Assessing machine learning for fair prediction of ADHD in school pupils using a retrospective cohort study of linked education and healthcare data
Objectives Attention deficit hyperactivity disorder (ADHD) is a prevalent childhood disorder, but often goes unrecognised and untreated. To improve access to services, accurate predictions of populations at high risk of ADHD are needed for effective resource allocation. Using a unique linked health...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMJ Publishing Group
2022-12-01
|
| Series: | BMJ Open |
| Online Access: | https://bmjopen.bmj.com/content/12/12/e058058.full |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832573912456626176 |
|---|---|
| author | Johnny Downs Robert Stewart Alice Wickersham Sumithra Velupillai Lucile Ter-Minassian Natalia Viani Lauren Cross |
| author_facet | Johnny Downs Robert Stewart Alice Wickersham Sumithra Velupillai Lucile Ter-Minassian Natalia Viani Lauren Cross |
| author_sort | Johnny Downs |
| collection | DOAJ |
| description | Objectives Attention deficit hyperactivity disorder (ADHD) is a prevalent childhood disorder, but often goes unrecognised and untreated. To improve access to services, accurate predictions of populations at high risk of ADHD are needed for effective resource allocation. Using a unique linked health and education data resource, we examined how machine learning (ML) approaches can predict risk of ADHD.Design Retrospective population cohort study.Setting South London (2007–2013).Participants n=56 258 pupils with linked education and health data.Primary outcome measures Using area under the curve (AUC), we compared the predictive accuracy of four ML models and one neural network for ADHD diagnosis. Ethnic group and language biases were weighted using a fair pre-processing algorithm.Results Random forest and logistic regression prediction models provided the highest predictive accuracy for ADHD in population samples (AUC 0.86 and 0.86, respectively) and clinical samples (AUC 0.72 and 0.70). Precision-recall curve analyses were less favourable. Sociodemographic biases were effectively reduced by a fair pre-processing algorithm without loss of accuracy.Conclusions ML approaches using linked routinely collected education and health data offer accurate, low-cost and scalable prediction models of ADHD. These approaches could help identify areas of need and inform resource allocation. Introducing ‘fairness weighting’ attenuates some sociodemographic biases which would otherwise underestimate ADHD risk within minority groups. |
| format | Article |
| id | doaj-art-9838b0378d0543a8b7cd45ec2303a585 |
| institution | Kabale University |
| issn | 2044-6055 |
| language | English |
| publishDate | 2022-12-01 |
| publisher | BMJ Publishing Group |
| record_format | Article |
| series | BMJ Open |
| spelling | doaj-art-9838b0378d0543a8b7cd45ec2303a5852025-02-02T02:10:13ZengBMJ Publishing GroupBMJ Open2044-60552022-12-01121210.1136/bmjopen-2021-058058Assessing machine learning for fair prediction of ADHD in school pupils using a retrospective cohort study of linked education and healthcare dataJohnny Downs0Robert Stewart1Alice Wickersham2Sumithra Velupillai3Lucile Ter-Minassian4Natalia Viani5Lauren Cross63 South London and Maudsley NHS Foundation Trust, NIHR Maudsley Biomedical Research Centre, London, UKInstitute of Psychiatry, Psychology and Neuroscience, King`s College London, London, UKCAMHS Digital Lab, Department of Child and Adolescent Psychiatry, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, London, UKPsychological Medicine, Institute of Psychiatry, Psychology and Neuroscience, King`s College London, London, UKDepartment of Psychological Medicine, King’s College London, London, UKDepartment of Psychological Medicine, King’s College London, London, UKDepartment of Psychological Medicine, King’s College London, London, UKObjectives Attention deficit hyperactivity disorder (ADHD) is a prevalent childhood disorder, but often goes unrecognised and untreated. To improve access to services, accurate predictions of populations at high risk of ADHD are needed for effective resource allocation. Using a unique linked health and education data resource, we examined how machine learning (ML) approaches can predict risk of ADHD.Design Retrospective population cohort study.Setting South London (2007–2013).Participants n=56 258 pupils with linked education and health data.Primary outcome measures Using area under the curve (AUC), we compared the predictive accuracy of four ML models and one neural network for ADHD diagnosis. Ethnic group and language biases were weighted using a fair pre-processing algorithm.Results Random forest and logistic regression prediction models provided the highest predictive accuracy for ADHD in population samples (AUC 0.86 and 0.86, respectively) and clinical samples (AUC 0.72 and 0.70). Precision-recall curve analyses were less favourable. Sociodemographic biases were effectively reduced by a fair pre-processing algorithm without loss of accuracy.Conclusions ML approaches using linked routinely collected education and health data offer accurate, low-cost and scalable prediction models of ADHD. These approaches could help identify areas of need and inform resource allocation. Introducing ‘fairness weighting’ attenuates some sociodemographic biases which would otherwise underestimate ADHD risk within minority groups.https://bmjopen.bmj.com/content/12/12/e058058.full |
| spellingShingle | Johnny Downs Robert Stewart Alice Wickersham Sumithra Velupillai Lucile Ter-Minassian Natalia Viani Lauren Cross Assessing machine learning for fair prediction of ADHD in school pupils using a retrospective cohort study of linked education and healthcare data BMJ Open |
| title | Assessing machine learning for fair prediction of ADHD in school pupils using a retrospective cohort study of linked education and healthcare data |
| title_full | Assessing machine learning for fair prediction of ADHD in school pupils using a retrospective cohort study of linked education and healthcare data |
| title_fullStr | Assessing machine learning for fair prediction of ADHD in school pupils using a retrospective cohort study of linked education and healthcare data |
| title_full_unstemmed | Assessing machine learning for fair prediction of ADHD in school pupils using a retrospective cohort study of linked education and healthcare data |
| title_short | Assessing machine learning for fair prediction of ADHD in school pupils using a retrospective cohort study of linked education and healthcare data |
| title_sort | assessing machine learning for fair prediction of adhd in school pupils using a retrospective cohort study of linked education and healthcare data |
| url | https://bmjopen.bmj.com/content/12/12/e058058.full |
| work_keys_str_mv | AT johnnydowns assessingmachinelearningforfairpredictionofadhdinschoolpupilsusingaretrospectivecohortstudyoflinkededucationandhealthcaredata AT robertstewart assessingmachinelearningforfairpredictionofadhdinschoolpupilsusingaretrospectivecohortstudyoflinkededucationandhealthcaredata AT alicewickersham assessingmachinelearningforfairpredictionofadhdinschoolpupilsusingaretrospectivecohortstudyoflinkededucationandhealthcaredata AT sumithravelupillai assessingmachinelearningforfairpredictionofadhdinschoolpupilsusingaretrospectivecohortstudyoflinkededucationandhealthcaredata AT lucileterminassian assessingmachinelearningforfairpredictionofadhdinschoolpupilsusingaretrospectivecohortstudyoflinkededucationandhealthcaredata AT nataliaviani assessingmachinelearningforfairpredictionofadhdinschoolpupilsusingaretrospectivecohortstudyoflinkededucationandhealthcaredata AT laurencross assessingmachinelearningforfairpredictionofadhdinschoolpupilsusingaretrospectivecohortstudyoflinkededucationandhealthcaredata |