errors between Sph ± 10D and Cyl ± 5D. Individuals with significant media opacity (e.g., corneal opacity, advanced cataract, vitreous hemorrhage), posterior segment pathology affecting visual acuity, amblyopia, recent eye surgery within 2 weeks, an active eye infection or inflammation, abnormal pupillary responses, history of refractive surgery, cataract surgery, corneal pathology such as keratoconus or ectasia, or any other procedure affecting corneal integrity were excluded from the study. Additionally, those who were on medication interfering with visual acuity, unstable medical conditions, and pregnant women were also not enrolled and all participants were confirmed to have no prior ocular surgeries or corneal disease.

The study device, Instaref R20, incorporates a Shack–Hartmann wavefront aberrometer that samples the retinal reflex using a microlenslet array (12.92 × 8.75 mm, 1.55 mm thickness) across the optical zone of 5 mm with ~400 measurements points. It functions with pupils ≥2 mm and is supported by pupil-centering and tilt-warning mechanisms to ensure precise alignment.

All participants underwent refractive error measurements in both eyes followed by a thorough ophthalmic examination of the anterior and posterior segments. The study protocol required participants to undergo four rounds of refraction measurements. Initially, an objective assessment of refractive error for both eyes was conducted using the tabletop auto refractometer (Biomedix Auto refractometer - Model: URK-800). Subsequently, SR was performed by an experienced optometrist (Examiner I) to determine the BCVA, with the starting point of SR being the auto refractometer.

Following this, a second optometrist, trained specifically in operating handheld autorefractors, performed the measurements (Examiner II) and employed the handheld study devices Instaref R20 and Instaref R20 LB in a randomized order. The device was held at the eye level, ensuring the crosshair on the screen aligned with the pupil. Each eye was tested using a 6/60 Snellen chart placed 6 meters away. Three averaged readings were obtained per eye, ensuring proper distance fixation to minimize accommodation. The same process was repeated for the left eye. Subsequently, for the Instaref R20 LB, participants were directed to wear light-blocking glasses featuring a brown polaroid over the fixating eye by Examiner II. This setup was designed to limit light entry, fostering a naturally dim atmosphere that encouraged pupil dilation, crucial for precise refractive error measurement. All measurements were obtained using a similar alignment protocol as with the Instaref R20, and the average of three readings was considered for analysis.

Each measurement was consistently executed in an environment where the lighting was bright and uniform, preserving a consistent level of illumination (400–650 lux) throughout the entire process. A 5-minute interval was allotted between each test to mitigate the effects of accommodation (if any). All the examiners were blinded from each other’s observations, eliminating any potential influence or bias.

All the refraction measurements were transformed using vector decomposition to spherical and cylindrical scale values using the below formulae. The spherical equivalent was calculated as sphere value + cylindrical value/2. According to vector principles, astigmatism has two orthogonal vectors known as J0 and J45. J0 is the horizontal and vertical component of astigmatism, and J45 is the oblique component of astigmatism. The following equations were used to calculate the vectors:

J0= (−C/2) cos θ (2a), J45= (−C/2) sin θ (2a) where C = cylinder power, a = axis of astigmatism

The agreement between the study device, SR, and tabletop autorefractor (URK-800, Biomedix) measurements was evaluated using a Bland–Altman analysis with 95% limits of agreement (LOA) on each power vector component (M, J0, and J45). The bias between the methods was compared. Agreement within thresholds of 0.5 D for M, J0, and J45 was evaluated. A mean difference of ≤ 0.50 diopters between methods was considered clinically acceptable for refractive error measurements based on previous studies. The light barrier likely enhances measurements by reducing ambient illumination, minimizing pupil constriction and glare, and thereby stabilizing pupil size, although this was not formally measured. Differences between devices in different groups were evaluated using the Wilcoxon signed rank test. Statistical analysis was conducted using JASP 0.18.1.0, while data visualization was achieved through Microsoft Excel.

Results

A total of 200 adult subjects with an average age of 44.33 ± 13.97 years (range: 18 to 68 years) were included. Gender distribution revealed that 44.44% were females (n = 88, 45.48 ± 11.21 years) and 55.55% were males (n = 110, 48.38 ± 11.54 years). Of the 200 participants, the Instaref R20 successfully captured refractive measurements in 194 subjects, whereas the Instaref R20 LB obtained readings in all 200 subjects. The tabletop autorefractometer also successfully captured readings in all participants.

Table 1 describes the mean and standard deviation (Mean ± SD in D) of M, J0, and J45 derived from refractive error measurements obtained by different methods. The corresponding median and interquartile range (IQR) values are presented in Supplementary Table 1. Statistical analysis indicated no significant discrepancies between right and left eye measurements for sphere, M, J0, and J45 along with strong inter-eye correlations (r = 0.91). Therefore, analysis was performed on the right eye readings only.

Table 1.

Mean±SD values of refractive outcomes measured using different devices

Device	M (in D) (Mean±SD)	J0 (in D) (Mean±SD)	J45 (in D) (Mean±SD)
Instaref R20	-0.46±2.33	-0.13±0.62	-0.03±0.42
Instaref R20 LB	-0.54±2.51	-0.12±0.57	0.04±0.39
AR	-0.68±2.52	-0.12±0.46	0.05±0.28
Subjective Refraction	-0.64±2.23	-0.06±0.33	0.03±0.18

SD, Standard Deviation; D, Dioptre; M, Spherical Equivalent; J0 and J45 - Cylindrical components

Supplementary Table 1.

Median±IQR values of refractive outcomes measured using different devices

Device		M (in D) (Median±IQR)		J0 (in D) (Median±IQR)		J45 (in D) (Median±IQR)
Instaref R20		0.13±1.91		-0.09±0.72		0.00±0.36
Instaref R20 LB		0.13±2.13		-0.08±0.58		0.00±0.33
AR		-0.06±3.03		-0.12±0.49		0.00±0.25
Subjective Refraction		0.00±1.84		0.00±0.23		0.00±0.00

IQR, Interquartile Range; D, Dioptre; M, Spherical Equivalent; J0 and J45 - Cylindrical components

For M values: The mean values for M, J0, and J45 of the Instaref R20 LB, Instaref R20, AR, and SR are presented in Table 1. On individual comparison with SR values, the mean difference based on Bland and Altman analysis for M was 0.17D (95% CI: -1.29 to 1.65) for Instaref R20, 0.10D (95% CI: -1.74 to 1.94) for Instaref R20 LB, and -0.03D (95% CI: -1.55 to 1.47) for AR as shown in Fig. 2. Notably, statistically significant differences were observed between Instaref R20 and SR (mean difference = 0.17D, P < 0.001), and between Instaref R20 and AR (mean difference = 0.21D, P < 0.001). However, no significant differences were found between Instaref R20 and Instaref R20 LB (mean difference = -0.07D, P = 0.20), Instaref R20 LB and SR (mean difference = 0.10D, P = 0.13), or Instaref R20 LB and AR (mean difference = -0.13D, P = 0.02). In the assessment of M across various devices compared to subjective refraction, it was observed that InstaRef and Instaref R20 LB exhibited concordance rates of approximately 64% within a 0.50D margin of agreement, with approximately 34% within a 0.25D range. Conversely, the AR demonstrated higher levels of agreement, with 71.17% of M values falling within the 0.50D range and 49.41% within the 0.25D range.

Figure 2.

Bland–Altman plots showing bias and 95% limit of agreement comparing InstaRef R20, Instaref R20 LB, and AR with SR for M values (a, d and g), J0 (b, e and h), and J45 (c, f and i)

For J0 and J45: Analysis of J0 values across the different devices demonstrated varying degrees of agreement. The mean J0 values for Instaref R20, Instaref R20 LB, AR, and SR were [-0.13D], [-0.12D], [-0.12D], and [-0.06], respectively. Statistically significant differences were observed between Instaref R20 and SR (mean difference = -0.07, P = 0.01), and between Instaref R20 and AR (mean difference = -0.01, P = 0.91). However, no significant differences were found between Instaref R20 and Instaref R20 LB (mean difference = -0.01, P = 0.24), Instaref R20 LB and SR (mean difference = -0.05, P = 0.05), or Instaref R20 LB and AR (mean difference = 0.00, P = 0.35).

When comparing J45 values, a statistically significant difference was found between Instaref R20 and SR (mean difference = -0.06, P = 0.04), while insignificant differences were observed between Instaref R20 and Instaref R20 LB (mean difference = -0.07, P = 0.02), Instaref R20 LB and SR (mean difference = 0.01, P = 0.44), or Instaref R20 LB and AR (mean difference = 0.00, P = 0.68). Additionally, in comparisons with SR, over 76% of subjects for Instaref R20, 85.35% for Instaref R20 LB, and 87% for AR exhibited disparities less than 0.50D for J0 and AR resulted in 87.05% of J45 values within a 0.25D range, while both Instaref R20 and Instaref R20 LB showed a lower percentage, with only 62% of J45 values falling within the 0.25D range [Table 2].

Table 2.

Agreement between different refractive error measurement devices represented using mean difference (95% CI), percentage of subjects with 0.25D and 0.50D difference and P values of refractive outcomes measured using different devices

Device		Metric		Mean Difference with 95% CI (D)		Percentage of Subjects with 0.25 D difference		Percentage of Subjects with 0.50 D difference		P
Instaref R20 Vs Subjective		M		0.17 (-1.29 to 1.65)		33.33		64.14		<0.001
		J0		-0.07 (-1.04 to 0.89)		56.06		77.77		0.01
		J45		-0.06 (-0.83 to 0.71)		61.11		82.32		0.04
Instaref R20 LB Vs Subjective		M		0.10 (-1.74 to 1.94)		34.34		64.64		0.13
		J0		-0.05 (-0.87 to 0.75)		61.61		85.35		0.05
		J45		0.01 (-0.64 to 0.68)		71.21		89.39		0.44
Instaref R20 Vs AR		M		0.21 (-1.31 to 1.74)		36.36		55.05		<0.001
		J0		-0.01 (-0.86 to 0.84)		52.02		76.76		0.91
		J45		-0.08 (-0.87 to 0.69)		61.61		79.29		0.03
Instaref R20 LB Vs AR		M		-0.13 (-1.84 to 1.57)		43.43		58.58		0.02
		J0		0.00 (-0.66 to 0.67)		60.10		83.33		0.35
		J45		0.00 (-0.64 to 0.62)		68.18		86.86		0.68
Instaref R20 Vs Instaref R20 LB		M		-0.07 (-1.51 to 1.35)		53.03		70.70		0.20
		J0		-0.01 (-0.07-0.04)		59.09		82.82		0.29
		J45		-0.07 (-0.13--0.01)		62.62		78.78		0.02

In a subgroup analysis based on refractive error, the myopic group (defined as Myopia < −0.50D, n = 66) and the hyperopic group (defined as Hyperopia > +0.50D, n = 73) were compared. It was observed that in the myopic group, the Instaref R20 LB displayed negligible differences across all metrics (M, J0, J45) with no significant deviation from subjective measurements. However, the AR showed a significant mean difference in M (-0.28, P < 0.001). In the hyperopic group, both Instaref R20 and Instaref R20 LB showed small, clinically insignificant differences in M and J0, while the AR exhibited significant discrepancies in M (0.20, P < 0.001) and J0 (-0.12, P < 0.001). Detailed tables [Tables 3 and 4] are provided in supplementary sheets.

Table 3.

Agreement between refractive outcomes measured of Instaref R20, Instaref R20 LB, and AR with subjective refraction in the myopic group (n=66)

Device		Metric		Mean Difference with 95% CI		P
Instaref R20 Vs Subjective		M		0.19 (-1.54 to 1.92)		0.085
		J0		-0.07 (-1.05 to 0.91)		0.24
		J45		-0.09 (-0.71 to 0.52)		0.019
Instaref R20 LB Vs Subjective		M		0.02 (-2.21 to 2.24)		0.92
		J0		0.06 (-0.69 to 0.82)		0.17
		J45		0.04 (-0.75 to 0.83)		0.98
AR Vs Subjective		M		-0.28 (-1.73 to 1.17)		<0.001
		J0		-0.01 (-0.49 to 0.47)		0.83
		J45		-0.01 (-0.37 to 0.35)		0.67

Table 4.

Agreement between refractive outcomes measured of Instaref R20, Instaref R20 LB, and AR with subjective refraction in the myopic group (n=73)

Device		Metric		Mean Difference with 95% CI		P
Instaref R20 Vs Subjective		M		0.12 (-0.86 to 1.10)		0.01
		J0		-0.07 (-0.85 to 0.70)		0.03
		J45		-0.04 (-0.80 to 0.71)		0.32
Instaref R20 LB Vs Subjective		M		0.07 (-1.08 tO 1.23)		0.17
		J0		-0.16 (-1.04 to 0.70)		<0.001
		J45		0.02 (-0.58 to 0.63)		0.36
AR Vs Subjective		M		0.20 (-0.73 to 1.14)		<0.001
		J0		-0.12 (-0.55 to –0.30)		<0.001
		J45		0.03 (-0.31 to 0.37)		0.11

Supplementary sheet table:

For Myopia Group:

For Hyperopic Group:

When comparing the reading acquisition capabilities, the Instaref R20 device was unable to capture readings in six individuals under bright lighting conditions. In contrast, the Instaref R20 LB successfully obtained stable readings in all study participants, indicating superior reliability in high-illumination environments.

Discussion

This prospective clinical study was initiated after observing that the Instaref R20 had difficulty obtaining stable readings under bright ambient lighting. The results show that adding a light barrier accessory significantly improved the device’s performance—in terms of both measurement accuracy for refractive error and its ability to capture readings in well-lit environments.

The primary mechanism underlying this improvement is the light barrier’s ability to reduce ambient light entering the eye, which in turn minimizes pupil constriction and glare. This likely leads to greater pupil stability and improved wavefront-based measurement quality. Although pupil diameter was not formally measured in this study, it is reasonable to infer that the reduced ambient illumination contributed to slightly larger pupil sizes, thereby facilitating easier and more reliable measurements. The proposed mechanism involving larger pupil size under the light barrier remains speculative because pupil diameter was not directly measured

The Instaref R20 LB showed a 3% (6 subjects) higher success rate in capturing measurements compared to the Instaref R20 alone. Both the Instaref R20 LB and the Auto Refractometer (AR) demonstrated strong agreement with subjective refraction (SR), with differences within clinically acceptable limits. While using the LB increased the successful capture rate under bright ambient lighting, reaching 100% in our study, practical factors may influence real-world application. In our study, all six subjects whom the standard Instaref R20 failed to measure were successfully measured with the light barrier, indicating that this accessory effectively reduces the impact of ambient light on pupil size and glare. However, variability in user handling, subject characteristics (such as very small pupils or ocular media clarity), and lighting conditions in diverse environments may still affect performance. Therefore, when deploying this solution in community-based settings, it is advisable to ensure minimal user training and standardized lighting protocols to maintain high capture success rates and measurement reliability.

A study conducted by Rao et al.[13,14] comparing a Shack–Hartmann-based handheld auto-refractometer (AR) with the gold standard subjective refraction (SR) and open-field AR found that the device showed agreement within 0.5D in the majority of the cohort—84% for M. In contrast, the present study found agreement of 64.64% for M with the Instaref R20 LB and 63.63% for M with the conventional AR. This discrepancy may be attributed to differences in the study populations’ age and variations in spherical equivalent between the two studies. Variations in study methodology could play a role, including differences in how subjective refraction (SR) was conducted, the specific protocols for autorefractor calibration, and the ambient lighting conditions during measurement, all of which can affect the accuracy and comparability of results.

Numerous prior investigations have demonstrated varying degrees of agreement between autorefractors. While SE and J0 measurements generally exhibit high concordance across devices, discrepancies often arise concerning J45, especially at oblique axes. These inconsistencies are attributed to differences in device algorithms, measurement techniques, and calibration processes, which are well documented in the literature.[15,16] In the current study, Instaref R20 demonstrated a mean difference of 0.17D, Instaref R20 LB - 0.10D, and the AR -0.03D, with statistically significant disparities observed between Instaref R20 and both SR and the AR. However, no significant differences were found between Instaref R20 LB and subjective refraction, indicating comparable performance. The spherical equivalent and cylindrical values of wavefront aberrometer-based handheld and tabletop ARs are validated against other techniques in previous studies.[17] Previous studies comparing traditional autorefractors and wavefront aberrometer found that there is a tendency for some wavefront aberrometers to over minus refractions owing to instrument myopia, but contrary results were found in this study. Both Instaref R20 and Instaref R20 LB underestimated myopia as compared to AR. This discrepancy in the findings could be due to the number of points measured through each wavefront and the AR varies. Although mean differences were small, the limits of agreement were wide (e.g., −1.74 to + 1.94 D for M), indicating substantial inter-subject variability.

The study had some limitations that warrant discussion. First, the starting point for SR was based on autorefraction readings. This approach may have inadvertently biased the results toward a stronger agreement between autorefraction and subjective refraction. Because subjective refraction was initiated using autorefraction readings, incorporation bias may have reduced the apparent differences between methods. We also acknowledge that subjective refraction performed using retinoscopy findings as the starting point could have minimized this bias; however, retinoscopy readings were not available for this study. Second, the study exclusively focused on adult participants, which may limit the generalizability of the findings to other age groups, particularly children and adolescents. Non-cycloplegic refraction may permit residual accommodation, particularly in younger adults and hyperopes, potentially influencing comparisons. Including a diverse range of age groups in future studies would offer a more comprehensive understanding of the devices’ performance across various demographic segments. Third, the study was unable to assess the performance of the Instaref R20 LB in comparison to the Instaref R20 under fully bright conditions, such as those encountered in eye screening camps, which may impact the applicability of the results in such settings.

Conclusion

The findings of the study suggest that the performance of the Instaref R20 LB surpassed that of the Instaref R20. The incorporation of polarized filter glasses not only expanded the capacity of the Instaref R20 to accurately capture readings in bright light conditions, making it suitable for use in outdoor screening camps, but also resulted in improved overall performance in refractive error assessment. Specifically, the Instaref R20 LB demonstrated refractive error measurements in the adult population that were clinically comparable to those obtained with both autorefraction and subjective refraction. In summary, with the addition of a light barrier, the Instaref R20 proves to be a promising solution for quick, portable, and cost-effective refractive error measurement. Its improved performance as a starting point for SR highlights its potential to simplify clinical procedures and enhance access to precise vision assessment and refraction, especially in settings with limited resources.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.