Prevalence and determinants of refractive errors at Sekororo District Hospital in Limpopo Province, South Africa

.1–23.0), hyperopia 8.1 % (95 % CI: 5.2–12.4) and astigmatism 26.9 % (95 % CI: 21.6–33.0). Conclusion: Across the samples (2018 and 2019) and to laterality (right and left eyes), myopia prevalence ranged from 10 % to 17.5 % while hyperopia ranged from 3.8 % to 8.1 % . Astigmatism was most prevalent (ranging from 26.9 % to 43.4 % ). Contribution: This article provides useful information about the prevalence of REs in the district hospital setting. The Department of Health Authority may use the results for policy decisions.

Common causative factors for URE include longer or shorter axial lengths rather than the average axial length of the eyeball, which is approximately 24.4 mm. 1,2,4,6An increased axial length (>24.4 mm) is often associated with an increased risk of myopia, while decreased axial length is associated with an increased risk of hyperopia.Other refractive elements of the eyeball such as abnormally strong or weak corneal refractive powers might be relevant, and asymmetric curvatures of the cornea and/or crystalline lens of the eye are associated with a greater risk for astigmatism. 4her causative factors for uncorrected refractive errors include genetic, environmental and ocular degenerative or inflammatory factors associated with diseases such as diabetes, keratoconus or glaucoma where glaucoma prevalence is associated with the refractive state, increasing gradually with increasing myopia. 1,2,3,4,5Genetic factors related to an affected child's family background with parents and/or grandparents with myopia may also produce myopia in the child. 6Childhood (or school) myopia is often at schoolgoing age (6 years and above, but before 18 years). 1nvironmental changes such as growing up in modern society, whereby people spend more time indoors and less time outdoors, are believed to be another causative factor in the increasing prevalence of myopia in children. 7,8In modern society, people spend extensive time viewing electronic devices such as computers, smartphones and others for recreational, educational and professional purposes. 7Again, decreased exposure to vitamin D or perhaps less outdoor exercise with more distance viewing as compared to closer viewing indoors may also lead to increased risk for myopia, which is more common in persons with vitamin D deficiency as opposed to those without such deficiency. 7,8Retinal degenerative or other changes with pathological myopia associated with ocular axial length elongation can sometimes lead to retinal damage including maculopathy and retinal detachment (RD). 2,4,6Such degenerative disease processes can progress from childhood to early adulthood, sometimes causing progressive myopia with VI that might not be fully amenable to compensation or correction.
Presbyopia is another common eye disorder with ageing that occurs because of physiological insufficiency in the amplitude of ocular accommodation causing blurred near vision when reading fine print or perhaps threading needles. 1,4Presbyopia relates to the ageing factors and sclerosis of the ocular crystalline lens. 1 In the sub-Saharan African (SSA) region, uncompensated presbyopia is believed to be the second leading cause of near VI, after cataracts as the leading cause (mainly because of insufficient healthcare resources leading to an inability to promptly perform the necessary surgeries). 3,9etropia or URE mainly occurs because of natural structural defects of the eyeball, and although they sometimes cannot be easily prevented, they can be diagnosed, measured and treated in clinical settings by healthcare professionals, such as optometrists, opticians, ophthalmologists and other trained healthcare professionals. 1,2,3Management of URE can be achieved through optical means using spectacles or contact lenses and surgical treatment. 1,2,3,4The prevalence of preventable distance and/or near VI because of URE and presbyopia differs from one country or region to another.In low-income countries, the prevalence of VI because of URE is four times more than in high-income countries. 3,9However, high-income countries still have a relatively high prevalence of VI because of the prevalence of refractive errors despite having ample resources and the latest medical technology to diagnose and treat ametropia. 3For example, in China and other Asian countries, prevalence of moderate-to-severe myopia is large with a high risk for possible complications such as maculopathy and RD leading to irreversible moderate or severe VI. 3,10 Comorbidities such as diabetes, hypertension and cardiovascular conditions also complicate refractive error and the risk for VI and this applies throughout the world. 3obally, as of 2022, at least 2.2 billion people are living with some form of VI or blindness, of which at least 1 bn are visually impaired because of URE and/or presbyopia.Of this 1 bn, URE affects at least 123 million people with 826 m presbyopes; thus, at least, 1 bn could be assisted with simple spectacles. 3The remaining 1.2 bn people are visually impaired or blind because of other causes, such as cataracts, glaucoma, corneal opacities, diabetic retinopathy and trachoma, and this is especially so in low-income countries. 3e primary purpose of this study was to investigate URE and VI among African participants visiting the optometry clinic in a rural environment in the Mopani District of the Limpopo Province, South Africa.As stated earlier, URE is a common disorder of the eye, leading to VI in the absence of correction or compensation.To the best knowledge of the researchers, this study is the first retrospective study of URE and VI conducted in a public hospital setting for patients consulted at a rural clinic over an extended period of 2 years from January 2018 to December 2019 based on history records archived for clinical purpose rather than this research.The samples were compared over 2 years to investigate similarities or differences per annum.

Method and materials
The data were collected retrospectively in the optometry clinic at Sekororo District Hospital in the Mopani District of Limpopo Province in South Africa.Historical records were extracted from the clinical archive for the patients who had consultations at the clinic between 01 January 2018 and 31 December 2019.(The year 2020 was excluded because of coronavirus disease 2019 [COVID-19].)This study was ethically approved (FREC-1170-2021) by the Faculty Research and Ethics Committee (FREC) in the Faculty of Health Sciences at the University of Johannesburg, South Africa.Permission to conduct the study at the selected district hospital was granted by the Provincial Health Research and Ethics Committee of the Limpopo Department of Health (South Africa) 11,12 and by the Chief Executive Officer (CEO) of Sekororo Hospital.
The records of the patients from the clinic were randomly selected using a probability-stratified random sampling method with the aid of a random function: f x = rand (num) of an Excel spreadsheet (Microsoft 365) for Windows 11 where num is the patient record number.The restriction and stratification techniques were applied to avoid bias and confounding effects by excluding the sampled records with incomplete information and then stratifying sampled records into two subgroups or strata (2018 and 2019) according to the year in which the patients consulted at the clinic.The statistical formula of Cochrane shown in Equation 1 was used to determine the required minimum sample size: whereby, n = the required or minimum sample size.P = the estimated proportion of refractive errors in a population.The degree of variability is not known.A proportion of 50% is assumed in this study for a maximum level of variability elected, suggesting that half of the sample would be emmetropic and the other half would be ametropic.
e = the margin of error, which is the risk the researcher is willing to accept, which could be because of factors such as missing or incomplete clinical records in the study.Z = the probability value at a significant level of 0.05: here Z = 1.96 obtained from the published probability table of z-scores 13 corresponding to the level of confidence required (here 95%).Therefore, the required minimum sample size for the study was 384 records.The stratified formula (sample size for the entire study divided by population size × stratum size) 14 was used to spread the required minimum sample size (384) into two strata (i.e.2018 and 2019).The required sample size for 2018 is 384/1140 × 706 = 238.For 2019, the required sample size is 384/1140 × 434 = 146.To further increase the statistical power, 200 additional records were added to each stratum; therefore, the records for 2018 increased from 238 to 438, while the records for the 2019 sample increased from 146 to 346.Furthermore, extracted clinical records with missing information needed to be excluded from the analysis, so 159 records for 2018 and 112 records for the 2019 sample were excluded from the study.Thus, the final sample sizes for analysis for the 2018 sample comprised 279 records and 234 records for the 2019 sample, or a grand total of 513 records over the period concerned.
For refractive error, the prevalence of myopia, hyperopia and presbyopia for the right and left eyes was determined using the spherical equivalent refractive (SER) powers of ≤ -0.25 diopter (D), ≥ 0.50 D and 1 D, respectively, but the prevalence of astigmatism was determined using the cylindrical equivalent refractive power of ≤ -0.25 D. 12 The determinants or associated risk factors for refractive errors were determined between the dependent variable (the refractive errors) and specific independent variables, i.e. age, and gender using univariate and multivariate logistic regression analysis reported as the odds ratio (OR) with 95% confidence intervals (CI).

Statistical analysis
The

Ethical considerations
An application for the full ethical clearance to conduct this study was obtained from the University of Johannesburg Faculty of Health Sciences Research Ethics Committee on 06 August 2021.The clearance number is REC-1170-2021.

Results
Table 1 shows demographic data of the clinical records extracted from the archive for patients who attended the optometry clinic.Table 2 shows the prevalence of myopia, hyperopia and astigmatism, as well as emmetropia in the 2018 and 2019 samples.Astigmatism was the most prevalent refractive error in both samples, and astigmatism was more in the right eyes than in the left eyes.The least prevalent refractive error for the right and left eyes in both samples was hyperopia, but the amount of hyperopia for the right and left eyes in the 2018 sample was similar.In contrast, hyperopia for the right and left eyes in the 2019 sample differed where the left eyes were more hyperopic than the right eyes.Emmetropia in both samples was more in the right eyes than in the left eyes.Table 4 shows that astigmatism was the most prevalent refractive error in the 2018 sample for females and males, and it was more in females than in males.However, more males were emmetropic as compared to females.For the 2019 sample, astigmatism remained the most prevalent refractive error in female and male patients, but males were more emmetropic than females.Moreover, the least prevalent refractive error was hyperopia in both samples for 2018 and 2019.The amount of hyperopia for males and females in both samples differed.
In    sample but not statistically significant.Gender and UPFS classifications are not a determinant of any refractive errors for the right and left eyes but are statistically significant (p = 0.001 for gender and p < 0.001 for UPFS).However, the multivariate logistic regression analysis results further revealed that age, gender and UPFS are not the determinants of refractive error for the right and left eyes.For the 2019 sample, univariate logistic regression analysis revealed that the age of a patient is a determinant of (any) refractive error for the right and left eyes and is statistically significant (OR = 1.00; 95% CI = 0.83-1.40;p = 0.020).Again, gender and UPFS are not a determinant of refractive errors for the right and left eyes, but the UPFS is statistically significant (p = 0.013).Multivariate logistic regression analysis in Table 6 revealed that age is a determinant of refractive errors for the right and left eyes, statistically significant (OR = 1.44; 95% CI = 0.23-2.65;p = 0.020).Gender and UPFS are not the determinants of refractive errors for the right and left eyes, but the UPFS is statistically significant (p = 0.019).

Discussion
This study is the first retrospective cross-sectional study based on the clinical records extracted from the archive of Sekororo District Hospital aimed at investigating the prevalence and determinants of refractive errors among all patients who consulted a rural optometry clinic over 2 years from January 2018 to December 2019.
The study comprised two stratified random samples (2018 and 2019).The results of the present study are comparable to those of previous studies from different parts of the world.
The overall prevalence of refractive errors for the right eyes and left eyes ranging from 50.8% to 63.8% for the 2018 and 2019 samples is not consistent with that of other previous studies, which could be because of different samples, settings and methodology used to gather the data.The prevalence of refractive errors for the right and left eyes of the present study in both samples is larger than that of other previous  related studies in Africa, where the overall prevalence ranges from 4.7% to 48.8% 15,16,17,18 and smaller than that of other studies conducted outside Africa with the prevalence range of 69.7% -78.0% because of the different settings. 19,20The prevalence of refractive errors for the right and left eyes of the present study in two samples is smaller than that of other previous studies in Africa, where the prevalence ranges from 96.2% to 97.5% 21,22 and larger than that of other previous studies in Africa with the prevalence range of 12.9% -97.5%. 23,24,25,26,27,28,29,30,31,32,33,34The difference in the prevalence of refractive errors for present and previous studies could be the result of a larger or smaller population, as well as the research design.
The prevalence of myopia in the right (10.0%) and left eyes (16.1%) of the present study with a range of 10.0% -17.5% in both samples did not necessarily agree with that of other previous studies.For example, the prevalence of myopia in the right and left eyes is larger than that for previous studies with the prevalence ranging from 4% to 7% conducted inside and outside Africa 19,22,27,35,36,37,38,39,40,41 but is smaller than that of other studies where the prevalence ranged from 27% to 33.8%. 15,22,27,30,37,42,43,44,45,46e prevalence of hyperopia for the right and left eyes of the present study in both stratified random samples (for 2018 and 2019) is not similar where the prevalence of hyperopia ranges from 3.8% to 7.53% to that of other studies conducted outside and inside the border of Africa.The prevalence of hyperopia for the right and left eyes of the present study in two samples is larger than that of other studies where the prevalence ranges from 4.5% to 6.5% 15,18,26,28,30,34,35,36,39,43,45,47,48 but is smaller than that of other studies that were conducted outside the African continent (with the prevalence ranging from 8% to 53.3%). 15,19,22,23,25,27,35,36,37,38,40,44,45,46,49,50,51,52,53e prevalence of astigmatism in the right and left eyes of the present study in samples ranged from 33.8% to 43.4%.The prevalence of astigmatism in the right and left eyes of the present study is not consistent with other previous studies conducted outside and inside Africa and is smaller than that of some studies where the prevalence ranges from 45.6% to 68.1% 19,22,27,38,52 but is larger than that of other studies conducted outside Africa where the prevalence ranges from 2.8% to 25.7%, which could be because of different population and settings. 15,22,23,25,26,29,34,35,36,39,42,46,53us, the results of the present study show that astigmatism is the most prevalent refractive error for the right and left eyes for both stratified random samples, and these results are consistent with previous systematic review and meta-analysis as reported by Hashemi et al. 49 and are not consistent with some studies conducted both outside 15,37,42,46,51 and inside the African continent 23,25,30,34,54 but are consistent with some studies outside Africa 19,35,36,52,55 and inside Africa. 22,27,38The reason for these inconsistencies could include differences in definitions of astigmatism or study designs.Females had a large prevalence of astigmatism for the right and left eyes in both samples of the present study compared to males, and this is consistent with one previous study 36 but is not consistent with another previous study.The prevalence of presbyopia for the right and left eyes in the 2018 sample of the present study is similar at 50.5% but the prevalence of presbyopia for the right and left eyes is similar at 47.0% for the 2019 sample.The prevalence of presbyopia in the right and left eyes in both 2018 samples is not consistent with that of other previous studies conducted in different parts of the world.The prevalence of presbyopia for the right of the present study in the 2018 sample is larger than that of other studies conducted outside Africa where the prevalence ranges from 30.4% to 42.9% 56,57 and is smaller than that of other studies conducted in Africa with the prevalence of 59%. 46The prevalence of presbyopia for the left eye of the present study in the 2018 sample is larger than that of other studies where the prevalence ranges from 30.4% to 42.9%. 56,57For the 2019 sample, the prevalence of presbyopia is smaller than that of another study (59%). 47e age of patients consulted at a rural optometry clinic at Sekororo District Hospital in 2018 is associated with determinants of refractive errors based on the results of univariate logistic regression analysis as shown in Table 6, but with multivariate logistic analysis, age was not a determinant of refractive errors and not statistically significant.For the 2019 sample, age is the determinant of refractive errors based on the univariate and multivariate logistic regression analysis and statistically significant for the right and left eyes but not the right eyes for the 2019 sample.These results of the present study are consistent with the results of other studies from different parts of the world, including those conducted within the African continent but with all previous studies.The age of the patients consulted at the clinic is the determinant of refractive errors of the present study and is consistent with other studies conducted outside Africa 42,52,58 and inside Africa 24,35,49,53 including South Africa 34 but not consistent with other studies outside Africa 15,17,40,42,45,46,58,59 and inside Africa. 23,24,25,32,53The present study assessed if the UPFS classifications were determinants of refractive errors or not.The UPFS classifications refer to the level of subsidisation of healthcare fees including assistive devices like spectacles for patients using government hospitals as defined by the South African government. 11n Table 6, the UPFS classifications were not determinants of any refractive errors but were statistically significant.

Limitations of the study
The present study was limited to one rural optometry clinic at a district hospital based on the historical clinical records extracted from the archive.The results of this study can only be generalised in the district hospital setting not in the general population.Another limitation of the present study is that the patients who consulted at the clinic were presenting with visual acuities, which could lead to an overestimation of the overall refractive errors.The study could also be susceptible to confounding effects because of the presence of associated factors of URE of the eye, such as the presence of cataracts and other ocular conditions where reduced vision could be restored after treatment.

Conclusions and recommendations
The most prevalent refractive error for the right and left eyes for both stratified random samples (2018 and 2019) is astigmatism, followed by myopia.Hyperopia for the right and left eyes was the least prevalent refractive error in both samples.In the 2018 sample, the prevalence of presbyopia for the right and left eyes was similar, but the 2018 sample had a larger prevalence of presbyopia compared to that of the 2019 sample.Females had a larger prevalence of refractive errors than males, but more females also attended the clinic.The optometry clinic provides a useful service to the rural community of the Maruleng subdistrict in the Mopani District, but more optometrists are required to expand and diversify the services on offer.Promotional activities will enhance the reach of the community service on offer.

Table 5 ,
hyperopia was the most common refractive error for both the right and left eyes in the age group ≤30 years in both samples (2018 and 2019), but the age group 31-60 years was more myopic, and the age group 61-90 years was more astigmatic.Emmetropia in the right and left eyes for both samples was more in the age group 31-60 years.However, the age group 61-90 years in two stratified samples was less emmetropic in which the right eyes were more emmetropic than the left eyes.Table6summarises the results of the univariate and multivariate logistic regression analysis to identify determinants of refractive errors for the right and left eyes of all patients consulted at the optometry clinic at Sekororo District Hospital from 2018 to 2019.The table also includes the uniform patient fee scheme (UPFS) as a variable for refractive errors.The categories H0 and H1 indicate the applicable subsidy for services that patients receive.

TABLE 3 :
Summary of the prevalence of presbyopia for the right and left eyes for the 2018 and 2019 samples.

TABLE 1 :
Summary of the demographic profiles for the 2018 and 2019 samples.: The table shows the sample sizes (n) and age groups presented in three intervals. Note

TABLE 2 :
Prevalence of emmetropia and refractive errors such as myopia, hyperopia and astigmatism of the right and left eyes for the 2018 and 2019 samples.

TABLE 4 :
Prevalence of refractive errors (myopia, hyperopia and astigmatism) and emmetropia based on gender with 95% confidence intervals (CI) of the right and left eyes for the 2018 and 2019 samples.
Note: Prevalence of refractive errors based on gender groups.CI, confidence interval; D, diopter.

TABLE 5 :
Prevalence of myopia, hyperopia, astigmatism and emmetropia for the right and left eyes with 95% confidence intervals (CI) showing the lower and upper limits based on age groups in 30-year intervals of the patients consulted at the clinic in 2018 and 2019.
Note: Prevalence of refractive errors based on age group.CI, confidence interval; D, diopter. 15

TABLE 6 :
Summary of logistic regression analysis between refractive errors and categorical variables for the right and left eyes for 2018 and 2019.