ObjectiveTo determine how machine learning has been applied to prediction applications in population health contexts. Specifically, to describe which outcomes have been studied, the data sources most widely used and whether reporting of machine learning predictive models aligns with established reporting guidelines.DesignA scoping review.Data sourcesMEDLINE, EMBASE, CINAHL, ProQuest, Scopus, Web of Science, Cochrane Library, INSPEC and ACM Digital Library were searched on 18 July 2018.Eligibility criteriaWe included English articles published between 1980 and 2018 that used machine learning to predict population-health-related outcomes. We excluded studies that only used logistic regression or were restricted to a clinical context.Data extraction and synthesisWe summarised findings extracted from published reports, which included general study characteristics, aspects of model development, reporting of results and model discussion items.ResultsOf 22 618 articles found by our search, 231 were included in the review. The USA (n=71, 30.74%) and China (n=40, 17.32%) produced the most studies. Cardiovascular disease (n=22, 9.52%) was the most studied outcome. The median number of observations was 5414 (IQR=16 543.5) and the median number of features was 17 (IQR=31). Health records (n=126, 54.5%) and investigator-generated data (n=86, 37.2%) were the most common data sources. Many studies did not incorporate recommended guidelines on machine learning and predictive modelling. Predictive discrimination was commonly assessed using area under the receiver operator curve (n=98, 42.42%) and calibration was rarely assessed (n=22, 9.52%).ConclusionsMachine learning applications in population health have concentrated on regions and diseases well represented in traditional data sources, infrequently using big data. Important aspects of model development were under-reported. Greater use of big data and reporting guidelines for predictive modelling could improve machine learning applications in population health.Registration numberRegistered on the Open Science Framework on 17 July 2018 (available at https://osf.io/rnqe6/).
BackgroundThe contribution of beverages to overall diet is of increasing interest to researchers and policymakers, particularly in terms of consumption of drinks high in added sugars; however, few tools to assess beverage intake have been developed and evaluated. This study aimed to evaluate the relative validity of a new online Beverage Frequency Questionnaire (BFQ) among young adults in Canada.MethodsA cross-sectional relative validation study was conducted among young adults aged 16–30 years (n = 50). Participants completed a 17-item BFQ, a 7-day food record (7dFR), and a single-item measure of sugary drink intake. Pearson correlations and paired t-tests were used to evaluate correlation and agreement between the measures for 17 individual beverage categories, total drink consumption, total alcoholic beverage consumption, and two definitions of drinks with excess sugars. Cognitive interviews were conducted to examine participant interpretation and comprehensiveness of the BFQ.ResultsEstimates of beverage intake based on the BFQ and the 7dFR were highly correlated, specifically for the total number and volume of beverages consumed, total alcoholic beverage consumption, sugary drink intake, and each of the 17 beverage categories with 3 exceptions: coffee or tea with sugar or cream, specialty coffees, and hard alcohol with caloric mix. Paired t-tests between the BFQ and the 7dFR indicated that the average reported volume was significantly different only for sweetened fruit drinks. The single-item measure of sugary beverage intake was not significantly correlated with the 7dFR. Cognitive interviewing demonstrated high comprehension levels, and confirmed the appropriateness of the BFQ beverage categories and sizes.ConclusionsOverall, the results suggest that the BFQ performed well relative to a 7dFR and had high usability among this study population, indicating its promise for collecting population-level data on beverage intake, including sugar-sweetened beverages, which are known indicators of diet and health.Electronic supplementary materialThe online version of this article (10.1186/s12937-018-0380-8) contains supplementary material, which is available to authorized users.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.