Evaluation of a Telemedicine Model for Following Keratoconus Patients in the Era of COVID-19 Pandemic
Introduction
Keratoconus is a progressive bilateral corneal ectasia resulting in myopia, irregular astigmatism, and, eventually, loss of corrected distance visual acuity. The estimated prevalence rate is 1 in 500–2000 people.1 Identifying keratoconus at an early stage is highly significant, as treatments such as corneal crosslinking (CXL) are able to halt disease progression and preserve good vision.2 The natural history of this disorder requires repeated outpatient visits for monitoring for disease progression. The main imaging techniques for patient diagnosis and follow-up include corneal topography with Placido disk-based imaging systems and corneal tomography, such as Scheimpflug or optical coherence tomography imaging.
During the era of the coronavirus disease 2019 (COVID-19) pandemic, it is advised to decrease in-person consultation to maintain social distancing. Maintaining social distancing measures holds a challenge for ophthalmologists, particularly during slit lamp examination. This provides a unique opportunity for embracing telemedicine, ideal for mitigating overcrowding of hospitals and clinics, decreasing exposure, and still promoting delivery of care. Telemedicine also decreases the burden on the health care system.
Telemedicine has been used widely in ophthalmology, most commonly for the screening and monitoring of diabetic retinopathy,3,4 age-related macular degeneration (AMD),5,6 and glaucoma.7,8 These diseases share similar features in which visual acuity measurements and routine imaging can accurately identify patients who require further care. Interestingly, keratoconus, which also shares this feature for monitoring, has not been studied for the feasibility of telemedicine service.
The purpose of this study was to assess the diagnostic accuracy and reliability of a telemedicine approach for detecting keratoconus patients’ progression.
Methods
A retrospective study included 204 eyes of 102 consecutive keratoconus patients who were followed and treated in our keratoconus outpatient clinic between November 2017 and September 2020.
We included patients who had at least one clinic visit before the visit included in the study and have completed manifest refraction and corneal topographies as described hereunder. Patients were excluded from the analysis if they were newly diagnosed keratoconus patients, had a history of ocular disease or surgery (including previous CXL before the study period), and those missing complete data on both eyes.
The study was approved by the institutional review board (IRB) of the Tel Aviv Medical Center, Tel Aviv, Israel, and complies with the tenets of the Declaration of Helsinki.
The outpatient clinical examination included measurement of the best spectacle corrected visual acuity (BSCVA) and manifest refraction performed by a certified optometrist and a full ophthalmic examination by a cornea surgeon. All patients underwent corneal topography (topographic modeling system-4, Tomey Corporation, Phoenix, AZ; and Galilei G4, Ziemer, Switzerland). The risk for keratoconus progression and the indication for CXL were determined. Progression of keratoconus was defined as an increase of maximum keratometry (Kmax), mean central K-readings, decrease in mean central corneal thickness, increase of myopia and/or astigmatism, or a subjective loss of vision.9–11
Data were abstracted for demographics, BSCVA with manifest refraction, and office diagnostic decision.
Evaluation of the accuracy of the telemedicine model was performed by having the two corneal specialists make a remote assessment of their keratoconus patients. The minimal time elapsed between office and remote assessment was 4 weeks.
During the remote assessments, first patient demographics and topographies were revealed to the ophthalmologist and a decision of keratoconus progression or stability was made. Second, manifest refraction and clinical findings obtained in the office visit were revealed to the examiner, and a second decision was made. The purpose of this two-step procedure was to determine the additive value of the clinical examination, performed only at the office visit.
The ophthalmologists determined the indication for keratoconus progression and referral for CXL using the same data and criteria at both the office examination and remote assessment. We refer to the decisions obtained at the remote assessment as remote diagnostic decisions 1 and 2.
As the clinic examination provided with the gold-standard diagnosis, we compared the decision obtained at the clinic with the remote diagnostic decisions and calculated sensitivity, specificity, and positive and negative predictive value (PPV and NPV, respectively). A schematic presentation of the study design and comparison of data are shown in Figure 1.
Fig. 1. Schematic representation of study design and goals.
Results
A total of 204 eyes of 102 keratoconus patients were included in the study. The mean patient age was 29 years (range, 10–53 years) and 67 (66%) were male. Patient baseline and clinical data are presented in Table 1.
| PATIENTS (102) | EYES (204) | |
|---|---|---|
| Age | 29.36 ± 8.6 (10–53) | |
| Male, n (%) | 67 (66%) | |
| Topography | ||
| Kmax, D | 49.35 ± 5.4 (38.4–74.1) | |
| Cyl, D | 3.99 ± 2.9 (0–15.1) | |
| CCT min, μm | 487.64 ± 52.3 (310–599) | |
| Refraction | ||
| Sphere, D | −0.17 ± 2.6 (−13.5–6) | |
| Cyl, D | −3.20 ± 2.8 (−13–1.25) | |
| BCVA (LogMar) | 0.17 ± 0.2 (0–2) | |
During the office examination, 13 (6%) eyes were referred to corneal CXL due to keratoconus progression. During remote evaluations, 17 (8%) eyes were referred to corneal CXL. In no case was remote diagnostic decision 2 different from remote decision 1 (i.e., by revealing manifest refraction and clinical findings).
There was an agreement of assessment between the office and remote diagnostic decisions in 192 (94%) of the eyes. Among the remaining 12 eyes, 8 (3.9%) eyes had false-positive diagnoses and 4 (1.9%) eyes had false-negative diagnoses. There was a calculated false-positive rate of 4.2% and a false-negative rate of 30.8%. Sensitivity and specificity were 69.2% (95% confidence interval [CI] 38.57% to 90.91%) and 95.8% (95% CI 91.9% to 98.2%), respectively. The PPV and NPV were 52.9% (95% CI 34.3% to 70.8%) and 97.9% (95% CI 95.3% to 99%), respectively. These results are shown in Table 2.
| OFFICE DECISION | NEGATIVE & POSITIVE PREDICTIVE VALUE, % (95% CI) | |||
|---|---|---|---|---|
| STABLE, N (%) | PROGRESSION, N (%) | TOTAL | ||
| Remote decision | ||||
| Stable | TN 183 (89.7) | FN 4 (1.9) | 187 | 52.9 (34.3–70.8) |
| Progression | FP 8 (3.9) | TP 9 (4.4) | 17 | 97.9 (95.3–99) |
| Total | 191 | 13 | 204 | |
| Specificity & sensitivity, % (95% CI) | 95.8 (91.9–98.2) | 69.2 (38.57–90.91) | ||
Among the eight eyes with false-positive diagnoses, referral for CXL treatment during remote assessment was indicated due to progression in both Kmax and cylinder in four eyes, progression of only Kmax in three eyes, and for the remaining patient an overview of all previous topographies revealed a trend in dynamics that was considered borderline in the office examination. Regarding the four false-negative diagnoses, two patients were believed to have an unreliable corneal topography during the remote assessment, therefore, not indicating true progression, and two patients showed progression in Kmax values, however, other parameters remained stable.
Discussion
The importance of continuous long-term follow-up of keratoconus patients to timely identify signs of progression has set a burden for both patients and clinicians, as well as encouraged to identify alternative options for keratoconus patient monitoring.
The concept of “virtual clinic” is increasingly being adopted in the care of glaucoma12–14 and AMD5,6 patients, using visual acuity test and different imaging techniques. The COVID-19 pandemic has led to a unique opportunity for investigating whether telemedicine may be a beneficial alternative for periodic office visits for keratoconus patients. The idea is that an optometrist would record manifest refraction and trained nurses or technicians would obtain corneal topographies, whereas the cornea specialist will determine whether treatment is indicated. This allows for a rapid assessment of data and diagnostic images without the patient present, maintaining social distance requirements and reducing the time for both the patient and the physician.
Our results showed that this model, at this time point, could not be applicable for keratoconus patient follow-up, as remote assessment had high specificity but low sensitivity.
Using the current telemedicine model would deprive a relatively high percentage of patients from the benefit of early treatment, halting their disease progression. Although only four patients had a false-negative diagnosis, due to the low rates of referral for CXL in our cohort, this value beholds a high false-negative rate.
Despite the high specificity calculated for the telemedicine model, we would have falsely treated eight patients. Overtreatment exposes the patient to iatrogenic complications as well as apposes an economic burden.
Of the patients with false-positive diagnoses, most had objective topographical signs of progression that were addressed as higher suspicious cases during office visits and these patients were requested for closer evaluations. It seems that there is higher ophthalmologist tolerance for small changes in progressive parameters during the in-person examination. This suggests that a telemedicine approach may encourage a thorough review of previous scans for the assessment of corneal topography dynamics and increase the confidence to treat in doubtful cases.
Limited publications exist on telemedicine for anterior segment diseases. Several attempts for telediagnosis of anterior segment pathology through portable cameras showed highly variable sensitivity and specificity, based on the particular pathology.15–17 Saleem et al.18 discussed that cornea specialists were less likely to use telemedicine unless the pathology is obvious externally, as virtually they will not be able to examine the ocular surface.
The fact that the remote decision was not affected by revealing the clinical examination findings highlights that this slowly progressive disorder is suitable for a telemedicine approach. Most false-positive cases were indicted as higher risk patients and warranted closer follow-up. This suggests that the high false-positive case is affected by other factors such as a more conservative approach during the office visit, lack of time for a thorough comparison of previous topographies, or other factors not considered in our retrospective evaluation such as patient consideration and compliance. It is also likely that the high false-negative rate and low sensitivity derive from the low rate of CXL referral, as numerically the low numbers reached statistical significance.
Limitations of this study are its retrospective nature and the difficulty in comparing our results with similar studies. This includes the lack of a standardized protocol to use for keratoconus assessment in our study. As opposed to pathologies such as diabetic retinopathy, with clear guidelines and protocols,19 there remain many controversial aspects of disease definition and diagnosis.9 Currently, there is no consistent or clear definition of ectasia progression, with high variability in the decision making for the medical and surgical management of these patients. To avoid interobserver variability, the same clinician performed the office and remote evaluation. This brings another potential bias, avoided by the encryption of the topographies as well as the time elapsed from the clinic examination to the remote assessment.
In addition, this study did not show that telemedicine is more efficient than the traditional methods of keratoconus patient follow-up. Further analysis is needed to determine the economic impact of this model and the cost of overtreatment of false–positive cases.
Conclusion
The incorporation of telemedicine for keratoconus patient follow-up is premature and not yet suitable for application. Additional work is warranted to improve the sensitivity to meet telemedicine ophthalmic standards.
Disclosure Statement
No competing financial interests exist.
Funding Information
No funding was received for this article.
References
- 1. Age-specific Incidence and prevalence of keratoconus: A nationwide registration study. Am J Ophthalmol 2017;175:169–172. Crossref, Medline, Google Scholar
- 2. . Updates on managements for keratoconus. J Curr Ophthalmol 2018;30:110–124. Crossref, Medline, Google Scholar
- 3. . Accuracy and reliability of teleophthalmology for diagnosing diabetic retinopathy and macular edema: A review of the literature. Diabetes Technol Ther 2006;8:102–111. Link, Google Scholar
- 4. Long-term comparative effectiveness of telemedicine in providing diabetic retinopathy screening examinations: A randomized clinical trial. JAMA Ophthalmol 2015;133:518–525. Crossref, Medline, Google Scholar
- 5. The concept of virtual clinics in monitoring patients with age-related macular degeneration. Acta Ophthalmol 2016;94:e353–e355. Crossref, Medline, Google Scholar
- 6. Evaluation of a telemedicine model to follow up patients with degeneration. Retina 2016;36:279–284. Crossref, Medline, Google Scholar
- 7. The effectiveness of Teleglaucoma versus in-patient examination for Glaucoma screening: A systematic review and meta-analysis. PLoS One 2014;9:e113779. Crossref, Medline, Google Scholar
- 8. Virtual clinics in glaucoma care: Face-to-face versus remote decision-making. Br J Ophthalmol 2017;101:892–895. Crossref, Medline, Google Scholar
- 9. Global consensus on keratoconus and ectatic diseases. Cornea 2015;34:359–369. Crossref, Medline, Google Scholar
- 10. Corneal cross-linking as a treatment for keratoconus: Four-year morphologic and clinical outcomes with respect to patient age. Ophthalmology 2013;120:908–916. Crossref, Medline, Google Scholar
- 11. Refractive, topographic, tomographic, and aberrometric analysis of keratoconic eyes undergoing corneal cross-linking. Ophthalmology 2009;116:369–378. Crossref, Medline, Google Scholar
- 12. . Service innovation in glaucoma management: Using a web-based electronic patient record to facilitate virtual specialist supervision of a shared care glaucoma programme. Br J Ophthalmol 2015;99:313–317. Crossref, Medline, Google Scholar
- 13. Experiences with developing and implementing a virtual clinic for glaucoma care in an NHS setting. Clin Ophthalmol 2015;9:1915–1923. Crossref, Medline, Google Scholar
- 14. Acceptability and use of glaucoma virtual clinics in the UK: A national survey of clinical leads. BMJ Open Ophthalmol 2018;3:e000127. Crossref, Medline, Google Scholar
- 15. Telemedicine for ophthalmic consultation services: Use of a portable device and layering information for graders. J Telemed Telecare 2016;23:365–370. Crossref, Medline, Google Scholar
- 16. Teleophthalmic approach for detection of corneal diseases: Accuracy and reliability. Cornea 2017;36:1159–1165. Crossref, Medline, Google Scholar
- 17. Novel telemedicine device for diagnosis of corneal abrasions and ulcers in resource-poor settings. JAMA Ophthalmol 2014;132:894–895. Crossref, Medline, Google Scholar
- 18. Virtual ophthalmology: Telemedicine in a COVID-19 era. Am J Ophthalmol 2020;216:237–242. Crossref, Medline, Google Scholar
- 19.
Early Treatment Diabetic Retinopathy Study Research Group . Early Treatment Diabetic Retinopathy Study Design and baseline patient characteristics: ETDRS report number 7. Ophthalmology 1991;98:741–756. Crossref, Medline, Google Scholar