Accounting for model uncertainty in estimating global burden of disease

Vock, David M.; Atchison, Elizabeth A.; Legler, Julie M.; McClure, David R. J.; Carlyle, Jamie C.; Jeavons, Elysia N.; Burton, Anthony

doi:10.2471/blt.09.073577

Cited by 9 publications

(6 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…And second, MM effect sizes and predictions across all fitted models in the model space are averaged through a weighted averaging scheme. Such model averaging methods have been shown to be both theoretically [58,226,227] and empirically [228][229][230][231][232][233][234][235][236][237][238][239][240][241] superior to SM inference methods applied to health-related data analyses [228-233, 235,239-241].…”

Section: Multimodel (Mm) Search and Averaging Methodsmentioning

confidence: 99%

Statistical modeling methods: challenges and strategies

Henley

Golden

Kashner

2019

Biostatistics & Epidemiology

View full text Add to dashboard Cite

Statistical modeling methods are widely used in clinical science, epidemiology, and health services research to analyze data that has been collected in clinical trials as well as observational studies of existing data sources, such as claims files and electronic health records. Diagnostic and prognostic inferences from statistical models are critical to researchers advancing science, clinical practitioners making patient care decisions, and administrators and policy makers impacting the health care system to improve quality and reduce costs. The veracity of such inferences relies not only on the quality and completeness of the collected data, but also statistical model validity. A key component of establishing model validity is determining when a model is not correctly specified and therefore incapable of adequately representing the Data Generating Process (DGP). In this article, model validity is first described and methods designed for assessing model fit, specification, and selection are reviewed. Second, data transformations that improve the model's ability to represent the DGP are addressed. Third, model search and validation methods are discussed. Finally, methods for evaluating predictive and classification performance are presented. Together, these methods provide a practical framework with recommendations to guide the development and evaluation of statistical models that provide valid statistical inferences.

show abstract

Section: Multimodel (Mm) Search and Averaging Methodsmentioning

confidence: 99%

Statistical modeling methods: challenges and strategies

Henley

Golden

Kashner

2019

Biostatistics & Epidemiology

View full text Add to dashboard Cite

show abstract

“…Incomplete knowledge of system dynamics and mediating mechanisms may often lead to exclusion of important variables, interactions and improper choices during model design. The uncertainty derived from the multiple choices involved in data processing and model construction (e.g., type of model, dynamic aspects, spatio-temporal scale, discrete or continuous nature of variables, stochasticity, level of analysis and the various statistical choices) is also often unaccounted, despite their direct effects on model estimates [ 20 ]. Additional sources of uncertainty come into play when model outcomes are extrapolated into the future, including poor specification of hypothetical future scenarios and the uncertainty or suitability of assuming that the past can mirror the future reliably.…”

Section: Discussionmentioning

confidence: 99%

Beyond crystal balls: crosscutting solutions in global health to prepare for an unpredictable future

et al. 2015

View full text Add to dashboard Cite

BackgroundEfforts in global heath need to deal not only with current challenges, but also to anticipate new scenarios, which sometimes unfold at lightning speed. Predictive modeling is frequently used to assist planning, but outcomes depend heavily on a subset of critical assumptions, which are mostly hampered by our limited knowledge about the many factors, mechanisms and relationships that determine the dynamics of disease systems, by a lack of data to parameterize and validate models, and by uncertainties about future scenarios.DiscussionWe propose a shift from a focus on the prediction of individual disease patterns to the identification and mitigation of broader fragilities in public health systems. Modeling capabilities should be used to perform “stress tests” on how interrelated fragilities respond when faced with a range of possible or plausible threats of different nature and intensity. This system should be able to reveal crosscutting solutions with the potential to address not only one threat, but multiple areas of vulnerability to future health risks.SummaryActionable knowledge not based on a narrow subset of threats and conditions can better guide policy, build societal resilience and ensure effective prevention in an uncertain world.

show abstract

“…In the GBD 2019 study, 95% uncertainty intervals (95% UI) 16 are calculated for all the results that provide information about the variability of estimates resulting from errors due to the sampling process and non-sampling errors caused by adjustments in data sources and modelling. The intervals correspond to 2.5 and 97.5 percent of the values obtained after the repetition of 100 draws for each estimate, after each step of the estimation process, using the Monte Carlo simulation method 17 .…”

Section: Methodsmentioning

confidence: 99%

Health impacts caused by excessive sodium consumption in Brazil: results of the GBD 2019 study

Guedes

Felisbino-Mendes

Vegi

et al. 2022

Rev. Soc. Bras. Med. Trop.

View full text Add to dashboard Cite

INTRODUCTION: Excessive sodium consumption is associated with increased blood pressure, which is an important risk factor for non-communicable diseases (NCDs). This study therefore aimed to describe the burden of NCDs attributable to excessive sodium consumption among Brazilians. METHODS: This observational study used mortality and Disability Adjusted Life Years (DALY) rates , and their respective uncertainty intervals (UI), from the Global Burden of Disease Study 2019 (GBD 2019) . The burden was obtained by the population attributable fraction of each NCD, considering the minimum theoretical value of risk (intake of 0-3g of sodium/day); the excessive consumption proportion in the population, obtained through population inquiries; and the relative risks obtained through meta-analyses. RESULTS: Excessive sodium consumption was the third highest dietary risk contributing to deaths (30,814; 95% UI = 2,034 - 84,130) and DALYs (699,119; 95% UI= 43,130 - 1,914,066) in 2019. States from the Northeast region had the highest age-standardized rates of deaths and DALYs, and the male population was more affected by NCDs caused by excessive sodium consumption. Cardiovascular diseases were the main contributing factors in the burden attributable to excessive sodium consumption. CONCLUSIONS: Regardless of the progress in addressing NCDs related to this risk factor, the impact remains high, especially among men and in the Northeast region. More effective measures are needed to reduce sodium in industrialized products, such as health promotion actions to combat sodium consumption, in order to prevent and control NCDs in Brazil.

show abstract

Accounting for model uncertainty in estimating global burden of disease

Cited by 9 publications

References 16 publications

Statistical modeling methods: challenges and strategies

Statistical modeling methods: challenges and strategies

Beyond crystal balls: crosscutting solutions in global health to prepare for an unpredictable future

Health impacts caused by excessive sodium consumption in Brazil: results of the GBD 2019 study

Contact Info

Product

Resources

About