Bridging childhood experiences and adult development: Boston Early Adversity and Mortality Study

Mina Antic, BS

Massachusetts General Hospital, Boston Veteran’s Administration Healthcare System, Northwestern University
Bridging childhood experiences and adult development: Boston Early Adversity and Mortality Study

Scientific Abstract

Background: Longitudinal studies of aging often begin in midlife. Although they provide rich data on health trajectories in adulthood, the reliance on retrospective data on early conditions and a lack of “birth-to-death” data on intervening processes have limited progress in studying pathways linking early experiences to later-life health. The Boston Early Adversity and Mortality Study (BEAMS) aims to overcome these shortcomings by using administrative record linkage methods to augment three nearly complete longitudinal aging studies with prospective data on the cohorts’ early life conditions.

Methods: BEAMS comprises the all-male VA Normative Aging (n= 2280), Grant (n=456), and Glueck (n=268) Studies (74%-94% deceased). It extends coverage to siblings (including women) such that our combined sample is representative of the early 20th century Northeastern

U.S. population. BEAMS acquires prospective information on early-life socioeconomic (e.g., neighborhood poverty) and environmental (e.g., lead exposure) conditions, and later-life health through administrative record linkage. This multi-step, human review process involves connecting individual identifiers to records in databases such as the 1900-40 U.S. Census, vital statistics, military, and Medicare records.

Conclusions: Through record linkage, BEAMS will create a cradle-to-grave dataset with prospective data on early-life exposures and lifespan health data. Including siblings will enable within-family comparisons. Our approach of a coordinated analysis across studies will enhance the replicability and robustness of our results, and our methodology can be adopted by other longitudinal aging studies to advance knowledge about pathways linking early-life conditions to lifespan health outcomes.

SoundCloud Transcript

Hello, and thank you for your interest in this poster. We are Ashley Dorame, Mina Antic, Dan Horgan, and the RAs for the Boston Early Adversity and Mortality Study, or BEAMS. This poster introduces our study, as well as our administrative record linkage methodology. We are speaking on behalf of our project team, which is led by Dr. Lewina Lee from Boston University & VA Boston and Dr. Daniel Mroczek from Northwestern University.

A large body of work has linked childhood socioeconomic deprivation and psychosocial adversity (such as maltreatment by caretakers) to poor later-life health. Studies have also linked environmental hazards in childhood, such as lead exposure, to worse health in adulthood.

However, lifespan health research on early adversity historically faces many challenges

First, virtually all existing longitudinal studies of aging began in adulthood, thus measures of early adversity mostly rely on retrospective self-report.

 

Second, earlier research often focused on single categories of stressors, which ignores confounding from correlated risks.

 

And finally, rigorous testing of a lifespan model would require data on different types of exposures, later-life health outcomes, and ideally longitudinal data on mediating factors, which poses enormous logistical and financial challenges.

In designing BEAMS, our team aims to overcome these limitations by augmenting existing data from three long-running aging studies. Our two main goals are:

One, to obtain prospective data on the early-life context and later-life health of our participants and their siblings from multiple administrative databases.

And two, to evaluate lifespan processes that transmit the impact of early life adversity onto later-life health by creating a prospective, birth-to-death dataset.

 

Our sample includes participants from 3 Greater Boston based longitudinal studies: the Glueck study, the Grant Study, and the VA Normative Aging Study, or NAS. At enrollment around 1938, participants in the Glueck study were inner-city school boys, who served as non-delinquent controls to a cohort of juvenile delinquents. Participants in the Grant study were a group of Harvard College sophomores in the classes of 1939-42. NAS enrolled healthy middle-aged men in the 1960s. In BEAMS, we extend our samples to siblings, including women, which yields a combined sample that is more representative of the Northeastern U.S. population in the early 1900s.

 

Administrative linkage refers to the process of connecting individuals’ identifiers to their records in various administrative databases (such as the National Death Index or U.S. Federal Census). “Hand-linkage,” a human review process, is one of several types of linkage that we use in BEAMS to connect identifiers to potential matches in publicly available administrative databases. Although hand-linkage is very time- and labor-intensive, it results in much higher match rates when compared to algorithm-based approaches.

 

Regarding records presented in the poster: In order to protect the confidentiality of our participants, all records shown here are publicly accessible examples from Ancestry.com. They do not belong to our participants.

 

To conduct linkage, we first compile available participant identifiers (such as name and date of birth). Then, we gather any known identifiers for siblings of the participants from our longitudinal datasets. Next, we make the connection, or linkage, to records of participants and their siblings in multiple administrative databases.

 

BEAMS targets two types of databases – those that inform early-life contexts; and those that inform later-life health of our cohorts.

 

For example, the U.S. Federal Census gives us prospective data on family structure, and socioeconomic status (as it includes information on residential address, parental occupation, census-level income, etc.). Using data we collect from one administrative database (like childhood residential addresses from census), we can bridge our samples to additional data (like childhood lead exposure from residential lead piping, obtained from the Boston Water and Sewer Commission). 

 

At the end of this process, we will have a prospective, multidimensional, “birth-to-death” dataset, repeated psychosocial and biological measures throughout adulthood, and data on morbidity and mortality. We will then test several scientific hypotheses on lifespan associations linking early-life experiences to later-life health outcomes.

 

We aim to contribute to the science of early adversity through this dataset that can strengthen causal inference, expand on how different forms of early-life adversity may increase vulnerability to later-life morbidity and mortality, and inform research on interventions that may mitigate the negative health impact of early adversity.

 

Thank you so much for listening to our presentation! We greatly appreciate the time you took to learn about BEAMS. We are grateful for the NAS, Grant, and Glueck participants and staff for their long-term commitment to our studies for many decades. We would like to thank NIA for their funding and support of BEAMS and the original cohort studies. Thank you.

Live Zoom Session – April 21st

research Areas

Authors

Mina Antic*, Ashley N. Dorame*, Daniel E. Horgan*, Maria Lopes, Arnold C. Castro, Delilah S. Harounian, Joseph P. Ferrie, PhD, Robert J. Waldinger, MD, Avron Spiro III, PhD, Daniel K. Mroczek, PhD, Lewina O. Lee, PhD *Mina Antic, Ashley N. Dorame, & Daniel E. Horgan equally contributed as first authors

Principal Investigator

Daniel K. Mroczek, PhD, and Lewina O. Lee, PhD