METHODS: We assessed fruit and vegetable consumption using data from country-specific, validated semi-quantitative food frequency questionnaires in the Prospective Urban Rural Epidemiology (PURE) study, which enrolled participants from communities in 18 countries between Jan 1, 2003, and Dec 31, 2013. We documented household income data from participants in these communities; we also recorded the diversity and non-sale prices of fruits and vegetables from grocery stores and market places between Jan 1, 2009, and Dec 31, 2013. We determined the cost of fruits and vegetables relative to income per household member. Linear random effects models, adjusting for the clustering of households within communities, were used to assess mean fruit and vegetable intake by their relative cost.

FINDINGS: Of 143 305 participants who reported plausible energy intake in the food frequency questionnaire, mean fruit and vegetable intake was 3·76 servings (95% CI 3·66-3·86) per day. Mean daily consumption was 2·14 servings (1·93-2·36) in low-income countries (LICs), 3·17 servings (2·99-3·35) in lower-middle-income countries (LMICs), 4·31 servings (4·09-4·53) in upper-middle-income countries (UMICs), and 5·42 servings (5·13-5·71) in high-income countries (HICs). In 130 402 participants who had household income data available, the cost of two servings of fruits and three servings of vegetables per day per individual accounted for 51·97% (95% CI 46·06-57·88) of household income in LICs, 18·10% (14·53-21·68) in LMICs, 15·87% (11·51-20·23) in UMICs, and 1·85% (-3·90 to 7·59) in HICs (ptrend=0·0001). In all regions, a higher percentage of income to meet the guidelines was required in rural areas than in urban areas (p<0·0001 for each pairwise comparison). Fruit and vegetable consumption among individuals decreased as the relative cost increased (ptrend=0·00040).

INTERPRETATION: The consumption of fruit and vegetables is low worldwide, particularly in LICs, and this is associated with low affordability. Policies worldwide should enhance the availability and affordability of fruits and vegetables.

METHODS: We did a parallel, two-arm, prospective observational study of opioid-dependent individuals aged 18 years and older who were treated in Malaysia in the Klang Valley in two settings: CDDCs and VTCs. We used sequential sampling to recruit individuals. Assessed individuals in CDDCs were required to participate in services such as counselling sessions and manual labour. Assessed individuals in VTCs could voluntarily access many of the components available in CDDCs, in addition to methadone therapy. We undertook urinary drug tests and behavioural interviews to assess individuals at baseline and at 1, 3, 6, 9, and 12 months post-release. The primary outcome was time to opioid relapse post-release in the community confirmed by urinary drug testing in individuals who had undergone baseline interviewing and at least one urine drug test (our analytic sample). Relapse rates between the groups were compared using time-to-event methods. This study is registered at ClinicalTrials.gov (NCT02698098).

FINDINGS: Between July 17, 2012, and August 21, 2014, we screened 168 CDDC attendees and 113 VTC inpatients; of these, 89 from CDDCs and 95 from VTCs were included in our analytic sample. The baseline characteristics of the two groups were similar. In unadjusted analyses, CDDC participants had significantly more rapid relapse to opioid use post-release compared with VTC participants (median time to relapse 31 days [IQR 26-32] vs 352 days [256-unestimable], log rank test, p<0·0001). VTC participants had an 84% (95% CI 75-90) decreased risk of opioid relapse after adjustment for control variables and inverse propensity of treatment weights. Time-varying effect modelling revealed the largest hazard ratio reduction, at 91% (95% CI 83-96), occurs during the first 50 days in the community.

INTERPRETATION: Opioid-dependent individuals in CDDCs are significantly more likely to relapse to opioid use after release, and sooner, than those treated with evidence-based treatments such as methadone, suggesting that CDDCs have no role in the treatment of opioid-use disorders.

METHODS: We assembled 1155 geographical records of yellow fever virus infection in people from 1970 to 2016. We used a Poisson point process boosted regression tree model that explicitly incorporated environmental and biological explanatory covariates, vaccination coverage, and spatial variability in disease reporting rates to predict the relative risk of apparent yellow fever virus infection at a 5 × 5 km resolution across all risk zones (47 countries across the Americas and Africa). We also used the fitted model to predict the receptivity of areas outside at-risk zones to the introduction or reintroduction of yellow fever transmission. By use of previously published estimates of annual national case numbers, we used the model to map subnational variation in incidence of yellow fever across at-risk countries and to estimate the number of cases averted by vaccination worldwide.

FINDINGS: Substantial international and subnational spatial variation exists in relative risk and incidence of yellow fever as well as varied success of vaccination in reducing incidence in several high-risk regions, including Brazil, Cameroon, and Togo. Areas with the highest predicted average annual case numbers include large parts of Nigeria, the Democratic Republic of the Congo, and South Sudan, where vaccination coverage in 2016 was estimated to be substantially less than the recommended threshold to prevent outbreaks. Overall, we estimated that vaccination coverage levels achieved by 2016 avert between 94 336 and 118 500 cases of yellow fever annually within risk zones, on the basis of conservative and optimistic vaccination scenarios. The areas outside at-risk regions with predicted high receptivity to yellow fever transmission (eg, parts of Malaysia, Indonesia, and Thailand) were less extensive than the distribution of the main urban vector, A aegypti, with low receptivity to yellow fever transmission in southern China, where A aegypti is known to occur.

INTERPRETATION: Our results provide the evidence base for targeting vaccination campaigns within risk zones, as well as emphasising their high effectiveness. Our study highlights areas where public health authorities should be most vigilant for potential spread or importation events.

METHODS: We assessed use of antiplatelet, cholesterol, and blood-pressure-lowering drugs in 8492 individuals with self-reported cardiovascular disease from 21 countries enrolled in the Prospective Urban Rural Epidemiology (PURE) study. Defining one or more drugs as a minimal level of secondary prevention, wealth-related inequality was measured using the Wagstaff concentration index, scaled from -1 (pro-poor) to 1 (pro-rich), standardised by age and sex. Correlations between inequalities and national health-related indicators were estimated.

FINDINGS: The proportion of patients with cardiovascular disease on three medications ranged from 0% in South Africa (95% CI 0-1·7), Tanzania (0-3·6), and Zimbabwe (0-5·1), to 49·3% in Canada (44·4-54·3). Proportions receiving at least one drug varied from 2·0% (95% CI 0·5-6·9) in Tanzania to 91·4% (86·6-94·6) in Sweden. There was significant (p<0·05) pro-rich inequality in Saudi Arabia, China, Colombia, India, Pakistan, and Zimbabwe. Pro-poor distributions were observed in Sweden, Brazil, Chile, Poland, and the occupied Palestinian territory. The strongest predictors of inequality were public expenditure on health and overall use of secondary prevention medicines.

INTERPRETATION: Use of medication for secondary prevention of cardiovascular disease is alarmingly low. In many countries with the lowest use, pro-rich inequality is greatest. Policies associated with an equal or pro-poor distribution include free medications and community health programmes to support adherence to medications.

METHODS: In this large-scale prospective cohort study, we recruited adults aged between 35 years and 70 years from 367 urban and 302 rural communities in 20 countries. We collected data on families and households in two questionnaires, and data on cardiovascular risk factors in a third questionnaire, which was supplemented with physical examination. We assessed socioeconomic status using education and a household wealth index. Education was categorised as no or primary school education only, secondary school education, or higher education, defined as completion of trade school, college, or university. Household wealth, calculated at the household level and with household data, was defined by an index on the basis of ownership of assets and housing characteristics. Primary outcomes were major cardiovascular disease (a composite of cardiovascular deaths, strokes, myocardial infarction, and heart failure), cardiovascular mortality, and all-cause mortality. Information on specific events was obtained from participants or their family.

FINDINGS: Recruitment to the study began on Jan 12, 2001, with most participants enrolled between Jan 6, 2005, and Dec 4, 2014. 160 299 (87·9%) of 182 375 participants with baseline data had available follow-up event data and were eligible for inclusion. After exclusion of 6130 (3·8%) participants without complete baseline or follow-up data, 154 169 individuals remained for analysis, from five low-income, 11 middle-income, and four high-income countries. Participants were followed-up for a mean of 7·5 years. Major cardiovascular events were more common among those with low levels of education in all types of country studied, but much more so in low-income countries. After adjustment for wealth and other factors, the HR (low level of education vs high level of education) was 1·23 (95% CI 0·96-1·58) for high-income countries, 1·59 (1·42-1·78) in middle-income countries, and 2·23 (1·79-2·77) in low-income countries (pinteraction<0·0001). We observed similar results for all-cause mortality, with HRs of 1·50 (1·14-1·98) for high-income countries, 1·80 (1·58-2·06) in middle-income countries, and 2·76 (2·29-3·31) in low-income countries (pinteraction<0·0001). By contrast, we found no or weak associations between wealth and these two outcomes. Differences in outcomes between educational groups were not explained by differences in risk factors, which decreased as the level of education increased in high-income countries, but increased as the level of education increased in low-income countries (pinteraction<0·0001). Medical care (eg, management of hypertension, diabetes, and secondary prevention) seemed to play an important part in adverse cardiovascular disease outcomes because such care is likely to be poorer in people with the lowest levels of education compared to those with higher levels of education in low-income countries; however, we observed less marked differences in care based on level of education in middle-income countries and no or minor differences in high-income countries.

INTERPRETATION: Although people with a lower level of education in low-income and middle-income countries have higher incidence of and mortality from cardiovascular disease, they have better overall risk factor profiles. However, these individuals have markedly poorer health care. Policies to reduce health inequities globally must include strategies to overcome barriers to care, especially for those with lower levels of education.

METHODS: Data were from the 10/66 Study. Individuals aged 65 years or older and without dementia at baseline were selected from China, Cuba, the Dominican Republic, Mexico, Peru, Puerto Rico, and Venezuela. Dementia incidence was assessed over 3-5 years, with diagnosis according to the 10/66 Study diagnostic algorithm. Discrimination and calibration were tested for five models: the Cardiovascular Risk Factors, Aging and Dementia risk score (CAIDE); the Study on Aging, Cognition and Dementia (AgeCoDe) model; the Australian National University Alzheimer's Disease Risk Index (ANU-ADRI); the Brief Dementia Screening Indicator (BDSI); and the Rotterdam Study Basic Dementia Risk Model (BDRM). Models were tested with use of Cox regression. The discriminative accuracy of each model was assessed using Harrell's concordance (c)-statistic, with a value of 0·70 or higher considered to indicate acceptable discriminative ability. Calibration (model fit) was assessed statistically using the Grønnesby and Borgan test.

FINDINGS: 11 143 individuals without baseline dementia and with available follow-up data were included in the analysis. During follow-up (mean 3·8 years [SD 1·3]), 1069 people progressed to dementia across all sites (incidence rate 24·9 cases per 1000 person-years). Performance of the models varied. Across countries, the discriminative ability of the CAIDE (0·52≤c≤0·63) and AgeCoDe (0·57≤c≤0·74) models was poor. By contrast, the ANU-ADRI (0·66≤c≤0·78), BDSI (0·62≤c≤0·78), and BDRM (0·66≤c≤0·78) models showed similar levels of discriminative ability to those of the development cohorts. All models showed good calibration, especially at low and intermediate levels of predicted risk. The models validated best in Peru and poorest in the Dominican Republic and China.

INTERPRETATION: Not all dementia prediction models developed in HICs can be simply extrapolated to LMICs. Further work defining what number and which combination of risk variables works best for predicting risk of dementia in LMICs is needed. However, models that transport well could be used immediately for dementia prevention research and targeted risk reduction in LMICs.