£6,000 PhD Scholarship | APPLY NOW £6,000 PhD Scholarship | APPLY NOW

Developing a risk stratification tool to detect ADHD in children and adolescents


   College of Medicine, Veterinary and Life Sciences

  Dr Michael Fleming  Applications accepted all year round  Self-Funded PhD Students Only

About the Project

Background: ADHD is associated with adverse impacts on health, education, and employment however there are currently delays, of sometimes years, before it is diagnosed and therefore managed, causing preventable distress to the child, family, and teachers as well as lasting psychological, educational, and social disadvantage. We hypothesize that development of a risk stratification tool will enable ADHD to be detected and managed earlier; thereby reducing the adverse impact on affected children and their families. 

Aims: We will undertake individual-level record linkage of several Scotland-wide education and health databases. Education records (including exam results, absenteeism, exclusion, additional support needs and leaver destination) for all pupils attending school in Scotland between 2009 and 2020 will be linked to prescribing data, maternity records, neonatal admissions, child health records, acute and psychiatric hospitalisations, and deaths. We will ascertain cases of ADHD using prescribing data to identify children dispensed one or more medications approved solely for the treatment of ADHD: methylphenidate, dexamphetamine, atomoxetine, or lisdexamphetamine. The Scottish pupil census holds information on all children attending local authority maintained primary, secondary, and special schools in Scotland which covers 95% of the school aged (4-19 years) population. Accessing data between 2009 and 2020 will yield linked records pertaining to over 1 million schoolchildren. Our previous work investigating educational and health outcomes of schoolchildren treated for ADHD uncovered an ADHD prevalence of 1.0%; therefore, using the same methodology, we expect to identify more than 10,000 children with ADHD. After initial data cleaning, merging and recoding, we will firstly determine the risk factors associated with ADHD including maternal medication, maternal antecedents (smoking, age, parity, previous abortions), pregnancy outcomes (birthweight and intrauterine growth restriction, Apgar score, mode of delivery, gestational age), early life hospitalisations (neonatal, acute, psychiatric), early life growth trajectories and development (pre-school cognitive measures), early life injury/trauma (hospitalisations), childhood medication for other chronic conditions (depression, anxiety, asthma), sociodemographic factors, and school progress(absenteeism, exclusion, special educational need, attainment, and unemployment on leaving school).

To explore the development of a risk stratification tool we will randomly split the data into training, validation, and test datasets. After appropriate transformation and scaling of data, we will train and fine-tune several classifiers (e.g. logistic regression, linear discriminant analysis, support vector machines (SVM) and random forests) to predict the outcome of ADHD, using K-fold cross validation to reduce the risk of overfitting. Each classifier will be evaluated using the confusion matrix to derive estimates of precision (true positives divided by the sum of true and false positives) and recall (true positives divided by the sum of true positives and false negatives). This metric is preferred to receiver operating characteristic curves when the class that is being predicted is rare. We will select the appropriate threshold for classification based on inspection of precision-recall versus threshold plots and precision versus recall curves. Should the individual classifiers prove a mediocre fit, we will explore the possibility of further development and evaluation using ensemble methods, which often produce better predictions than one preferred classifier. The preferred model will be useful to clinicians to help identify children who require further investigation to enable earlier diagnoses of ADHD. Analyses will most likely be performed using R and Anaconda Python.

Training outcomes: The student will undergo training (via courses and self-learning) in the following: Safe researcher training, R programming, statistical methods, data linkage methods, analysing ‘big’ data, machine learning techniques, additional statistical programming packages (if needed) such as SPSS, Stata, SAS, and python.

Apply Now


References

1. https://jamanetwork.com/journals/jamapediatrics/fullarticle/2624340
2. https://journals.sagepub.com/doi/pdf/10.1177/1087054711427563
3. https://www.frontiersin.org/articles/10.3389/fnhum.2020.560021/full
4. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7259455/pdf/ndt-16-1331.pdf
PhD saved successfully
View saved PhDs