Multicenter, Head-to-Head, Real-World Validation Study of Seven Automated Artificial Intelligence Diabetic Retinopathy Screening Systems.

OBJECTIVE: With rising global prevalence of diabetic retinopathy (DR), automated DR screening is needed for primary care settings. Two automated artificial intelligence (AI)-based DR screening algorithms have U.S. Food and Drug Administration (FDA) approval. Several others are under consideration while in clinical use in other countries, but their real-world performance has not been evaluated systematically. We compared the performance of seven automated AI-based DR screening algorithms (including one FDA-approved algorithm) against human graders when analyzing real-world retinal imaging data. RESEARCH DESIGN AND METHODS: This was a multicenter, noninterventional device validation study evaluating a total of 311,604 retinal images from 23,724 veterans who presented for teleretinal DR screening at the Veterans Affairs (VA) Puget Sound Health Care System (HCS) or Atlanta VA HCS from 2006 to 2018. Five companies provided seven algorithms, including one with FDA approval, that independently analyzed all scans, regardless of image quality. The sensitivity/specificity of each algorithm when classifying images as referable DR or not were compared with original VA teleretinal grades and a regraded arbitrated data set. Value per encounter was estimated. RESULTS: Although high negative predictive values (82.72-93.69%) were observed, sensitivities varied widely (50.98-85.90%). Most algorithms performed no better than humans against the arbitrated data set, but two achieved higher sensitivities, and one yielded comparable sensitivity (80.47%, P = 0.441) and specificity (81.28%, P = 0.195). Notably, one had lower sensitivity (74.42%) for proliferative DR (P = 9.77 × 10-4) than the VA teleretinal graders. Value per encounter varied at $15.14-$18.06 for ophthalmologists and $7.74-$9.24 for optometrists. CONCLUSIONS: The DR screening algorithms showed significant performance differences. These results argue for rigorous testing of all such algorithms on real-world data before clinical implementation.

Uterotonic drugs to prevent postpartum haemorrhage: a network meta-analysis.

BACKGROUND: Postpartum haemorrhage (PPH) is the leading cause of maternal mortality worldwide. Prophylactic uterotonic drugs can reduce blood loss and are routinely recommended. There are several uterotonic drugs for preventing PPH, but it is still debatable which drug or combination of drugs is the most effective. OBJECTIVES: To identify the most effective and cost-effective uterotonic drug(s) to prevent PPH, and generate a ranking according to their effectiveness and side-effect profile. METHODS: The Cochrane Pregnancy and Childbirth's Trials Register (1 June 2015), ClinicalTrials.gov and the World Health Organization (WHO)'s International Clinical Trials Registry Platform (ICTRP) were searched for unpublished trial reports (30 June 2015). In addition, reference lists of retrieved studies (updated October 2017) were searched for randomised trials evaluating uterotonic drugs for preventing PPH. The study estimated relative effects and rankings for preventing PPH, defined as blood loss of ≥ 500 ml and ≥ 1000 ml. Pairwise meta-analyses and network meta-analysis were performed to determine the relative effects and rankings of all available drugs and combinations thereof [ergometrine, misoprostol (Cytotec®; Pfizer Inc., New York, NY, USA), misoprostol plus oxytocin (Syntocinon®; Novartis International AG, Basel, Switzerland), carbetocin (Pabal®; Ferring Pharmaceuticals, Saint-Prex, Switzerland), ergometrine plus oxytocin (Syntometrine®; Alliance Pharma plc, Chippenham, UK), oxytocin, and a placebo or no treatment]. Primary outcomes were stratified according to the mode of birth, prior risk of PPH, health-care setting, drug dosage, regimen and route of drug administration. Sensitivity analyses were performed according to study quality and funding source, among others. A model-based economic evaluation compared the relative cost-effectiveness separately for vaginal births and caesareans with or without including side effects. RESULTS: From 137 randomised trials and 87,466 women, ergometrine plus oxytocin, carbetocin and misoprostol plus oxytocin were found to reduce the risk of PPH blood loss of ≥ 500 ml compared with the standard drug, oxytocin [ergometrine plus oxytocin: risk ratio (RR) 0.69, 95% confidence interval (CI) 0.57 to 0.83; carbetocin: RR 0.72, 95% CI 0.52 to 1.00; misoprostol plus oxytocin: RR 0.73, 95% CI 0.6 to 0.9]. Each of these three strategies had 100% cumulative probability of being ranked first, second or third most effective. Oxytocin was ranked fourth, with an almost 0% cumulative probability of being ranked in the top three. Similar rankings were noted for the reduction of PPH blood loss of ≥ 1000 ml (ergometrine plus oxytocin: RR 0.77, 95% CI 0.61 to 0.95; carbetocin: RR 0.70, 95% CI 0.38 to 1.28; misoprostol plus oxytocin: RR 0.90, 95% CI 0.72 to 1.14), and most secondary outcomes. Ergometrine plus oxytocin and misoprostol plus oxytocin had the poorest ranking for side effects. Carbetocin had a favourable side-effect profile, which was similar to oxytocin. However, the analysis was restricted to high-quality studies, carbetocin lost its ranking and was comparable to oxytocin. The relative cost-effectiveness of the alternative strategies is inconclusive, and the results are affected by both the uncertainty and inconsistency in the data reported on adverse events. For vaginal delivery, when assuming no adverse events, ergometrine plus oxytocin is less costly and more effective than all strategies except carbetocin. The strategy of carbetocin is both more effective and more costly than all other strategies. When taking adverse events into consideration, all prevention strategies, except oxytocin, are more costly and less effective than carbetocin. For delivery by caesarean section, with and without adverse events, the relative cost-effectiveness is different, again because of the uncertainty in the available data. LIMITATIONS: There was considerable uncertainty in findings within the planned subgroup analyses, and subgroup effects cannot be ruled out. CONCLUSIONS: Ergometrine plus oxytocin, carbetocin and misoprostol plus oxytocin are more effective uterotonic drug strategies for preventing PPH than the current standard, oxytocin. Ergometrine plus oxytocin and misoprostol plus oxytocin cause significant side effects. Carbetocin has a favourable side-effect profile, which was similar to oxytocin. However, most carbetocin trials are small and of poor quality. There is a need for a large high-quality trial comparing carbetocin with oxytocin; such a trial is currently being conducted by the WHO. The relative cost-effectiveness is inconclusive, and results are affected by uncertainty and inconsistency in adverse events data. STUDY REGISTRATION: This study is registered as PROSPERO CRD42015020005; Cochrane Pregnancy and Childbirth Group (substudy) reference number 0871; PROSPERO-Cochrane (substudy) reference number CRD42015026568; and sponsor reference number ERN_13-1414 (University of Birmingham, Birmingham, UK). FUNDING: Funding for this study was provided by the National Institute for Health Research Health Technology Assessment programme in a research award to the University of Birmingham and supported by the UK charity Ammalife (UK-registered charity 1120236). The funders of the study had no role in study design, data collection, data synthesis, interpretation or writing of the report.

Postpartum haemorrhage (PPH) is the most common reason why mothers die in childbirth worldwide. Although most healthy women can cope well with blood loss after birth, some do not, and this can pose a serious risk to their health and even life. To reduce blood loss after birth, the routine administration of a drug to contract the uterus (uterotonic) has become standard practice across the world. This research seeks to identify which is the most effective and cost-effective drug. Different drugs have been used for reducing the occurrence of PPH. They include oxytocin, misoprostol, ergometrine, carbetocin, and combinations of these drugs, each with different effectiveness and side effects. The study synthesised the available evidence to compare all of these drugs and combinations thereof. After putting the results of all available comparisons together in a network, a ranking among them was calculated, and provided robust effectiveness and side-effect profiles for each drug and their associated costs. The study included 137 randomised trials, involving a total of 87,466 women. The results suggested that ergometrine plus oxytocin, carbetocin and misoprostol plus oxytocin are the most effective strategies for preventing PPH and are more effective than the currently recommended drug, oxytocin. Each of these three strategies had almost 100% probability of being ranked first, second or third most effective. Oxytocin was ranked fourth with an almost 0% probability of being ranked in the top three. Ergometrine plus oxytocin and misoprostol plus oxytocin were the worst drug combinations for side effects, with carbetocin having the most favourable side-effect profile. Carbetocin could prevent approximately one further event of PPH out of three in comparison with oxytocin. However, existing carbetocin studies were small and of poor quality. There is need for a large high-quality study comparing carbetocin with the current standard treatment of oxytocin for the prevention of PPH. The cost analyses of the alternative drug strategies remain inconclusive.

How can we teach EBM in clinical practice? An analysis of barriers to implementation of on-the-job EBM teaching and learning.

INTRODUCTION: Evidence-based medicine (EBM) improves the quality of health care. Courses on how to teach EBM in practice are available, but knowledge does not automatically imply its application in teaching. We aimed to identify and compare barriers and facilitators for teaching EBM in clinical practice in various European countries. METHODS: A questionnaire was constructed listing potential barriers and facilitators for EBM teaching in clinical practice. Answers were reported on a 7-point Likert scale ranging from not at all being a barrier to being an insurmountable barrier. RESULTS: The questionnaire was completed by 120 clinical EBM teachers from 11 countries. Lack of time was the strongest barrier for teaching EBM in practice (median 5). Moderate barriers were the lack of requirements for EBM skills and a pyramid hierarchy in health care management structure (median 4). In Germany, Hungary and Poland, reading and understanding articles in English was a higher barrier than in the other countries. CONCLUSION: Incorporation of teaching EBM in practice faces several barriers to implementation. Teaching EBM in clinical settings is most successful where EBM principles are culturally embedded and form part and parcel of everyday clinical decisions and medical practice.

Teaching trainers to incorporate evidence-based medicine (EBM) teaching in clinical practice: the EU-EBM project.

BACKGROUND: Evidence based medicine (EBM) is considered an integral part of medical training, but integration of teaching various EBM steps in everyday clinical practice is uncommon. Currently EBM is predominantly taught through theoretical courses, workshops and e-learning. However, clinical teachers lack confidence in teaching EBM in workplace and are often unsure of the existing opportunities for teaching EBM in the clinical setting. There is a need for continuing professional development (CPD) courses that train clinical trainers to teach EBM through on-the-job training by demonstration of applied EBM real time in clinical practice. We developed such a course to encourage clinically relevant teaching of EBM in post-graduate education in various clinical environments. METHODS: We devised an e-learning course targeting trainers with EBM knowledge to impart educational methods needed to teach application of EBM teaching in commonly used clinical settings. The curriculum development group comprised experienced EBM teachers, clinical epidemiologists, clinicians and educationalists from institutions in seven European countries. The e-learning sessions were designed to allow participants (teachers) to undertake the course in the workplace during short breaks within clinical activities. An independent European steering committee provided input into the process. RESULTS: The curriculum defined specific learning objectives for teaching EBM by exploiting educational opportunities in six different clinical settings. The e-modules incorporated video clips that demonstrate practical and effective methods of EBM teaching in everyday clinical practice. The course encouraged focussed teaching activities embedded within a trainer's personal learning plan and documentation in a CPD portfolio for reflection. CONCLUSION: This curriculum will help senior clinicians to identify and make the best use of available opportunities in everyday practice in clinical situations to teach various steps of EBM and demonstrate their applicability to clinical practice. Once fully implemented, the ultimate outcome of this pilot project will be a European qualification in teaching EBM, which will be used by doctors, hospitals, professional bodies responsible for postgraduate qualifications and continuing medical education.

Progesterone for the prevention of preterm birth: a critical evaluation of evidence.

A systematic review of the literature identified nine randomised trials that evaluated the effects of progestational agents in the prevention of preterm delivery. These studies were of variable quality. Meta-analyses showed reductions in delivery rates before 37 weeks (OR 0.42, 95% CI 0.31-0.57) and 34 weeks (OR 0.51, 95% CI 0.34-0.77) as well as in respiratory distress syndrome (OR 0.55, 95% CI 0.31-0.96) with progestational agents. A cumulative meta-analysis showed that the treatment benefit for the outcome of delivery before 37 weeks exceeded the conventional level of statistical significance in 1975 (p<0.01); by 1985, the p-value was <0.001, and by 2003, it was <0.0001. Another cumulative meta-analysis in which the studies were added to the pooled analysis by decreasing quality score showed significant benefit even when the analysis was limited to just the highest quality trials (OR 0.47, 95% CI 0.33, 0.66, p<0.0001). An exploration of the applicability of the effects across various baseline risks using a L'abbe plot found that the benefit was consistent across a range of risks. A comprehensive review of both trial and observational data on harm did not show any demonstrable evidence of harm to mother and baby. Women at high risk of preterm birth should be recommended progestational agent therapy.

Effectiveness of nifedipine versus atosiban for tocolysis in preterm labour: a meta-analysis with an indirect comparison of randomised trials.

OBJECTIVE: To explore the effectiveness of nifedipine compared with atosiban for tocolysis in preterm labour. DESIGN: A systematic review of randomised controlled trials with meta-analysis using adjusted indirect comparison. POPULATION: Six hundred and seventy-nine women recruited in nine randomised trials evaluating the effectiveness of nifedipine versus beta-agonists, and 852 women recruited in four trials of atosiban versus beta-agonists. There were no trials comparing nifedipine directly with atosiban. METHODS: We performed meta-analysis with a technique involving an adjusted indirect comparison between nifedipine and atosiban using beta-agonists as the common comparator. This approach preserves the benefit accrued by randomisation in the original comparisons. MAIN OUTCOME MEASURES: Reduction in neonatal respiratory distress syndrome and delay in delivery by 48 hours. RESULTS: Nifedipine tocolysis was associated with a significant reduction in respiratory distress syndrome compared with atosiban (OR 0.55, 95% CI 0.32-0.97). It also increased the number of women whose delivery was delayed by 48 hours (OR 1.20, 95% CI 0.73-1.95), although this result was not statistically significant. CONCLUSIONS: When indirectly compared with atosiban, nifedipine tocolysis is more effective. In the absence of a direct comparison, our analysis provides a way to explore the potential benefits of nifedipine versus atosiban.

Incorporating the views of obstetric clinicians in implementing evidence-supported labour and delivery suite ward rounds: a case study.

OBJECTIVES: To implement evidence-supported delivery suite ward rounds using strategies to overcome barriers. DESIGN: Qualitative case study to identify barriers and a quantitative assessment of outcomes on ward rounds. METHOD: After an introduction to the proposed evidence-supported delivery suite ward round reactions were elicited from 15 clinicians. Responses were categorized into barriers related to knowledge, skills, attitudes and behaviour. Using strategies to overcome these barriers, a weekly evidence-supported ward round was established. During patient-centred discussions, clinical questions were formulated. Following literature searches, evidence was acquired and appraised. RESULTS: Clinicians had limited awareness about the potential benefits of a clinical librarian's support during ward rounds. They were unsure of their skills in evidence-based practice. There was scepticism, fear of loss of autonomy, poor motivation and resistance to change. Clinicians identified several factors that could impinge on their practice. Using appropriate strategies when evidence-supported ward rounds were conducted, 67 questions were raised. However only seven articles were appraised in time to affect decisions on that ward round. CONCLUSIONS: Knowledge of barriers was helpful in developing plans for implementation of evidence-supported ward rounds. The pilot study indicates the feasibility of this approach.

Aspirin for prevention of preeclampsia in women with historical risk factors: a systematic review.

OBJECTIVE: To examine the effectiveness of aspirin in preventing perinatal death and preeclampsia in women with predisposing historical risk factors, such as previous history of preeclampsia, chronic hypertension, diabetes, and renal disease. DATA SOURCES: Searches were conductes in Medline, Embase, Cochrane Library, National Research Register, SCISEARCH, AND ISI Conference Proceedings without any language restrictions, using the following medical subject headings and text words: "aspirin," "antiplatelet*," "salicyl*," "acetylsalicyl*," "platelet aggregation inhibitors," "pre-eclamp*," "preeclamp*," and "hypertens*. METHODS OF STUDY SELECTION: We included all randomized trials that evaluated the effectiveness of aspirin compared with placebo or no treatment in women with predisposing historical risk factors and reported clinically relevant perinatal or maternal outcomes. Study selection, quality appraisal, and data extractions were performed independently and in duplicate.We identified 14 relevant trials, including a total of 12,416 women. Meta-analysis showed a significant benefit of aspirin therapy in reducing perinatal death (odds ratio [OR] 0.79, 95% confidence interval [CI] 0.64, 0.96) and preeclampsia (OR 0.86, 95% CI 0.76, 0.96). Aspirin was also associated with a reduction in rates of spontaneous preterm birth (OR 0.86, 95% CI 0.79, 0.94), and an increase of 215 g in mean birth weight (weighted mean difference 215, 95% CI 90, 341). There was no increase in the risk of placental abruption with aspirin (OR 0.98, 95% CI 0.79, 1.21). Funnel plot analysis indicated that publication and related biases were unlikely (Egger test, P =.84). CONCLUSION: Aspirin reduces the risk of perinatal death and preeclampsia in women with historical risk factors. Given the importance of these outcomes and the safety and low cost of aspirin, aspirin therapy should be considered in women with historical risk factors.

Women's attitudes to HIV screening in pregnancy in an area of low prevalence.

OBJECTIVES: To identify factors that influence acceptance of the human immunodeficiency virus (HIV) screening test by pregnant women. DESIGN: Prospective survey using questionnaires. SETTING: An antenatal clinic at a tertiary referral hospital. POPULATION: 200 pregnant women of multiethnic origin residing in a low prevalence area. METHODS: Two hundred consecutive women attending the antenatal booking clinic were interviewed. We collected data on reasons for opting in or out and factors that could influence the rate of acceptance, such as ethnicity, age, parity, level of education, first language and marital or relationship status. MAIN OUTCOME MEASURES: Attitudes of pregnant women to HIV screening. The rate of acceptance of HIV screening was 160/200(80%). The most frequent reasons for opting in were the perceived benefits of screening (92/160, 58%) and the fact that the test is now routinely offered (87/160, 54%). The most frequent reason for refusal of screening was perceived low risk of HIV (27/40, 68%). Logistic regression analysis showed that ethnicity, age, parity and level of formal education did not play a role in influencing uptake of the test. However, women whose first language was English were more likely to opt in compared with those whose first language was not English (P = 0.016) and those who were married or in a stable relationship were more likely to opt in than those who were not (P < 0.001). CONCLUSION: We have not reached the national target of 90% in our region. This appears to be due to a combination of factors, the main one being a perception of low risk. Greater knowledge of the ease of transmission of HIV may alter this perception, which may improve acceptance of the test.

Oxytocin antagonists for tocolysis in preterm labour -- a systematic review.

BACKGROUND: Oxytocin antagonists have been shown to inhibit uterine contractions and delay preterm delivery. Our objective was to examine the clinical effectiveness and safety of oxytocin antagonists for tocolysis in preterm labour. MATERIAL/METHODS: We searched MEDLINE, EMBASE, the Cochrane Controlled Trials Register and Science Citation Index using the following Medical Subject Headings and textwords: oxytocin (antagonists and inhibitors), atosiban, antocin, oxytocin antagonists, oxytocin receptor antagonists and oxytocin inhibitors. All randomised controlled trials that compared effectiveness and safety of atosiban with a placebo or another tocolytic in women with threatened or actual preterm labour were included. The primary outcome measure was the proportion of women undelivered by 48 hours from the commencement of treatment. RESULTS: Six articles met the inclusion criteria - two compared atosiban to placebo and four atosiban to a beta-agonist. Meta-analysis showed a significant increase in the proportion of women undelivered by 48 hours in women receiving atosiban compared to placebo (RR 1.13, 95%CI 1.02, 1.26). When compared with beta-agonists, atosiban increased the proportion of women undelivered by 48 hours, but this trend did not reach statistical significance (RR 1.07, 95%CI 0.98, 1.17). Side effect profile was substantially better for atosiban compared to beta-agonists. CONCLUSIONS: Oxytocin antagonists appear to be effective and safe for tocolysis in preterm labour.

Prediction of vaginal delivery following caesarean section for failure to progress based on the initial aberrant labour pattern.

OBJECTIVE: To determine whether the pattern of cervical dilatation prior to caesarean section for 'failure to progress' can predict the outcome of subsequent labour. STUDY DESIGN: Single hospital case note review of 171 women delivered by caesarean section for failure to progress and subsequently delivering at the same hospital. Cervicograms were categorized into one of the four patterns by an assessor blinded to the subsequent outcome. Statistical analysis was done by analysis of variance. RESULTS: The incidence of vaginal delivery did not significantly differ between the groups. CONCLUSION: Categorisation of failure to progress by partographic abnormality does not predict subsequent successful vaginal delivery.