Prediction of ADHD symptoms from prenatal data in two large population-based cohorts
Introduction The association between low birth weight and attention problems in childhood has been replicated many times (e.g. Momany, Kamradt, & Nikolas, 2018). However birth weight is unlikely the aetiological start-point of this association, as birth weight is itself the product of many pren...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Cambridge University Press
2022-06-01
|
| Series: | European Psychiatry |
| Subjects: | |
| Online Access: | https://www.cambridge.org/core/product/identifier/S0924933822003844/type/journal_article |
| _version_ | 1797616282652114944 |
|---|---|
| author | N. Dooley M. Cannon D. Cotter M. Clarke |
| author_facet | N. Dooley M. Cannon D. Cotter M. Clarke |
| author_sort | N. Dooley |
| collection | DOAJ |
| description |
Introduction
The association between low birth weight and attention problems in childhood has been replicated many times (e.g. Momany, Kamradt, & Nikolas, 2018). However birth weight is unlikely the aetiological start-point of this association, as birth weight is itself the product of many prenatal factors e.g. gestational complications, maternal toxin exposure during pregnancy and basic demographics.
Objectives
We explore (1) which prenatal factors best predict attention problems in two independant population-based cohorts of children (2) which associations, if any, are moderated by sex and (3) we report accuracy statistics of our prenatal prediction algorithm for attention problems.
Methods
Participants were children aged 9 from ABCD study from the United States (N > 9,000) and the Growing Up in Ireland (GUI) study from Ireland (N > 6,000). Selected variables included familial pscyhiatric history, maternal smoking during gestation, prescription and non-prescription drug-use during gestation and a variety of gestational complications. All interactions with sex were also included. We used 5-fold cross-validation and elastic net regression (glmnet) to identify the optimal predictors of attention problems (measured by CBCL and SDQ).
Results
Strongest predictors of attention problems in the U.S. cohort included male sex, number of drugs used during pregnancy, number of family members with a history of mental illness, and number of gestational complications. Sex interacted with several of these risks. Protective factors included being a twin/triplet, being Asian, having higher household income and higher parental education level.
Conclusions
Several risk factors for childhood attention problems were identified across both cohorts, supporting their generalizabilty. Other findings were cohort-specific.
Disclosure
No significant relationships.
|
| first_indexed | 2024-03-11T07:38:56Z |
| format | Article |
| id | doaj.art-496f6268f0da4cdbb22b632f639fcae9 |
| institution | Directory Open Access Journal |
| issn | 0924-9338 1778-3585 |
| language | English |
| last_indexed | 2024-03-11T07:38:56Z |
| publishDate | 2022-06-01 |
| publisher | Cambridge University Press |
| record_format | Article |
| series | European Psychiatry |
| spelling | doaj.art-496f6268f0da4cdbb22b632f639fcae92023-11-17T05:08:56ZengCambridge University PressEuropean Psychiatry0924-93381778-35852022-06-0165S142S14210.1192/j.eurpsy.2022.384Prediction of ADHD symptoms from prenatal data in two large population-based cohortsN. Dooley0M. Cannon1D. Cotter2M. Clarke3Royal College of Surgeons in Ireland, Psychiatry, Dublin, IrelandRoyal College of Surgeons in Ireland, Psychiatry, Dublin, IrelandRoyal College of Surgeons in Ireland, Psychiatry, Dublin, IrelandRoyal College of Surgeons in Ireland, Psychiatry, Dublin, Ireland Introduction The association between low birth weight and attention problems in childhood has been replicated many times (e.g. Momany, Kamradt, & Nikolas, 2018). However birth weight is unlikely the aetiological start-point of this association, as birth weight is itself the product of many prenatal factors e.g. gestational complications, maternal toxin exposure during pregnancy and basic demographics. Objectives We explore (1) which prenatal factors best predict attention problems in two independant population-based cohorts of children (2) which associations, if any, are moderated by sex and (3) we report accuracy statistics of our prenatal prediction algorithm for attention problems. Methods Participants were children aged 9 from ABCD study from the United States (N > 9,000) and the Growing Up in Ireland (GUI) study from Ireland (N > 6,000). Selected variables included familial pscyhiatric history, maternal smoking during gestation, prescription and non-prescription drug-use during gestation and a variety of gestational complications. All interactions with sex were also included. We used 5-fold cross-validation and elastic net regression (glmnet) to identify the optimal predictors of attention problems (measured by CBCL and SDQ). Results Strongest predictors of attention problems in the U.S. cohort included male sex, number of drugs used during pregnancy, number of family members with a history of mental illness, and number of gestational complications. Sex interacted with several of these risks. Protective factors included being a twin/triplet, being Asian, having higher household income and higher parental education level. Conclusions Several risk factors for childhood attention problems were identified across both cohorts, supporting their generalizabilty. Other findings were cohort-specific. Disclosure No significant relationships. https://www.cambridge.org/core/product/identifier/S0924933822003844/type/journal_articleadhdFoetal GrowthPrenatal Risksmachine learning |
| spellingShingle | N. Dooley M. Cannon D. Cotter M. Clarke Prediction of ADHD symptoms from prenatal data in two large population-based cohorts European Psychiatry adhd Foetal Growth Prenatal Risks machine learning |
| title | Prediction of ADHD symptoms from prenatal data in two large population-based cohorts |
| title_full | Prediction of ADHD symptoms from prenatal data in two large population-based cohorts |
| title_fullStr | Prediction of ADHD symptoms from prenatal data in two large population-based cohorts |
| title_full_unstemmed | Prediction of ADHD symptoms from prenatal data in two large population-based cohorts |
| title_short | Prediction of ADHD symptoms from prenatal data in two large population-based cohorts |
| title_sort | prediction of adhd symptoms from prenatal data in two large population based cohorts |
| topic | adhd Foetal Growth Prenatal Risks machine learning |
| url | https://www.cambridge.org/core/product/identifier/S0924933822003844/type/journal_article |
| work_keys_str_mv | AT ndooley predictionofadhdsymptomsfromprenataldataintwolargepopulationbasedcohorts AT mcannon predictionofadhdsymptomsfromprenataldataintwolargepopulationbasedcohorts AT dcotter predictionofadhdsymptomsfromprenataldataintwolargepopulationbasedcohorts AT mclarke predictionofadhdsymptomsfromprenataldataintwolargepopulationbasedcohorts |