Diabetic retinopathy is a microvascular complication of diabetes mellitus that is usually asymptomatic in the early stages. Therefore, its timely detection and treatment are essential. First pilot projects exist to establish a smartphone-based and AI-supported screening of DR in primary care. This study explored health professionals’ perceptions of potential barriers and enablers of using a screening such as this in primary care to understand the mechanisms that could influence implementation into routine clinical practice. Semi-structured telephone interviews were conducted and analysed with the help of qualitative analysis of Mayring. The following main influencing factors to implementation have been identified: personal attitude, organisation, time, financial factors, education, support, technical requirement, influence on profession and patient welfare. Most determinants could be relocated in the behaviour change wheel, a validated implementation model. Further research on the patients’ perspective and a ranking of the determinants found is needed.

Background Diabetic retinopathy (DR) affects 10–24% of patients with diabetes mellitus type 1 or 2 in the primary care (PC) sector. As early detection is crucial for treatment, deep learning screening methods in PC setting could potentially aid in an accurate and timely diagnosis. Purpose The purpose of this meta-analysis was to determine the current state of knowledge regarding deep learning (DL) screening methods for DR in PC. Data sources A systematic literature search was conducted using Medline, Web of Science, and Scopus to identify suitable studies. Study selection Suitable studies were selected by two researchers independently. Studies assessing DL methods and the suitability of these screening systems (diagnostic parameters such as sensitivity and specificity, information on datasets and setting) in PC were selected. Excluded were studies focusing on lesions, applying conventional diagnostic imaging tools, conducted in secondary or tertiary care, and all publication types other than original research studies on human subjects. Data extraction The following data was extracted from included studies: authors, title, year of publication, objectives, participants, setting, type of intervention/method, reference standard, grading scale, outcome measures, dataset, risk of bias, and performance measures. Data synthesis and conclusion The summed sensitivity of all included studies was 87% and specificity was 90%. Given a prevalence of DR of 10% in patients with DM Type 2 in PC, the negative predictive value is 98% while the positive predictive value is 49%. Limitations Selected studies showed a high variation in sample size and quality and quantity of available data.

Objective Diabetic retinopathy (DR) may lead to irreversible damage to the eye and cause blindness if diagnosed in its advanced stages. Artificial intelligence (AI) may support screening and contribute to a timely diagnosis. The aim of this study was to evaluate factors that might influence the success of implementing AI-supported devices for DR screenings in general practice. Methods A questionnaire with modules on attitudes toward digital solutions, technical factors, perceived patient perspectives, and sociodemographic data was constructed and 2100 general practitioners (GPs) in Germany were invited to participate via a personal letter. Results Two hundred nine physicians participated in the survey (10% response rate, mean age = 54 years, 46% women). Acquisition costs (mean = 1.37), remuneration (mean = 1.46), and running costs (mean = 1.40) were considered particularly relevant in the context of AI-based screening tools. GPs indicated that a mean of €27.00 (SD = 19) was considered to be an appropriate reimbursement for an AI-based screening for DR in their practice. Less relevant factors were availability of a smartphone used in the practice (mean = 2.53) and time until the examination result was available (mean = 2.29). Important technical factors were practicability of the device (mean = 1.27), unproblematic installation of any necessary software (mean = 1.34), and the integrability into the practice information system (mean = 1.44). Considering the patient welfare, physicians rated the accuracy of the examination, omission of pupil dilation, and the duration of the examination as the most important factors. Participants ranked the factors broadening the scope of care, strengthening the primary care (PC) range, and signs of modern medical practice as the most important factors for making an AI-based screening tool attractive for their practice. Conclusions These findings serve as a basis for a successful implementation of AI-assisted screening devices in PC and might facilitate early screenings for ophthalmological diseases in general practice. The most relevant barriers that need to be overcome for a successful implementation of such tools include clarification of the costs and reimbursement policies.