This study demonstrates that a photographer-led eyelid lesion clinic is a safe and effective way of assessing patients referred to secondary care with an eyelid lesion which is felt likely to be benign. We demonstrate that such a service can significantly reduce the time patients wait for assessment and interventional treatment if required. We show that it is effective at the identification of eyelid malignancies with a rate of missed malignancy of 0.13%. We consider a 6 months re-referral rate of 9% to be reasonable, especially given around 20% of these re-referrals were for an unrelated pathology. However, our appointment non-attendance rate of 8.3% was slightly higher than the 6.8% recorded in Scotland in the first quarter of 2019/20 , the last period for which data is available.
Our figure of 60.9% correlation between clinical and histopathological diagnosis is lower than that reported in series such as Banerjee et al.  who reported 90.2% concordance for benign lesions and 67.6% for malignant lesions. This was however based on an in person evaluation of eyelid lesions. Our rate of missed malignancy of 0.13% compares favourably with a study by Izzettinoglu et al. who reported a 2.2% rate of lesions diagnosed as benign clinically which were found to be malignant on histopathology within a series of 408 eyelid lesions . Further none of the 330 patients seen in the first six months of the study period were re-referred with a suspected malignancy. This is also a surrogate marker of the extremely low rate of missed malignancies.
This study was not specifically designed to assess cost implications of our IBELMS service however we feel the service is likely to be more cost-effective than a traditional face-to-face model. Prior to the introduction of the service, a proportion of these patients would have been seen in a face-to-face consultant-led clinic and a proportion would have been referred directly to a nurse-led minor operations clinic. Many patients would have had an image of their eyelid lesion recorded for documentation. In the IBELMS model around 3.5 times as many patients can be reviewed by a consultant in a session than in a face-to-face model (40 IBELMS patients vs 12 patients in a standard clinic profile). The number of patients booked for surgery who do not need or qualify for surgery is likely to be lower with prior secondary care evaluation. These savings are partially offset by the cost of increased medical illustration staff and time in the IBELMS model however. There were no capital costs associated with the IBELMS service as it uses existing equipment and software. We would advocate a study designed to specifically assess cost implications of this model to support introduction in other units.
Several studies have reported on the utility and safety of telemedicine for the assessment of eyelid lesions. Kang et al.  in a study of 44 patients showed 91% agreement in diagnosis between in person nurse specialist review and remote consultant review of images and 82% agreement with management. No lesion felt to be malignant on in person review was deemed benign by the remote reviewer. Ah-Kye et al.  recently published data from synchronous telemedicine consultations for patients with presumed benign eyelid lesions. In their unit, patients send a self-taken photograph of their eyelid and then receive a video consultation with a clinician to discuss the diagnosis and management. They had a slightly higher rate of patients directly listed for surgery than in our series (57.3% vs 47.9%) however they discharged fewer patients (22.8% vs 33.2%) . They report a very low non-attendance rate at 2.57%. We propose that not having to attend a healthcare facility for any aspect of the consultation is likely to lower non-attendance rates versus our model. However, patient-supplied images and a reliance on the quality of a patient’s device camera versus a medical photographer image may reduce the confidence of a clinician in discharging a patient hence the lower discharge and higher clinic review rate in their series.
We believe this study to be the largest to date looking at asynchronous remote evaluation of eyelid lesions. Whilst it was performed at a single centre, there were a number of photographers and consultant reviewers who evaluated patients so we feel these findings could be replicated at other units and the results are not likely influenced by any individual photographer or clinician. We feel that this study would be representative of the UK population as a whole as it would have captured the majority of referrals with eyelid lesions within the Greater Glasgow area during this period, as no other public sector provider treats patients for periocular lesions in the region. In planning an image-based eyelid lesion service, the reported 7–8% rate of re-referrals must be factored in.
Asynchronous and synchronous telemedicine both have advantages and disadvantages. As demonstrated by Ah-Kye et al.’s study  synchronous telemedicine with patients not being required to attend a healthcare facility, non-attendance rates are likely to be lower. However, this model risks excluding the 10% of patients who do not use the internet  whereas we feel that our model is accessible to a broader range of patients. It does involve patients attending a healthcare site though so whilst capacity in consultant lead face-to-face clinics is improved, the benefits in sustainability as demonstrated by Ah-Kye et al.  would not be replicated with this model. In the future, clinical photographs could be taken in optometry practices which would combine the benefits of a professional standard photograph with a reduction in patient travel. Preliminary data from cloud-based referral platforms suggest the potential to reduce hospital attendance by up to 52% . With synchronous telemedicine, clinicians need to undertake consultations during a set period when appointments are booked whereas with asynchronous telemedicine, clinicians can review records at a time that suits them, facilitating flexible working. Patients do not need to set aside a specific time for an appointment as they do with synchronous telemedicine consultations which many prefer.
Limitations include the fact that the study was retrospective and recording was not complete for all patients. 6 months follow up data was only available for less than half of the studied patients and, given that not all patients had a biopsy or excision procedure performed, the exact diagnosis of their lid lesions cannot be confirmed. The study did not evaluate patient satisfaction rates in this cohort; however surrogate measures for satisfaction, including re-referral rates, were low.
In conclusion, we demonstrate that an image-based asynchronous teleconsultation model for the assessment of patients with presumed benign eyelid lesions is efficient and safe. It Is able to reduce waiting times for an assessment appointment, allow clinicians to work flexibly and minimise use of clinic consultation space whilst accurately identifying periocular malignancies requiring prompt treatment. We recommend that other units with an increasing waiting list of patients with eyelid lesions or restricted by face-to-face clinic capacity consider adopting this model.