Automated Diabetic Retinopathy Image Assessment Software: Diagnostic Accuracy and Cost-Effectiveness Compared with Human Graders.

OBJECTIVE: With the increasing prevalence of diabetes, annual screening for diabetic retinopathy (DR) by expert human grading of retinal images is challenging. Automated DR image assessment systems (ARIAS) may provide clinically effective and cost-effective detection of retinopathy. We aimed to determine whether ARIAS can be safely introduced into DR screening pathways to replace human graders. DESIGN: Observational measurement comparison study of human graders following a national screening program for DR versus ARIAS. PARTICIPANTS: Retinal images from 20 258 consecutive patients attending routine annual diabetic eye screening between June 1, 2012, and November 4, 2013. METHODS: Retinal images were manually graded following a standard national protocol for DR screening and were processed by 3 ARIAS: iGradingM, Retmarker, and EyeArt. Discrepancies between manual grades and ARIAS results were sent to a reading center for arbitration. MAIN OUTCOME MEASURES: Screening performance (sensitivity, false-positive rate) and diagnostic accuracy (95% confidence intervals of screening-performance measures) were determined. Economic analysis estimated the cost per appropriate screening outcome. RESULTS: Sensitivity point estimates (95% confidence intervals) of the ARIAS were as follows: EyeArt 94.7% (94.2%-95.2%) for any retinopathy, 93.8% (92.9%-94.6%) for referable retinopathy (human graded as either ungradable, maculopathy, preproliferative, or proliferative), 99.6% (97.0%-99.9%) for proliferative retinopathy; Retmarker 73.0% (72.0 %-74.0%) for any retinopathy, 85.0% (83.6%-86.2%) for referable retinopathy, 97.9% (94.9%-99.1%) for proliferative retinopathy. iGradingM classified all images as either having disease or being ungradable. EyeArt and Retmarker saved costs compared with manual grading both as a replacement for initial human grading and as a filter prior to primary human grading, although the latter approach was less cost-effective. CONCLUSIONS: Retmarker and EyeArt systems achieved acceptable sensitivity for referable retinopathy when compared with that of human graders and had sufficient specificity to make them cost-effective alternatives to manual grading alone. ARIAS have the potential to reduce costs in developed-world health care economies and to aid delivery of DR screening in developing or remote health care settings.

An observational study to assess if automated diabetic retinopathy image assessment software can replace one or more steps of manual imaging grading and to determine their cost-effectiveness.

BACKGROUND: Diabetic retinopathy screening in England involves labour-intensive manual grading of retinal images. Automated retinal image analysis systems (ARIASs) may offer an alternative to manual grading. OBJECTIVES: To determine the screening performance and cost-effectiveness of ARIASs to replace level 1 human graders or pre-screen with ARIASs in the NHS diabetic eye screening programme (DESP). To examine technical issues associated with implementation. DESIGN: Observational retrospective measurement comparison study with a real-time evaluation of technical issues and a decision-analytic model to evaluate cost-effectiveness. SETTING: A NHS DESP. PARTICIPANTS: Consecutive diabetic patients who attended a routine annual NHS DESP visit. INTERVENTIONS: Retinal images were manually graded and processed by three ARIASs: iGradingM (version 1.1; originally Medalytix Group Ltd, Manchester, UK, but purchased by Digital Healthcare, Cambridge, UK, at the initiation of the study, purchased in turn by EMIS Health, Leeds, UK, after conclusion of the study), Retmarker (version 0.8.2, Retmarker Ltd, Coimbra, Portugal) and EyeArt (Eyenuk Inc., Woodland Hills, CA, USA). The final manual grade was used as the reference standard. Arbitration on a subset of discrepancies between manual grading and the use of an ARIAS by a reading centre masked to all grading was used to create a reference standard manual grade modified by arbitration. MAIN OUTCOME MEASURES: Screening performance (sensitivity, specificity, false-positive rate and likelihood ratios) and diagnostic accuracy [95% confidence intervals (CIs)] of ARIASs. A secondary analysis explored the influence of camera type and patients' ethnicity, age and sex on screening performance. Economic analysis estimated the cost per appropriate screening outcome identified. RESULTS: A total of 20,258 patients with 102,856 images were entered into the study. The sensitivity point estimates of the ARIASs were as follows: EyeArt 94.7% (95% CI 94.2% to 95.2%) for any retinopathy, 93.8% (95% CI 92.9% to 94.6%) for referable retinopathy and 99.6% (95% CI 97.0% to 99.9%) for proliferative retinopathy; and Retmarker 73.0% (95% CI 72.0% to 74.0%) for any retinopathy, 85.0% (95% CI 83.6% to 86.2%) for referable retinopathy and 97.9% (95% CI 94.9 to 99.1%) for proliferative retinopathy. iGradingM classified all images as either 'disease' or 'ungradable', limiting further iGradingM analysis. The sensitivity and false-positive rates for EyeArt were not affected by ethnicity, sex or camera type but sensitivity declined marginally with increasing patient age. The screening performance of Retmarker appeared to vary with patient's age, ethnicity and camera type. Both EyeArt and Retmarker were cost saving relative to manual grading either as a replacement for level 1 human grading or used prior to level 1 human grading, although the latter was less cost-effective. A threshold analysis testing the highest ARIAS cost per patient before which ARIASs became more expensive per appropriate outcome than human grading, when used to replace level 1 grader, was Retmarker £3.82 and EyeArt £2.71 per patient. LIMITATIONS: The non-randomised study design limited the health economic analysis but the same retinal images were processed by all ARIASs in this measurement comparison study. CONCLUSIONS: Retmarker and EyeArt achieved acceptable sensitivity for referable retinopathy and false-positive rates (compared with human graders as reference standard) and appear to be cost-effective alternatives to a purely manual grading approach. Future work is required to develop technical specifications to optimise deployment and address potential governance issues. FUNDING: The National Institute for Health Research (NIHR) Health Technology Assessment programme, a Fight for Sight Grant (Hirsch grant award) and the Department of Health's NIHR Biomedical Research Centre for Ophthalmology at Moorfields Eye Hospital and the University College London Institute of Ophthalmology.

Cohort profile: Examining Neighbourhood Activities in Built Living Environments in London: the ENABLE London-Olympic Park cohort.

PURPOSE: The Examining Neighbourhood Activities in Built Living Environments in London (ENABLE London) project is a natural experiment which aims to establish whether physical activity and other health behaviours show sustained changes among individuals and families relocating to East Village (formerly the London 2012 Olympics Athletes' Village), when compared with a control population living outside East Village throughout. PARTICIPANTS: Between January 2013 and December 2015, 1497 individuals from 1006 households were recruited and assessed (at baseline) (including 392 households seeking social housing, 421 seeking intermediate and 193 seeking market rent homes). The 2-year follow-up rate is 62% of households to date, of which 57% have moved to East Village. FINDINGS TO DATE: Assessments of physical activity (measured objectively using accelerometers) combined with Global Positioning System technology and Geographic Information System mapping of the local area are being used to characterise physical activity patterns and location among study participants and assess the attributes of the environments to which they are exposed. Assessments of body composition, based on weight, height and bioelectrical impedance, have been made and detailed participant questionnaires provide information on socioeconomic position, general health/health status, well-being, anxiety, depression, attitudes to leisure time activities and other personal, social and environmental influences on physical activity, including the use of recreational space and facilities in their residential neighbourhood. FUTURE PLANS: The main analyses will examine the changes in physical activity, health and well-being observed in the East Village group compared with controls and the influence of specific elements of the built environment on observed changes. The ENABLE London project exploits a unique opportunity to evaluate a 'natural experiment', provided by the building and rapid occupation of East Village. Findings from the study will be generalisable to other urban residential housing developments, and will help inform future evidence-based urban planning.

Prevalence of Active Long-term Problems in Patients With Anorectal Malformations: A Systematic Review.

BACKGROUND: Anorectal malformations are a spectrum of congenital anomalies of the rectum with high infantile survival rates and variable outcomes. Long-term (>10 years old) active problems associated with this condition have been poorly investigated. OBJECTIVE: The purpose of this review was to systematically define the prevalence of the most common active long-term problems in patients with a history of anorectal malformation repair. DATA SOURCES: MEDLINE, EMBASE, and the Cochrane Library were searched electronically using the OVID search platform. STUDY SELECTION: Original articles from August 1, 1994, to October 20, 2015, that included outcome data for patients aged ≥10 years with anorectal malformation. Cloaca was excluded from the study. INTERVENTIONS: Prevalence estimates of anorectal malformations were obtained from published articles. CIs were ascertained in the logit scale after transforming prevalence into log odds and were then transformed into the original scale. The same method was used for subgroup analysis investigating high and low anorectal malformations. MAIN OUTCOME MEASURES: The overall prevalences of fecal, urinary, and sexual dysfunction were analyzed. RESULTS: Twelve studies including 455 patients with a history of anorectal malformation repair were included for analysis. The range of reported prevalence of long-term active problems was as follows: fecal incontinence, 16.7% to 76.7%; chronic constipation, 22.2% to 86.7%; urinary incontinence, 1.7% to 30.5%; ejaculatory dysfunction, 15.6% to 41.2%; and erectile dysfunction, 5.6% to 11.8%. LIMITATIONS: The study was limited by its retrospective, small size; multiple complex associated anomalies often not reported; and heterogeneous composition of patients with limited stratification analysis. CONCLUSIONS: There is an overall high prevalence of active long-term issues in adolescents and young adults with anorectal malformations. Additional multicenter research is needed to define characteristics and predictors of long-term outcome, to implement effective follow-up, and to transition to adult health care.

Global variations and time trends in the prevalence of childhood myopia, a systematic review and quantitative meta-analysis: implications for aetiology and early prevention.

The aim of this review was to quantify the global variation in childhood myopia prevalence over time taking account of demographic and study design factors. A systematic review identified population-based surveys with estimates of childhood myopia prevalence published by February 2015. Multilevel binomial logistic regression of log odds of myopia was used to examine the association with age, gender, urban versus rural setting and survey year, among populations of different ethnic origins, adjusting for study design factors. 143 published articles (42 countries, 374 349 subjects aged 1-18 years, 74 847 myopia cases) were included. Increase in myopia prevalence with age varied by ethnicity. East Asians showed the highest prevalence, reaching 69% (95% credible intervals (CrI) 61% to 77%) at 15 years of age (86% among Singaporean-Chinese). Blacks in Africa had the lowest prevalence; 5.5% at 15 years (95% CrI 3% to 9%). Time trends in myopia prevalence over the last decade were small in whites, increased by 23% in East Asians, with a weaker increase among South Asians. Children from urban environments have 2.6 times the odds of myopia compared with those from rural environments. In whites and East Asians sex differences emerge at about 9 years of age; by late adolescence girls are twice as likely as boys to be myopic. Marked ethnic differences in age-specific prevalence of myopia exist. Rapid increases in myopia prevalence over time, particularly in East Asians, combined with a universally higher risk of myopia in urban settings, suggest that environmental factors play an important role in myopia development, which may offer scope for prevention.

Global variations and time trends in the prevalence of primary open angle glaucoma (POAG): a systematic review and meta-analysis.

Systematic review of published population based surveys to examine the relationship between primary open angle glaucoma (POAG) prevalence and demographic factors. A literature search identified population-based studies with quantitative estimates of POAG prevalence (to October 2014). Multilevel binomial logistic regression of log-odds of POAG was used to examine the effect of age and gender among populations of different geographical and ethnic origins, adjusting for study design factors. Eighty-one studies were included (37 countries, 216 214 participants, 5266 POAG cases). Black populations showed highest POAG prevalence, with 5.2% (95% credible interval (CrI) 3.7%, 7.2%) at 60 years, rising to 12.2% (95% CrI 8.9% to 16.6%) at 80 years. Increase in POAG prevalence per decade of age was greatest among Hispanics (2.31, 95% CrI 2.12, 2.52) and White populations (1.99, 95% CrI 1.86, 2.12), and lowest in East and South Asians (1.48, 95% CrI 1.39, 1.57; 1.56, 95% CrI 1.31, 1.88, respectively). Men were more likely to have POAG than women (1.30, 95% CrI 1.22, 1.41). Older studies had lower POAG prevalence, which was related to the inclusion of intraocular pressure in the glaucoma definition. Studies with visual field data on all participants had a higher POAG prevalence than those with visual field data on a subset. Globally 57.5 million people (95% CI 46.4 to 73.1 million) were affected by POAG in 2015, rising to 65.5 million (95% CrI 52.8, 83.2 million) by 2020. This systematic review provides the most precise estimates of POAG prevalence and shows omitting routine visual field assessment in population surveys may have affected case ascertainment. Our findings will be useful to future studies and healthcare planning.

Body mass index in early and middle adult life: prospective associations with myocardial infarction, stroke and diabetes over a 30-year period: the British Regional Heart Study.

OBJECTIVES: Adiposity in middle age is an established risk factor for cardiovascular disease and type 2 diabetes; less is known about the impact of adiposity from early adult life. We examined the effects of high body mass index (BMI) in early and middle adulthood on myocardial infarction (MI), stroke and diabetes risks. DESIGN: A prospective cohort study. PARTICIPANTS: 7735 men with BMI measured in middle age (40-59 years) and BMI ascertained at 21 years from military records or participant recall. PRIMARY AND SECONDARY OUTCOME MEASURES: 30-year follow-up data for type 2 diabetes, MI and stroke incidence; Cox proportional hazards models were used to examine the effect of BMI at both ages on these outcomes, adjusted for age and smoking status. RESULTS: Among 4846 (63%) men (with complete data), a 1 kg/m(2) higher BMI at 21 years was associated with a 6% (95% CI 4% to 9%) higher type 2 diabetes risk, compared with a 21% (95% CI 18% to 24%) higher diabetes risk for a 1 kg/m(2) higher BMI in middle age (hazard ratio (HR) 1.21, 95% CI 1.18 to 1.24). Higher BMI in middle age was associated with a 6% (95% CI 4% to 8%) increase in MI and a 4% (95% CI 1% to 7%) increase in stroke; BMI at 21 years showed no associations with MI or stroke risk. CONCLUSIONS: Higher BMI at 21 years of age is associated with later diabetes incidence but not MI or stroke, while higher BMI in middle age is strongly associated with all outcomes. Early obesity prevention may reduce later type 2 diabetes risk, more than MI and stroke.

Incidence of Late-Stage Age-Related Macular Degeneration in American Whites: Systematic Review and Meta-analysis.

PURPOSE: To estimate incidence of age-related macular degeneration (AMD) by subtype in American whites aged ≥50 years. DESIGN: Systematic review and meta-analysis. SETTING: Prospective cohort studies of AMD incidence in populations of white European ancestry published in MEDLINE, EMBASE, and Web of Science. STUDY POPULATION: Fourteen publications in 10 populations that examined AMD incident cases were identified. OBSERVATION PROCEDURE: Data on age-sex-specific incidence of late AMD, geographic atrophy (GA) and neovascular AMD (NVAMD), year of recruitment, AMD grading method, and continent were extracted. MAIN OUTCOME MEASURE(S): Annual incidence of late AMD, GA, and NVAMD by age-sex in American whites aged ≥50 years from a Bayesian meta-analysis of incidence studies was compared with incidence extrapolated from published prevalence estimates. RESULTS: Incidence rates from the review agreed with those derived from prevalence, but the latter were based on more data, especially at older ages and by AMD subtypes. Annual incidence (estimated from prevalence) of late AMD in American whites was 3.5 per 1000 aged ≥50 years (95% credible interval 2.5, 4.7 per 1000), equivalent to 293 000 new cases in American whites per year (95% credible interval 207 000, 400 000). Incidence rates approximately quadrupled per decade in age. Annual incidence GA rates were 1.9 per 1000 aged ≥50 years, NVAMD rates were 1.8 per 1000. Late AMD incidence was 38% higher in women vs men (95% credible interval 6%, 82%). CONCLUSIONS: Estimating AMD incidence from prevalence allows better characterization at older ages and by AMD subtype where longitudinal data from incidence studies are limited.

A study of whether automated Diabetic Retinopathy Image Assessment could replace manual grading steps in the English National Screening Programme.

OBJECTIVES: Diabetic retinopathy screening in England involves labour intensive manual grading of digital retinal images. We present the plan for an observational retrospective study of whether automated systems could replace one or more steps of human grading. METHODS: Patients aged 12 or older who attended the Diabetes Eye Screening programme, Homerton University Hospital (London) between 1 June 2012 and 4 November 2013 had macular and disc-centred retinal images taken. All screening episodes were manually graded and will additionally be graded by three automated systems. Each system will process all screening episodes, and screening performance (sensitivity, false positive rate, likelihood ratios) and diagnostic accuracy (95% confidence intervals of screening performance measures) will be quantified. A sub-set of gradings will be validated by an approved Reading Centre. Additional analyses will explore the effect of altering thresholds for disease detection within each automated system on screening performance. RESULTS: 2,782/20,258 diabetes patients were referred to ophthalmologists for further examination. Prevalence of maculopathy (M1), pre-proliferative retinopathy (R2), and proliferative retinopathy (R3) were 7.9%, 3.1% and 1.2%, respectively; 4749 (23%) patients were diagnosed with background retinopathy (R1); 1.5% were considered ungradable by human graders. CONCLUSIONS: Retinopathy prevalence was similar to other English diabetic screening programmes, so findings should be generalizable. The study population size will allow the detection of differences in screening performance between the human and automated grading systems as small as 2%. The project will compare performance and economic costs of manual versus automated systems.

Esophageal atresia and transitional care--step 1: a systematic review and meta-analysis of the literature to define the prevalence of chronic long-term problems.

BACKGROUND: Esophageal atresia (EA) is a rare congenital anomaly with high infantile survival rates. The aim of this study was to outline the prevalence of common long-term problems associated with EA repair in patients older than 10 years of age. DATA SOURCES: Original papers were identified by systematic searching of MEDLINE and EMBASE databases from January 1993 to July 2014. Fifteen articles (907 EA patients) met inclusion criteria. CONCLUSIONS: This is the first systematic review aiming to quantify the prevalence of the long-term problems associated with EA. The main active medical conditions (pooled estimated prevalence) identified were the following: dysphagia (50.3%), gastroesophageal reflux disease with (40.2%) or without (56.5%) histological esophagitis, recurrent respiratory tract infections (24.1%), doctor-diagnosed asthma (22.3%), persistent cough (14.6%), and wheeze (34.7%). The prevalence of Barrett's esophagus (6.4%) was 4 and 26 times higher than the adult (1.6%) and pediatric (.25%) general populations. Adult and pediatric practitioners should focus on how to develop effective long-term follow-up and transitional care for these patients.

Adiposity in early, middle and later adult life and cardiometabolic risk markers in later life; findings from the British regional heart study.

OBJECTIVES: This research investigates the associations between body mass index (BMI) at 21, 40-59, 60-79 years of age on cardiometabolic risk markers at 60-79 years. METHODS: A prospective study of 3464 British men with BMI measured at 40-59 and 60-79 years, when cardiometabolic risk was assessed. BMI at 21 years was ascertained from military records, or recalled from middle-age (adjusted for reporting bias); associations between BMI at different ages and later cardiometabolic risk markers were examined using linear regression. Sensitive period, accumulation and mobility life course models were devised for high BMI (defined as BMI≥75th centile) and compared with a saturated BMI trajectory model. RESULTS: At ages 21, 40-59 and 60-79 years, prevalences of overweight (BMI≥25 kg/m2) were 12%, 53%, 70%, and obesity (≥30 kg/m2) 1.6%, 6.6%, and 17.6%, respectively. BMI at 21 years was positively associated with serum insulin, blood glucose, and HbA1c at 60-79 years, with increases of 1.5% (95%CI 0.8,2.3%), 0.4% (0.1,0.6%), 0.3% (0.1,0.4%) per 1 kg/m2, respectively, but showed no associations with blood pressure or blood cholesterol. However, these associations were modest compared to those between BMI at 60-79 years and serum insulin, blood glucose and HbA1c at 60-79 years, with increases of 8.6% (8.0,9.2%), 0.7% (0.5,0.9%), and 0.5% (0.4,0.7%) per 1 kg/m2, respectively. BMI at 60-79 years was also associated with total cholesterol and blood pressure. Associations for BMI at 40-59 years were mainly consistent with those of BMI at 60-79 years. None of the life course models fitted the data as well as the saturated model for serum insulin. A sensitive period at 50 years for glucose and HbA1c and sensitive period at 70 years for blood pressure were identified. CONCLUSIONS: In this cohort of men who were thin compared to more contemporary cohorts, BMI in later life was the dominant influence on cardiovascular and diabetes risk. BMI in early adult life may have a small long-term effect on diabetes risk.