Adoption Studies PDF 2024

Quantitative Genetics 3: Adoption Studies Dr Tom McAdams MSc DevPP 7PADDTMF: Nature Nurture 1 26/11/2024 1 Learning objectives: understand the logic Why use adoption studies – strengths and limitations Key approaches Adoptive parents and children Bio...

Quantitative Genetics 3: Adoption Studies Dr Tom McAdams MSc DevPP 7PADDTMF: Nature Nurture 1 26/11/2024 1 Learning objectives: understand the logic Why use adoption studies – strengths and limitations Key approaches Adoptive parents and children Biological parents and children Children conceived using in-vitro-fertilisation (IVF) Biological siblings reared apart Unrelated siblings reared together What each approach can or can’t tell you about the influence of genetics and/or environments on child development 2 What influences child development? When do parents influence children? When do genetics influence children? When do wider environments influence children? …When are children like their parents? …When are siblings similar? …When are siblings different? (Nurture) (Nature) 3 The basis of adoption designs. 4 What influences child development? In adoption studies you can When do parents influence children? When do genetics influence children? separate the influence of shared When do wider environments influence children? genes and environments, based …When are children like their parents? on known information about the …When are siblings similar? family. …When are siblings different? 5  Parents who adopted their child at birth: What does this control* for? * A control variable is an element held constant throughout an experiment or analysis, to assess the (potentially causal) relationship between other variables. A control variable is a variable or an element which is held constant throughout an experiment or study to assess the relationship between other variables. Since it remains constant, i.e., in an unchanging state, it enables researchers to test and better understand the causal relationship between other variables (i.e., helps you control for confounding). 6 Parents who used an egg and sperm donor for IVF: What does this control for? 7 Parents who’s biological child is born via surrogacy: What does this control for? 8 Parents who gave their child for adoption at birth: What does this control for? 9  Identical twins raised separately from birth: What does this control for? Don’t need the parents in an adoption design – if you have sibling pairs 10  Biological siblings raised separately from birth: What does this control for? Don’t need the parents in an adoption design – if you have sibling pairs 11  Unrelated siblings raised together from birth: What does this control for? Don’t need the parents in an adoption design – if you have sibling pairs 12  Summary Genetic relatives Genetic + environmental Adoption relatives Environmental relatives Page 75: Adoption is a natural experiment that creates “genetic” relatives (biological parents and their adopted-away offspring; siblings adopted apart) and “environmental” relatives (adoptive parents and adopted children; genetically unrelated siblings adopted into the same family). Resemblance for these “genetic” and “environmental” relatives can be used to test the extent to which resemblance between the usual “genetic + environmental” relatives is due to either nature or nurture. Families of parents and biological children: all family members are genetic and environmental relatives. Genetic relatives share genes but not family environment. For example, siblings adopted apart, biological parents and adopted away child. Environmental relatives share family environment but aren’t genetically related. For example, adoptee and adoptive parents, adoptee and their adoptive siblings. Studying genetic relatives is useful because trait resemblance implies genetic influences. Trait resemblance between environmental relatives implies influence of shared environment. 13 The origins and history of adoption designs. 14  Early adoption studies Study of IQ (Burks 1928) 214 foster children and 105 control children, with their parents Aged 5 – 14 years, in USA cities Measure intelligence in both generations “Geniuses and imbeciles, regardless of what environment they are brought up in, are the products of heredity, for environment simply cannot stretch or shrink the mind of a normal individual to the point of genius or imbecility. Heredity, on the other hand, is an influence so powerful…” Conclusions later critiqued (e.g., Rajan 1983) https://www.jstor.org/stable/pdf/44111476.pdf https://psycnet.apa.org/record/2006-20848-051 https://www.gwern.net/docs/genetics/heritable/1949-burks.pdf 15  Early adoption studies Study of IQ (Skodak and Skeels, 1949) First full adoption study including birth and adoptive parents. The same group of adopted children administered intelligence tests on four occasions. Highlight issues with “selective placement” – particularly in relation to socioeconomic status. Children adopted away had better IQ than their birth parents – suggesting benefits of the improved home environment. Selective placement: Children placed in adoptive homes similar to those of their birth parents (especially in relation to socio-economic status) – which violates assumptions of the adoption design. In adoption designs we assume NO selective placement. 16  Early adoption studies Study of IQ (Newman et al., 1937) Public appeal for information about any cases of identical twins reared apart – 19 pairs found and received national press interest. “offer came from a charitable organization, which was willing to give us a pair of identical-twin orphans a few weeks old, whom we could then separate and bring up in environments as different as we might wish. This seemed to us a fine opportunity, but the thought of having to wait twelve years for them to reach an age suitable for testing, and in the meantime acting as scientific parent to a pair of twins…” … but the thought of having to wait twelve years for them to reach an age suitable for testing, and in the meantime acting as scientific parent to a pair of twins, would have been enough to dissuade us, even without the refusal of our better half to countenance any such wild scheme. 17  Early adoption studies Dr Peter Neubauer’s secret study Infants twins and triplets were separated and studied in New York City in the 1960s and 70s. Adoptive parents did not know they were raising a child who had been separated from a twin. The twins' records remain sealed at Yale University until 2065. “The study was ethically defensible by the standards of its time” http://scienceandfilm.org/articles/3141/twins-reared-apart-from-birth-beyond-the- secret-study 18  Early adoption studies Minnesota Study of Twins Reared Apart (MISTRA) Launched 1979 by Bouchard et al., >100 sets of reared apart twins or triplets in USA, UK, Canada, China, Germany “…found that an identical twin reared away from his or her co-twin seems to have about an equal chance of being similar to the co- twin in terms of personality, interests, and attitudes as one who has been reared with his or her co-twin” “We can conclude that the environment, rather than making twins alike, makes them different.” Example of the ‘Jim twins’ separated by adoption at 4 weeks: Both named James, both married to a Betty, both divorced from women called Linda, both named their pet dog Toy… https://embryo.asu.edu/pages/sources-human-psychological-differences-minnesota- study-twins-reared-apart-1990-thomas-j 19  Early adoption studies Segal et al. 2018 Twins accidently switched / misplaced at birth, n=7 The Fullerton Study of Chinese Twins Reared Apart, n=23 Singleton twins adopted from China (related to 1979 one-child policy) Professor Nancy Segal https://www.cambridge.org/core/journals/twin-research-and-human- genetics/article/twins-reared-apart-and-twins-in-families-the-findings-behind-the- fascination/81CDA1FD091681E282C4EBDFC5B60397 20  Bigger, longitudinal, non-twin adoption designs. * General consensus: Unethical to separate twins! Not done these days. 21 Adoption cohort data Cohort Adoption cohort Cohort studies are a type of Following families who have adopted longitudinal study—an approach a child from birth. Typically collecting that follows research data on the adopted child, their participants over a period of adoptive parents, and their birth time (often many years). parents over a period of time (often Specifically, cohort studies many years). recruit and follow participants who share a common characteristic, such as a Adoption cohort data particular occupation or The data collected (e.g., demographic similarity. questionnaire or observation data) from an adoption cohort study. 22  USA: children adopted at birth Colorado Adoption Project Both report minimal Early Growth and (CAP) selective placement Development Study 40+ years of data and minimal (EGDS) 242 adoptive mothers, 237 adoptive fathers contact between 15+ years of data 243 control mothers, 244 control fathers adoptive and birth 360 adoptive parents 286 birth mothers, 60 birth fathers families. These are also 359 birth mothers, 114 birth fathers + adopted and non-adopted siblings controlled for in analyses + adopted and non-adopted siblings Mean age of placement 29 days as measured covariates. Mean age of placement 3 days https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3817005/ https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5575997/ ncbi.nlm.nih.gov/pmc/articles/PMC7056588/ 23  Early Growth and Development Study (EGDS) 6 years 7 years 8 years Adoptive Adoptive Adoptive father anxiety father anxiety father anxiety Adoptive Adoptive Adoptive mother anxiety mother anxiety mother anxiety Child Child Child anxiety anxiety anxiety * A Cross-Lagged Panel Model Birth parent internalising This is a “structural equation model” This type is called a “cross-lagged panel model” Explained in more detail later in the module by Dr Moritz Here. 24  Early Growth and Development Study (EGDS) 6 years 7 years 8 years Adoptive Adoptive Adoptive father anxiety father anxiety father anxiety.10**(.03,.18) Adoptive Adoptive Adoptive mother anxiety mother anxiety mother anxiety.16**(.08,.24) Child Child Child anxiety anxiety anxiety.01 (-.09,.11).06 (-.03,.16) -.04 (-.12,.04) * Controlled for measured covariates: - Obstetric complications - Adoption openness Birth parent - Child sex internalising https://acamh.onlinelibrary.wiley.com/doi/full/10.1111/jcpp.13068 Bold arrow shows result that was consistent across sensitivity analyses. Results shows child to parent mother anxiety prediction, free from confounding by genetic relatedness. No evidence for associations between birth parents’ anxiety/depression and child anxiety at these timepoints. 25  Early Growth and Development Study (EGDS) * Controlled for measured covariates: - Obstetric complications - Adoption openness https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4667033/ More evidence for bidirectional associations between parents and children, free from confounding by genetic relatedness. 26  Early Growth and Development Study (EGDS) + Early Parenting of Children (EPoCh) study 711 children from 414 households In middle childhood (around age 11): (1) Biological sibs* sharing a home “Genetically related sibs more similar in BMI (2) Biological sibs* in different homes than genetically unrelated sibs – especially (3) Non-biological sibs sharing a home when reared in the same home *Includes MZ twins, DZ twins, full-sibs, half-sibs environment” Total BMI variance attributed to: * Controlled for measured covariates: 63% heritable factors in sib pairs – Sex – Race & ethnicity – Home type (adoptive/birth) 31% common environment in sib pairs – At least one biological parent (yes/no) – & BMI scores already age-adjusted 6% unique environment and error https://pubmed.ncbi.nlm.nih.gov/32687510/ 27  UK: Cardiff IVF Study A 'prenatal cross-fostering' design where pregnant mothers are related or unrelated to their child as a result of in vitro fertilization (IVF) Conception type N Homologous IVF (parent’s own gametes) 386 (50%) Sperm donation 182 (24%) Embryo donation 153 (20%) Egg donation 27 (3.5%) Gestational Surrogacy 21 (2.7%) https://bmcmedresmethodol.biomedcentral.com/articles/10.1186/1471-2288-7-25 https://www.sciencedirect.com/science/article/pii/S0890856711000621 28  UK: Cardiff IVF Study Child’s birth Prenatal Child’s weight stress ADHD Prenatal Child’s b. Association only present in related mother-offspring stress gestational age pairs, consistent with genetic links. Child’s antisocial behaviour ***Importance of the measured Postnatal a. Associations equivalent in related and unrelated mother- stress covariate!*** child pairs, consistent with environmental links. * Controlled for measured covariates: – Child age – Child gender Prenatal Child’s – Family social class – Maternal age at child birth stress anxiety – Maternal height – Antenatal health – Postnatal anxiety/depression c. Association explained by exposure to postnatal stress. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2830085/pdf/S0033291709005911a.pdf 29 Using pre-existing datasets for adoption designs. 30 Registry data A registry Administrative data A place where registers or records are kept in an Information routinely gathered when registering organized system to store uniform data. E.g. people, carrying out transactions, or for record birth records are stored in a birth registry. keeping – usually when delivering a service. A national registry Useful in public services, e.g., schools, courts, benefits, or healthcare systems. A place where a nation (country) stores records (usually administrative) about all of the people All interactions stored for operational in the country. purposes (to carry out their day-to-day work, to monitor and improve their performance, and E.g. a national health registry (or national patient registry) would include health records keep providing services in an effective way). from everyone accessing healthcare in the country. A national birth registry would include An important data source for research… information on everyone born in the country. 31 Registry data Scandinavian countries hold the best national Can use these data to create population registry data cohorts, following groups of interest across time through their administrative records. Everyone has a unique personal identification number, links their administrative data across …Can find all kinds of adoptive families! By national public registries. data on who gave birth to who, and who lives with who, who went to family courts for Gives individual-level data on most facets of people’s lives, including health, education, adoption processes. And link to their health, housing, childbirth, etc. education, crime, etc data. Following slides mostly about Scandinavian studies, but the UK infrastructure for national registry data is catching up. Administrative Data Research UK is giving researchers access the UK’s public sector data. 32  Swedish registry data Siblings pairs matched for high genetic Census data + national criminal, medical and pharmacy risk for drug abuse were exposed to registers used to identify biological families where: different rearing environments. Parent(s) have records of Siblings reared in adoptive homes had drug abuse, a substantially reduced risk for drug alcohol abuse, abuse compared to their non-adopted or criminal behaviour. sibling. Full-sibling pairs where one Half-sibling pairs where one “High-quality rearing environments child adopted before age 5 child adopted before age 5 can substantively reduce risk for drug (n=1161) (n= 3085) abuse in those at high genetic risk” * Controlled for measured covariates: – Parent birth age – Parent gender – Drug abuse risk of non-shared parent https://pubmed.ncbi.nlm.nih.gov/26753502/ 33  Swedish registry data n = 19,715 n = 82,698 Genetic + 1. Adoptive parent: Environmental 2. Lived-with parent: environmental link divorce link divorce Adopted child: Biological child: divorce divorce Birth parent: Lived-apart parent: Genetic Genetic divorce divorce link link n = 8,523 – 53,097 * Controlled for measured covariates: 3. Genetic + Environmental environmental link – Parent education – Parent Externalising (alcohol use disorder, drug link abuse, criminal behaviour) – Offspring year of birth Biological siblings: Adoptive siblings: divorce divorce https://journals.sagepub.com/doi/full/10.1177/0956797617734864 34  Swedish registry data n = 19,715 n = 82,698 1. “Adoptees resembled their biological 2. “Offspring resembled their not-lived-with parents, rather than their adoptive parents, in fathers and their lived-with mothers, with their history of divorce” stronger resemblance to lived-with mothers” “0.13 heritability estimate for divorce across generations” n = 8,523 – 53,097 * Controlled for measured covariates: 3. “Adoptees resembled their biological siblings, not adoptive siblings, in their history – Parent education of divorce” – Parent Externalising (alcohol use disorder, drug abuse, criminal behaviour) – Offspring year of birth https://journals.sagepub.com/doi/full/10.1177/0956797617734864 35  Finnish registry data + clinical observations and interviews Adoptive family