download PDFfull article:
According to global estimates of visual impairment by WHO, 285 million people are visually impaired worldwide1. The distribution of visual impairment is not uniform throughout the world. The least developed and developing countries have the highest burden of visual impairment. Among these populations the distribution of visual impairment varies across different age groups and gender; adults older than 50 years and females are at higher risk2.
India is geographically, linguistically, ethnically and socioeconomically diverse, so the need for and uptake of health care vary significantly throughout the population3,4. Currently in India and other developing countries, the task of providing standard eye examinations to every individual is limited by the lack of qualified eyecare practitioners and proper infrastructure3. There are only 12 eyecare practitioners per million population in Asia5 and in India there is only one ophthalmologist per 100 000 population3,6. Comprehensive eye examinations using an effective screening method would be appropriate to detect populations at risk. In India, telemedicine activities have been emerging since 1999 and it is one of the feasible strategies to provide quality care7.
Currently there are many successful teleophthalmology models in India, such as the Sankara Nethralaya Tele-Ophthalmology Project (SNTOP) and Aravind Tele Network (ATN)8. Studies have reported that such models are effective in detecting the diseases early and provide timely intervention9. Currently these tele-eyecare models aim at detecting structural changes through imaging techniques. The basic functional parameter that is used to quantify visual impairment is visual acuity (VA). In India, and in countries with similar demographics where there is high demand for detection of avoidable blindness, basic VA screening is provided by trained personnel (a vision technician), school teachers and social workers. Owing to these logistics issues, quality of eye care can be impaired. A recent study assessing the accuracy of vision and refractive error estimation by a non-formally trained technician showed poor accuracy in refraction and good VA evaluation10. There is a need for innovative methods in tele-eye care that could aid in providing quality primary eye care to the population at large. The technology should also facilitate proper utilization of manpower and provide culturally competent service to meet the needs of a multilingual population11.
Visual acuity measurement is the most important clinical parameter in ocular examination. Systematic approach of VA measurement helps in effective screening of populations at risk. Exploring the possibility of incorporating the standard technique of VA measurement (real time, virtual platform) in a tele-eyecare model would be a step towards reducing the burden of visual impairment.
Computerized logMAR (COMPlog; http://www.complog-acuity.com) is a digital portable, user-friendly and validated VA testing system. Studies have shown that the efficacy of digital VA charts is as good as the conventional VA charts12. The authors of the present study have previously reported good agreement of telebased VA measurement among a normal young literate university population13. This computer-savvy, literate young student population responded well to the virtually monitored VA test. The results showed good agreement between virtual and face-to-face examination. Such a virtual vision examination model would face more challenges in rural applications due to various factors such as illiteracy, multilingual local vernacular languages and acceptance of telebased methods. Thus the current study evaluates the efficacy and applicability of virtually controlled telebased VA methods among a multilingual hospital-based population using the COMPlog VA system.
Subjects were recruited from the Outpatient Department of Optometry, School of Allied Health Sciences, Manipal University and Department of Ophthalmology, Kasturba Medical College, Manipal, India. In this cross-sectional study a sample size of 96 subjects was estimated, considering a mean difference of 0.01 logMAR and standard deviation of 0.16 logMAR, alpha error at 5% and power of 80%. Eligible subjects were enrolled for the study and categorized into three language groups: Malayalam-speaking group (n=32), English-speaking group (n=32) and Kannada-speaking group (n=32). This cross-sectional study was conducted between June 2012 and April 2013.
Subjects aged greater than 18 years with a presenting VA of better than 6/60 (1.0 logMAR) were included. Subjects were also recruited according to their language preference, which was based on dialect or mother tongue (Kannada, Malayalam or English language). Subjects with any history of ocular surgery were excluded.
Presenting VA was measured by both face-to-face and telemethods by two experienced optometrists. The order of testing was randomized and the examining optometrist was masked from the previous VA readings.
Visual acuity measurement
COMPlog is a VA measurement software operated through a laptop computer (Windows system). The patient views the optotypes on a secondary monitor (24 inch, 1600x1200 resolution LED flat panel). For this study, the secondary monitor was calibrated for a 3 m distance for performing VA measurements. The optometrist controlled the test and recorded on the laptop whether the patient's responses were correct or incorrect. The software runs an acuity screening to quickly identify the rough VA of the patient and further refines using a thresholding technique. The detailed methodology of VA estimation using COMPlog is available elsewhere12,13. In both the methods, VA was measured in the COMPlog system using Landolt 'C', a type of visual acuity chart used to estimate acuity among illiterate patients. Prior to the measurements all the subjects received uniform formatted instructions. Visual acuity was measured first for the right eye for all the subjects.
Face-to-face method: This was similar to the method that would be used at a conventional eye clinic. The VA examination was performed by an optometrist (SK) physically present in the clinic. Instruction was predominantly provided in English with a few dialects from the local vernacular language for better understanding of the test instructions. Such a limitation prevails in a regular clinic, as the clinician's proficiency towards the vernacular language has a significant effect, from history-taking to patient counselling.
Telemethod: Subjects were examined by another optometrist (masked) and located elsewhere (different premises). The optometrists (JJ, MD) were comfortable speaking and comprehending all three languages (Kannada, Malayalam and English). The mode of communications and instructions were made as per patients' choice of comfort. In telemode, the examining optometrist who was fluent in the respective languages communicated with the patient and completed the examination. Lync software (Microsoft; http://office.microsoft.com/en-us), laptops with specific internet protocol (IP) addresses and internet connectivity were used to create a virtual interface between the subject and examining optometrist elsewhere and to have remote control of the COMPlog system. Virtual (remotely operated) control of examination procedures and videoconferencing aided the optometrist positioned elsewhere (different physical location) to remotely operate the COMPlog VA chart and to interact with the subjects.
Patient feedback was assessed for both the methods using a questionnaire. The questions addressed clarity, understandability, quickness, acceptability and overall satisfaction of both the methods. The response was given using a five-point scale ranging from 0 to 4 (worse to best). The presenting VA measurement of the left eye of every subject was considered for analysis. Data collected for the analysis were logMAR VA, duration of measurement and questionnaire responses for both the procedures (face-to-face and telemethod).
Ethics approval
The study was approved by the Institutional Research Board, School of Allied Health Sciences, Manipal University (IEC175/2013). Written informed consent was obtained from all participants and the study was performed in accordance with the tenets of the Declaration of Helsinki.
The mean age of 96 subjects was 40.3 (standard deviation=14.1, range=18-71) years. There was no statistically significant variability in age distribution across language groups (p=0.345). Male-to-female ratio was 45:51. Gender distribution among the language groups was not statistically significant (p=0.806).
Table 1 shows the mean VA of 96 subjects and of the language groups using both face-to-face and telemethod. The mean difference in VA between the two methods was -0.00 (±0.16) logMAR. The difference in VA measurement using the two methods was not statistically significant for the overall data (p=0.844). One-way analysis of variance (ANOVA) was done to analyse the variation in the VA difference across language groups, which was not statistically significant (p=0.140; Table 1).
The two methods had good intra-class correlation (0.912, 95% confidence interval (CI) 0.868-0.941). A Bland-Altman plot showed good agreement between the two methods (F1). The 95% limit of agreement for the mean difference in VA ranged from 0.31 to -0.32 logMAR.
When classifying the subjects with VA better than 0.3 logMAR (6/12) and worse or equal to 0.3 logMAR, the two methods had good agreement, Kappa statistic 0.742 (95%CI 0.601-0.895). The VA (better or worse) across language groups also had good statistical agreement (Table 2).
The variability of VA measures with the three language subset population across various VA ranges is shown in Figure 2. The changes in VA for all three groups were well within the clinically accepted test-retest variability of ±0.2 logMAR values. Language-specific instruction showed the least variability for the Kannada and English groups. Among the Malayalam group there was improvement in telebased methods for poorer VA, but the association was not statistically significant.
Mean time in seconds taken for the face-to-face method (64.63±30.24) was less than for the telemethod (76.65±32.69) and showed statistical significance (p=0.001). The difference in time between the two procedures did not statistically vary across language groups (p=0.355). Seventy six percent of the participants completed all of the questionnaire. More than 87% reported high satisfaction with the telemethod and 89% were satisfied with the face-to-face method (p=0.149). Other parameters in the questionnaire did not suggest statistical significance between the two methods (p=0.157). Only one subject was dissatisfied with the acceptance of the telemethod. The medical records of that particular subject revealed moderate to severe hearing impairment.
Table 1: Comparison of face-to-face method and telemethod for measuring visual acuity
Table 2: Kappa statistic across the three language groups for visual acuity better than logMAR 0.3
Figure 1: Bland-Altman plot of the COMPlog visual acuity for face-to-face method and telemethod.
Figure 2: Change in COMPlog visual acuity between face-to-face method and telemethod
across three different language instruction groups. (Test-retest variability of ±0.2 logMAR)
Discussion
The present study examines various aspects (comparability, efficacy, time taken, easiness, acceptability and satisfaction during the procedure) of VA measurement that can be incorporated in a telebased model. The mean difference in VA between the two methods was 0.00 (SD±0.016) logMAR units. The VA measurements done in the telebased method were comparable to that of the face-to-face method. The 95% limit of agreement between the two methods ranged from +0.31 to -0.32 logMAR units. Srinivasan et al., who performed a similar study among young computer-literate university students and staff population, reported slightly less variability in 95% limits of agreement (+0.26 to -0.26 logMAR units)13. The variability reported could be due to the diverse and older population in the current study. However the COMPlog validation study done by Laidlaw et al12 reported a 95% limit of agreement (2 SD) on test-retest variability as ±0.10 logMAR. This large variability for the telemethod of VA may be attributed to the varying examining technique/methodology (face-to-face and telemethod) using the COMPlog system.
The kappa statistic showed a substantial agreement between the two methods for a referral cut-off of VA worse than 0.30 logMAR (6/12), demonstrating the applicability of the telemethod of VA measurement as an effective screening tool. The present study reports the clinical acceptability of the telebased VA measurement compared to face-to-face, with a variability (1 SD) of 0.16 logMAR unit. Rosser et al. reported that a test-retest variability in VA of logMAR 0.2 or greater is considered as clinically significant change using the ETDRS eye chart14. The changes in VA for all three groups were well within the clinically accepted test-retest variability of ±0.2 logMAR. The influence of language-specific instructions in the telemethod showed a positive effect on the variability of VA (Fig2).
The quality of VA examination and refraction performed by semi-trained tecnicians, even with the use of autorefractors, was reported to be limited10. Often the clinician gets assistance from colleagues or translators to communicate with patient in instances where language is an issue. The limitation of this would be availability of other clinicians at the premises and lack of knowledge about the condition among translators, resulting in poor translation ability. The limitation of the examining optometrist unable to speak the mother tongue or local dialect has an indirect effect on the quality of care provided in a multilingual Indian population.
The authors report on a novel method of telebased consultation providing the option of connecting an available professional eyecare provider (without geographical boundaries) who can perform a virtual eye examination. This telebased virtually controlled model explores the versatility of connecting clinician and patient based on language preferences, thus establishing a better customer relationship and thereby improve quality of eye care.
Studies have reported good patient satisfaction with tele-ophthalmology methods15. Multiple applicability of the telebased model is yet to be explored for effective utilization of human resources8,9,16. The scope of application includes businesses such as optical shops, satellite centres of base hospitals and rural outreach programs for effective eyecare delivery16. This particular virtually controlled tele-eyecare model has limited application in people with severe hearing deficit and non-verbal persons, hence the acceptability may be less in those population as reported in this study.
Virtual eye examination could eventually become a universal eyecare model, where an eyecare practitioner could perform comprehensive eye screening in a virtual platform without being physically present with a patient. The future of such a model would include refractive services through automated (virtually operated) refractometers and phoropters, anterior and posterior segment imaging technology, along with telebased VA examination. Incorporating vernacular instruction and native dialect during the examination procedure would increase acceptability and aid in culturally competent eyecare services in such a model in rural and remote areas.
The possible limitation of this model would probably be high capital investment, internet data connectivity and electricity supply in remote and rural locations. The authors hope that the capital investment would be compensated by low recurring cost, compared to the mobile tele-eye screening model; this cost-effectiveness needs to be studied. The growth in alternative energy sources would meet the needs for continuous power supply in rural regions, but would add to the initial capital cost. Growing interest of government (services such as National Knowledge Network, India) and private organisations towards better data connectivity could aid in implementing such virtual eyecare units.
Visual acuity measurements in the telemethod along with native dialect were comparable to conventional face-to-face methods in a hospital-based multilingual population. The COMPlog VA testing system can be effectively integrated in to a tele-eyecare model.
Acknowledgements
The authors wish to thank Satish Kamath and KB Varadharaj, Department of IT and Network Administration, Manipal University, for teleconnectivity support. They also wish to thank Mr Babu Noushad, Assistant Professor, Department of Optometry, School of Allied Health Sciences, Manipal University, for his mentoring and support during his tenure.
References
1. Pascolini D, Mariotti SP. Global estimates of visual impairment: 2010. British Jounal of Ophthalmology 2012; 96: 614-618.
2. Resnikoff S, Pascolini D, Etya'ale D, Mariotti SP, Pokharel GP. Global data on visual impairment in the year 2002. Bulletin World Health Organization 2004; 82: 844-851.
3. Thomas R, Paul P, Rao GN, Muliyil JP, Mathai A. Present status of eye care in India. Survey of Ophthalmology 2005; 50: 85-101.
4. Dandona R, Dandona L, Srinivas M, Giridhar P, Prasad MN, Vilas K et al. Moderate visual impairment in India: the Andhra Pradesh Eye Disease Study. British Jounal of Ophthalmology 2002; 86: 373-377.
5. Holden BA, Rao GN, Knox KM, Sulaiman SM. Visual impairment: a correctable global problem. Medical Journal of Australia 1997; 167: 351-352.
6. Holden BA , Resnikoff S . The role of optometry in vision 2020. Community Eye Health 2002; 15: 33-36.
7. Mishra SK, Kapoor L, Singh IP. Telemedicine in India: current scenario and the future. Telemedicine Journal and eHealth 2009; 15: 568-575.
8. Prathiba V, Rema M. Teleophthalmology: a model for eye care delivery in rural and underserved areas of India. International Journal of Family Medicine 2011; 4: article ID 683267.
9. Verma M, Raman R, Mohan RE. Application of tele-ophthalmology in remote diagnosis and management of adnexal and orbital diseases. Indian Journal of Ophthalmology 2009; 57: 381-384.
10. Zhou Z, Zeng J, Ma X, Pang X, Yi H, Chen Q, et al. Accuracy of rural refractionists in Western China. Investigative Ophthalmology and Vision Science 2014; 55: 154-161.
11. Ngo-Metzger Q, Telfair J, Sorkin DH, Weidmer B, Weech-Maldonado R, Hurtado M, et al. Commonwealth fund - cultural competency and quality of care: obtaining the patient's perspective. (Online) 2006. Available: http://www.commonwealthfund.org/~/media/Files/Publications/Fund%20Report/2006/Oct/Cultural%20Competency%20and%20Quality%20of%20Care%20%20Obtaining%20the%20Patients%20Perspective/Ngo%20Metzger_cultcompqualitycareobtainpatientperspect_963%20pdf.pdf (Accessed 13 June 2013).
12. Laidlaw DA, Tailor V, Shah N, Atamian S, Harcourt C. Validation of a computerised logMAR visual acuity measurement system (COMPlog): comparison with ETDRS and the electronic ETDRS testing algorithm in adults and amblyopic children. British Jounal of Ophthalmology 2008; 92: 241-244.
13. Srinivasan K, Ramesh SV, Babu N, Sanker N, Ray A, Karuna SM. Efficacy of a remote based computerised visual acuity measurement. British Jounal of Ophthalmology 2012; 96: 987-990.
14. Rosser DA, Cousens SN, Murdoch IE, Fitzke FW, Laidlaw DA. How sensitive to clinical change are ETDRS logMAR visual acuity measurements? Investigative Ophthalmology and Vision Science 2003; 44: 3278-3281.
15. Kumari Rani P, Raman R, Manikandan M, Mahajan S, Paul PG, Sharma T. Patient satisfaction with tele-ophthalmology versus ophthalmologist-based screening in diabetic retinopathy. Journal of Telemedicine and Telecare 2006; 12: 159-160.
16. World Health Organization. Telemedicine: opportunities and developments in Member States: report on the second global survey on eHealth. (Online) 2009. http://www.who.int/goe/publications/goe_telemedicine_2010.pdf (Accessed 13 June 2013).