Predictors of persisting symptoms after concussion in children following a traumatic brain injury: a longitudinal retrospective cohort study.

Abstract

Objectives: To identify predictors of persisting symptoms after concussion (PSaC) in children, following any medically attended traumatic brain injury (TBI).

Design: Retrospective cohort study.

Setting: Linked primary and secondary care data from UK Clinical Practice Research Datalink and Hospital Episode Statistics.

Participants: Children aged 1-17 years with a medically attended TBI between 2013 and 2017.

Main outcome measure: A binary indicator of PSaC or suspected PSaC, measured using either a clinical code for PSaC or medical attendances for one or more PSaC symptoms 3-12 months after TBI.

Results: We identified 137 873 children with a TBI; 4620 (3.4%) had PSaC or suspected PSaC. More females (3.8%) had PSaC than males (3.1%). Those with PSaC were older at the time of TBI compared with those without PSaC (8 vs 5.5 years). In a multivariable logistic regression model, older age (OR =1.02 per year increase in age, 95% CI 1.01 to 1.03), female sex (OR=1.20, 95% CI 1.13 to 1.28), being Asian (OR=1.37, 95% CI 1.22 to 1.54) or mixed ethnicity (OR=1.18, 95% CI 1.01 to 1.37) (compared with white ethnicity), having a history of headaches (OR=3.52, 95% CI 3.13 to 3.95), learning disabilities (OR=2.06, 95% CI 1.69 to 2.52), ADHD (OR=2.41, 95% CI 1.91 to 3.04), anxiety (OR=2.58, 95% CI 2.18 to 3.05), depression (OR=4.00, 95% CI 3.28 to 4.89) or sleep disorders (OR=2.35, 95% CI 1.99 to 2.78) were associated with increased odds of PSaC.

Conclusions: These results may be used to identify children more likely to develop PSaC following a TBI and those who may benefit from targeted healthcare for PSaC symptoms. Identifying cases of PSaC in primary care data was challenging as perhaps many children do not attend services for suspected PSaC or, if they did, are not diagnosed with PSaC. Furthermore, the clinical predictors are a measure of healthcare access for these symptoms; thus, results could be influenced by patient or carer's health-seeking behaviour.