Purpose: To assess the diagnostic accuracy measures such as sensitivity and specificity of smartphone-based artificial intelligence (AI) approaches in the detection of diabetic retinopathy (DR). Methods: A literature search of the EMBASE and MEDLINE databases (up to March 2020) was conducted. Only studies using both smartphone-based cameras and AI software for image analysis were included. The main outcome measures were pooled sensitivity and specificity, diagnostic odds ratios and relative risk of smartphone-based AI approaches in detecting DR (of all types), and referable DR (RDR) (moderate nonproliferative retinopathy or worse and/or the presence of diabetic macular edema). Results: Smartphone-based AI has a pooled sensitivity of 89.5% (95% confidence interval [CI]: 82.3%–94.0%) and pooled specificity of 92.4% (95% CI: 86.4%–95.9%) in detecting DR. For referable disease, sensitivity is 97.9% (95% CI: 92.6%-99.4%), and the pooled specificity is 85.9% (95% CI: 76.5%–91.9%). The technology is better at correctly identifying referable retinopathy. Conclusions: The smartphone-based AI programs demonstrate high diagnostic accuracy for the detection of DR and RDR and are potentially viable substitutes for conventional diabetic screening approaches. Further, high-quality randomized controlled trials are required to establish the effectiveness of this approach in different populations.

Introduction and objectivesInfluenza vaccination is offered annually in the UK to high-risk individuals such as those with asthma as a preventive measure against influenza infection and influenza-related complications. However, the effectiveness and safety of influenza vaccination in people with asthma is not well established.1 MethodsWe conducted a systematic review and meta-analysis assessing the overall quality of evidence using the GRADE methodology. Published literature was searched through 13 electronic databases from Jan 1970 to Jan 2016 for clinical trials and epidemiological studies. Unpublished or ongoing literature was searched through references and citations of key publications, and by contacting influenza vaccine manufacturers. The screening for eligible studies, data extraction and quality appraisal was conducted by two reviewers independently. Separate meta-analyses were undertaken for observational and experimental evidence using random-effects models.ResultsWe identified 35 eligible studies, and four contributed to the meta-analyses. Risk of bias was high for one randomised controlled trial (RCT), unclear for 11 RCTs, and low for eight RCTs. The quality of five non-RCTs, four cohorts, and two case-control studies was strong. Moderate quality was found for one non-RCT, and three cohort studies. In people with asthma, pooled vaccine effectiveness (VE) was 45% (OR: 0.55; 95% CI: 0.44 to 0.69; I2 = 0%) for laboratory confirmed influenza. Pooled effectiveness of live vaccines was 81% (RR: 0.19; 95% CI: 0.06 to 0.67; I2 = 0%) for influenza infection (confirmed by cell culture or rise in antibody titre) and 72% (RR: 0.28; 95%: 0.10 to 0.80; I2 = 0%) for influenza-like illness. VE was also observed against asthma attacks. No increased risk of vaccine-induced asthma symptoms and attacks was identified. The quality of the body of evidence was considered very low for all outcomes.ConclusionsEvidence on VE in people with asthma against influenza, asthma exacerbations, and other clinical outcomes is limited and of very low quality. Thus, better quality evidence is required, especially in adults with asthma. Vaccination with inactivated or live vaccines was found to be safe and well tolerated in patients with asthma.ReferenceCates CJ, Rowe BH. Vaccines for preventing influenza in people with asthma. Cochrane Database Sys Rev 2013;2:Cd000364.