Summary Background Radical measures are required to identify and reduce blindness due to diabetes to achieve the Sustainable Development Goals by 2030. Therefore, we evaluated the accuracy of an artificial intelligence (AI) model using deep learning in a population-based diabetic retinopathy screening programme in Zambia, a lower-middle-income country. Methods We adopted an ensemble AI model consisting of a combination of two convolutional neural networks (an adapted VGGNet architecture and a residual neural network architecture) for classifying retinal colour fundus images. We trained our model on 76 370 retinal fundus images from 13 099 patients with diabetes who had participated in the Singapore Integrated Diabetic Retinopathy Program, between 2010 and 2013, which has been published previously. In this clinical validation study, we included all patients with a diagnosis of diabetes that attended a mobile screening unit in five urban centres in the Copperbelt province of Zambia from Feb 1 to June 31, 2012. In our model, referable diabetic retinopathy was defined as moderate non-proliferative diabetic retinopathy or worse, diabetic macular oedema, and ungradable images. Vision-threatening diabetic retinopathy comprised severe non-proliferative and proliferative diabetic retinopathy. We calculated the area under the curve (AUC), sensitivity, and specificity for referable diabetic retinopathy, and sensitivities of vision-threatening diabetic retinopathy and diabetic macular oedema compared with the grading by retinal specialists. We did a multivariate analysis for systemic risk factors and referable diabetic retinopathy between AI and human graders. Findings A total of 4504 retinal fundus images from 3093 eyes of 1574 Zambians with diabetes were prospectively recruited. Referable diabetic retinopathy was found in 697 (22·5%) eyes, vision-threatening diabetic retinopathy in 171 (5·5%) eyes, and diabetic macular oedema in 249 (8·1%) eyes. The AUC of the AI system for referable diabetic retinopathy was 0·973 (95% CI 0·969–0·978), with corresponding sensitivity of 92·25% (90·10–94·12) and specificity of 89·04% (87·85–90·28). Vision-threatening diabetic retinopathy sensitivity was 99·42% (99·15–99·68) and diabetic macular oedema sensitivity was 97·19% (96·61–97·77). The AI model and human graders showed similar outcomes in referable diabetic retinopathy prevalence detection and systemic risk factors associations. Both the AI model and human graders identified longer duration of diabetes, higher level of glycated haemoglobin, and increased systolic blood pressure as risk factors associated with referable diabetic retinopathy. Interpretation An AI system shows clinically acceptable performance in detecting referable diabetic retinopathy, vision-threatening diabetic retinopathy, and diabetic macular oedema in population-based diabetic retinopathy screening. This shows the potential application and adoption of such AI technology in an under-resourced African population to reduce the incidence of preventable blindness, even when the model is trained in a different population. Funding National Medical Research Council Health Service Research Grant, Large Collaborative Grant, Ministry of Health, Singapore; the SingHealth Foundation; and the Tanoto Foundation.

Artificial intelligence using deep learning to screen for referable and vision-threatening diabetic retinopathy in Africa: a clinical validation study.

Background 
Despite the sharp decline in global under-5 deaths since 1990, uneven progress has been achieved across and within countries. In sub-Saharan Africa (SSA), the Millennium Development Goals (MDGs) for child mortality were met only by a few countries. Valid concerns exist as to whether the region would meet new Sustainable Development Goals (SDGs) for under-5 mortality. We therefore examine further sources of variation by assessing age patterns, trends, and forecasts of mortality rates.


Methods and findings 
Data came from 106 nationally representative Demographic and Health Surveys (DHSs) with full birth histories from 31 SSA countries from 1990 to 2017 (a total of 524 country-years of data). We assessed the distribution of age at death through the following new demographic analyses. First, we used a direct method and full birth histories to estimate under-5 mortality rates (U5MRs) on a monthly basis. Second, we smoothed raw estimates of death rates by age and time by using a two-dimensional P-Spline approach. Third, a variant of the Lee–Carter (LC) model, designed for populations with limited data, was used to fit and forecast age profiles of mortality. We used mortality estimates from the United Nations Inter-agency Group for Child Mortality Estimation (UN IGME) to adjust, validate, and minimize the risk of bias in survival, truncation, and recall in mortality estimation. Our mortality model revealed substantive declines of death rates at every age in most countries but with notable differences in the age patterns over time. U5MRs declined from 3.3% (annual rate of reduction [ARR] 0.1%) in Lesotho to 76.4% (ARR 5.2%) in Malawi, and the pace of decline was faster on average (ARR 3.2%) than that observed for infant (IMRs) (ARR 2.7%) and neonatal (NMRs) (ARR 2.0%) mortality rates. We predict that 5 countries (Kenya, Rwanda, Senegal, Tanzania, and Uganda) are on track to achieve the under-5 sustainable development target by 2030 (25 deaths per 1,000 live births), but only Rwanda and Tanzania would meet both the neonatal (12 deaths per 1,000 live births) and under-5 targets simultaneously. Our predicted NMRs and U5MRs were in line with those estimated by the UN IGME by 2030 and 2050 (they overlapped in 27/31 countries for NMRs and 22 for U5MRs) and by the Institute for Health Metrics and Evaluation (IHME) by 2030 (26/31 and 23/31, respectively). This study has a number of limitations, including poor data quality issues that reflected bias in the report of births and deaths, preventing reliable estimates and predictions from a few countries.


Conclusions 
To our knowledge, this study is the first to combine full birth histories and mortality estimates from external reliable sources to model age patterns of under-5 mortality across time in SSA. We demonstrate that countries with a rapid pace of mortality reduction (ARR ≥ 3.2%) across ages would be more likely to achieve the SDG mortality targets. However, the lower pace of neonatal mortality reduction would prevent most countries from achieving those targets: 2 countries would reach them by 2030, 13 between 2030 and 2050, and 13 after 2050.

/pdf/age-distribution-trends-and-forecasts-of-under-5-mortality-ujy2zq4skm.pdf

Age distribution, trends, and forecasts of under-5 mortality in 31 sub-Saharan African countries: A modeling study

Objectives Maternal health care improvement and reduction of maternal and child mortality are priorities of the global health agenda. In Angola, maternal mortality remains high and the risk of pregnancy-related death was 1 in 32 during 2015. This study aims to identify demographic and social factors influencing antenatal care and health facility delivery among women in Dande and to understand their impact on birth outcomes. Methods This study is based on community-based longitudinal data collected by the Dande Health and Demographic Surveillance System between 2009 and 2015. Data on pregnancy outcomes (10,289 outcomes of 8,066 women) were collected for all reported pregnancies, including sociodemographic information, health services utilisation and women's reproductive history. Logistic regression was used to investigate the determinants of birth outcomes, antenatal care attendance and institutionalised delivery. Findings Of the 10,289 pregnancy outcomes, 98.5% resulted in live births, 96.8% attended antenatal care, and 82.5% had four or more visits. Yet, 50.7% of the women delivered outside a health facility. Antenatal care attendance was a determinant of birth outcomes (stillbirth: unadjusted OR = 0.34 95% CI = 0.16-0.70; abortion: OR = 0.07 95% CI = 0.04-0.12). Older women, with lower education, living at a greater distance of a health facility and in rural areas, were less likely to use maternal health care. Having had previous pregnancies, namely resulting in live births, also decreased the likelihood of health care utilization by pregnant women. Conclusions The study identifies relevant social determinants for the utilisation of antenatal care, place of delivery and their impact on birth outcome, thereby providing insight on how best to address inequities in health care utilization.

/pdf/determinants-of-maternal-health-care-and-birth-outcome-in-284u3wqwzr.pdf

Determinants of maternal health care and birth outcome in the Dande Health and Demographic Surveillance System area, Angola.

In Eritrea, facility delivery rates show great discrepancy within urban centers. This study was conducted in Akordet, a multi-cultural lowland town of Gash-Barka Region, aiming at assessing the factors influencing facility delivery. A community based analytical cross-sectional study was conducted among a total of 282 mothers who gave birth within the 2 years preceding the data collection time. Data collection was carried out by going house-to-house and interviewing the mothers using a structured closed ended questionnaire. Bivariate and multivariate logistic regressions were used to determine the magnitude of the relationship between place of delivery and the explanatory variables (Religion, Ethnicity, Mother’s educational level, Husband’s Educational level, Place of delivery preceding last pregnancy, Birth order of last child, Any complications during previous delivery, First ANC Visit during last pregnancy, Number of ANC visits during last pregnancy and Any complication during last pregnancy.). For this study, p-value ≤0.05 was considered as statistically significant. The rate of facility delivery in this setting was found to be 82.3%. Almost all (96.1%) the mothers had at least one ANC visit during their last pregnancy, with the majority (59.7%) visiting ANC clinics during second trimester for the first time. Mothers whose educational level is junior and above (AOR 8.8, CI: 1.18–65.64), whose husband’s educational level is junior and above (AOR 3.92, CI: 1.03–14.54), who gave birth in health facility before the last pregnancy (AOR 8.16, CI: 3.41–19.48), and those who had complications during last pregnancy (AOR 2.24, CI: 1.04–4.82) were more likely to deliver in a health facility. Mothers whose last child’s birth order was 4th -6th were less likely (AOR 0.24, CI: 0.090.62) to deliver at health facility. Early initiation of ANC and regularity in attendance should be emphasized. Health educations given to pregnant mothers should try to persuade the mothers that each pregnancy and ensuing delivery is unique. Empowering the community in general and women in particular by increasing the level of participation in education might payoff in high level of facility delivery.

/pdf/factors-determining-choice-of-place-of-delivery-analytical-1kqnr6yr4p.pdf

Factors determining choice of place of delivery: analytical cross-sectional study of mothers in Akordet town, Eritrea.

Background Africa, and sub-Saharan Africa in particular, remains one of the regions with modest improvements to maternal and newborn survival and morbidity. Good quality intrapartum and early postpartum care in a health facility as well as delivery under the supervision of trained personnel is associated with improved maternal and newborn health outcomes and decreased mortality. We describe and contrast recent time trends in the scale and socio-economic inequalities in facility-based and private facility-based childbirth in sub-Saharan Africa. Methods We used Demographic and Health Surveys in two time periods (2000-2007 and 2008-2016) to analyse levels and time trends in facility-based and private facility-based deliveries for all live births in the five-year survey recall period to women aged 15-49. Household wealth quintiles were used for equity analysis. Absolute numbers of births by facility sector were calculated applying UN Population Division crude birth rates to the total country population. Results The percentage of all live births occurring in health facilities varied across countries (5%-85%) in 2000-2007. In 2008-2016, this ranged from 22% to 92%. The lowest percentage of all births occurring in private facilities in 2000-2007 period was in Ethiopia (0.3%) and the highest in the Democratic Republic of Congo at 20.5%. By 2008-2016, this ranged from 0.6% in Niger to 22.3% in Gabon. Overall, the growth in the absolute numbers of births in facilities outpaced the growth in the percentage of births in facilities. The largest increases in absolute numbers of births occurred in public sector facilities in all countries. Overall, the percentage of births occurring in facilities was significantly lower for poorest compared to wealthiest women. As the percentage of facility births increased in all countries over time, the extent of wealth-based differences had reduced between the two time periods in most countries (median risk ratio in 2008-2016 was 2.02). The majority of countries saw a narrowing in both the absolute and relative difference in facility-based deliveries between poorest and wealthiest. Conclusions The growth in facility-based deliveries, which was largely driven by the public sector, calls for increased investments in effective interventions to improve service delivery and quality of life for the mother and newborn. The goal of universal health coverage to provide better quality services can be achieved by deploying interventions that are holistic in managing and regulating the private sector to enhance performance of the health care system in its entirety rather than interventions that only target service delivery in one sector.

/pdf/time-trends-in-facility-based-and-private-sector-childbirth-2qljapqgb6.pdf

Time trends in facility-based and private-sector childbirth care: analysis of Demographic and Health Surveys from 25 sub-Saharan African countries from 2000 to 2016

Sub-Saharan Africa remains one of the regions with modest health outcomes; and evidenced by high maternal mortality ratios and under-5 mortality rates. There are complications that occur during and following pregnancy and childbirth that can contribute to maternal deaths; most of which are preventable or treatable. Evidence shows that early and regular attendance of antenatal care and delivery in a health facility under the supervision of trained personnel is associated with improved maternal health outcomes. The aim of this study is to assess changes in and determinants of health facility delivery using nationally representative surveys in sub-Saharan Africa. This study also seeks to present renewed evidence on the determinants of health facility delivery within the context of the Agenda for Sustainable Development to generate evidence-based decision making and enable deployment of targeted interventions to improve health facility delivery and maternal and child health outcomes. We used pooled data from 58 Demographic and Health Surveys (DHS) conducted between 1990 and 2015 in 29 sub-Saharan African countries. This yielded a total of 1.1 million births occurring in the 5 years preceding the surveys. Descriptive statistics were used to describe the counts and proportions of women who delivered by place of delivery and their background characteristics at the time of delivery. We used multilevel logistic regression model to estimate the magnitude of association in the form of odds ratios between place of delivery and the predictors. Results show that births among women in the richest wealth quintile were 68% more likely to occur in health facilities than births among women in the lowest wealth quintile. Women with at least primary education were twice more likely to give birth in facilities than women with no formal education. Births from more recent surveys conducted since 2010 were 85% more likely to occur in facilities than births reported in earliest (1990s) surveys. Overall, the proportion of births occurring in facilities was 2% higher than would be expected; and varies by country and sub-Saharan African region. Proven interventions to increase health facility delivery should focus on addressing inequities associated with maternal education, women empowerment, increased access to health facilities as well as narrowing the gap between the rural and the urban areas. We further discuss these results within the agenda of leaving no one behind by 2030.

/pdf/health-facility-delivery-in-sub-saharan-africa-successes-4j1gjsx35a.pdf

Sangwani Nkhana-Salimu

Papers

Health facility delivery in sub-Saharan Africa: successes, challenges, and implications for the 2030 development agenda