Periodic Reporting for period 3 - ECHO (Early conditions, delayed adult effects and morbidity, disability and mortality in modern human populations)
Reporting period: 2022-03-01 to 2023-08-31
What is the explanation for this transformation of the health landscape of human populations?
ECHO is a research project designed to provide answers to this question. Over the past two decades there has been important research on the nature of delayed effects of conditions experienced in early life on the health status of adults. This field of research is known as the Developmental Origins of Adult Health and Disease (DOHaD). Increasing empirical evidence suggests that some of the mechanisms that could explain adult conditions, including obesity, T2D, CAD, kidney disease and others, are early life exposures, including those that occurred in utero and affected parents even before conception. The ECHO project will contribute to this area of study by proposing new formal demographic models of health, disability and mortality. We are formulating a microsimulation model that considers stochastic epigenetic changes triggered by early exposures and modified by subsequent individual behaviour. These models are aimed at forecasting future prevalence of obesity, T2D, CAD, disability and mortality. These will be tailored so that estimates of future health costs can be easily obtained.
Work Package No 1 includes activities to develop a formal demographic model to represent early frailty, critical ages, age patterns of chronic illness, excess morbidity, disability, and mortality. We carry out empirical testing using, among other databases, the Human Mortality Database and the Latin American Mortality Database.
The following are tasks carried out under this Work Package:
A.1. Completed initial design and implementation of a microsimulation model to forecast adult prevalence of chronic diseases and disability as a function of prevalence of adverse early conditions. This is one of two precursors of ECHOsim.
A.2. Completed analysis of historic data on cohorts born between 1900 and 1980 to identify cohort mortality patterns and retrieve estimates of excess mortality associated with exposure to adverse early conditions. Utilized state of the art Age Period Cohort (APC) models using LAMBdA and HMD. Estimates of parameters will be used in a follow up of (A.1) to enrich the microsimulation.
A.3. To complement (A.2) we analysed relations between distinct patterns of mortality decline and physical growth of birth cohorts born after 1930 in the US, Puerto Rico, and Mexico. We used Waaler surfaces to estimate excess risks of mortality in those cohorts and demonstrate that these are consistent estimates in (A.2). and those employed in the simulation model in (A.1).
A.4. As part of efforts to generate credible estimates of effects of adverse early conditions on adult health, disability and mortality to be used in ECHOsim, we completed analysis of several data sets
i. Four large data sets (MHAS, PREHCO, SABEColombia, CRELES) and estimated multistate models to assess the impact of a number of self-reported adverse early conditions on adult obesity, Type 2 Diabetes, cardiovascular disease, disability, and mortality
ii. a large data set from Argentina that enabled us to compute estimates of effects of early poverty on disability
iii. a unique US data set that facilitated estimation of place of birth on adult mortality statistics.
A.5. An important part of Work Package No 1 is research on empirical estimation of large shocks experienced by individuals. These estimates can serve as benchmarks in the ECHOsim model. To move forward in this area, we pursued two activities
i. Completed work on the effects of the 1918 flu on the nutritional status of survivors and demonstrated that survivors who experienced the flu while in utero, are deviant in physical growth markers that are set before age 10.
ii. Completed assessment of trends in healthy life expectancy among Spanish cohorts between 2012 and 2017. Identified deviant roles of birth cohorts who experienced the Spanish Civil War as infants and children.
A.7. We formulate a new model to estimate biological age (BA) from chronological age (CA) and biomarkers routinely collected in health surveys. We use structural equation models, estimate basic parameters, and then use the difference between predicted BA and CA to assess its effects on health conditions and mortality.
II. Work Package No 2: Component II: ECHOsim Multistate hazard model
Work Package No 2 includes activities to perform meta-analyses to retrieve key parameters for use in ECHOsim, estimation of multistate models for onset of chronic disease, disability and mortality, preliminary testing of ECHOsim and its precursors, multistate demographic projections, comparative analyses of multistate parameters and Approximate Bayesian Computation of stochastic forecasts
The following are tasks carried out under this Work Package:
A.1. Completed work on formulation of a model to understand excess obesity and T2D in migrant population to the US and Europe. This model is the backbone of our work on testing hypotheses from the Developmental Origins of Adult Health and Disease (DOHaD). Empirical data to test the model will provide important inputs needed for ECHOsim
A.2. Completed initial stages of an Agent Base Model (ABM), a second precursor of ECHOsim. The initial model is set to simulate trajectories of obesity and T2D using parameters for genetic and sociocultural heredity, assortative mating, and differential net fertility. As part of Work Package No 2 and No 3 we will augment this initial model by introducing new dimensions, including epigenetic changes and Gene x Environment interactions.
An extension of the preliminary ABM model described above enabled us to focus on a problem that has recently attracted a lot of attention: the impact of socioeconomic mobility in a place of residence on an individual's adult health. This area of research is important for ECHO as it belongs to consideration regarding human behaviors that have deleterious impacts on health:
A.3. Completed work analyzing data from USA counties linking county-based social mobility indicators and adult health conditions and health related behaviors of individuals who lived in those counties as children.
A.4. To introduce additional complexity in the ABM model referred to in (A.3) we are investigating estimates of Gene x Environment interactions in relation to genetic predisposition to obesity and T2D.
a. Formulated a new model to estimate biological age Work Package No 1, A.7
b. Formulated an ABM model to track the trajectory of obesity prevalence in human populations. Work Package No 2, A.1
c. Completed design of a microsimulation model of delayed effects. Work Package No 1, A.1.