The volume and complexity of diagnostic imaging is increasing at a pace faster than the availability of human expertise to interpret it. Artificial intelligence has shown great promise in classifying two-dimensional photographs of some common diseases and typically relies on databases of millions of annotated images. Until now, the challenge of reaching the performance of expert clinicians in a real-world clinical pathway with three-dimensional diagnostic scans has remained unsolved. Here, we apply a novel deep learning architecture to a clinically heterogeneous set of three-dimensional optical coherence tomography scans from patients referred to a major eye hospital. We demonstrate performance in making a referral recommendation that reaches or exceeds that of experts on a range of sight-threatening retinal diseases after training on only 14,884 scans. Moreover, we demonstrate that the tissue segmentations produced by our architecture act as a device-independent representation; referral accuracy is maintained when using tissue segmentations from a different type of device. Our work removes previous barriers to wider clinical use without prohibitive training data requirements across multiple pathologies in a real-world setting.

Two decades after the Safe Motherhood campaign's 1987 launch in India, half a million women continue to die from pregnancy-related causes every year. Key health-care interventions can largely prevent these deaths, but their use is limited in developing countries, and is reported to vary between population groups. We reviewed the use of maternal health-care interventions in developing countries to assess the extent, strength and implications of evidence for variations according to women's place of residence and socioeconomic status. Studies with data on use of a skilled health worker at delivery, antenatal care in the first trimester of pregnancy and medical settings for delivery were assessed. We identified 30 eligible studies, 12 of which were of high or moderate quality, from 23 countries. Results of these studies showed wide variation in use of maternal health care. Methodological factors (e.g. inaccurate identification of population in need or range of potential confounders controlled for) played a part in this variation. Differences were also caused by factors related to health-care users (e.g. age, education, medical insurance, clinical risk factors) or to supply of health care (e.g. clinic availability, distance to facility), or by an interaction between such factors (e.g. perceived quality of care). Variation was usually framed by contextual issues relating to funding and organization of health care or social and cultural issues. These findings emphasize the need to investigate and assess context-specific causes of varying use of maternal health care, if safe motherhood is to become a reality in developing countries.

Overall uptake rates in this organized screening programme were encouraging, but nonetheless there was low uptake in the most ethnically diverse areas and a striking gradient by SES. Action to promote equality of uptake is needed to avoid widening inequalities in cancer mortality.