This study was funded by Shire Pharma Canada ULC, now part of Takeda
ADHD is frequently managed in primary care; however, primary care providers often have limited formal training in managing mental health conditions in children and adolescents (Sultan et al, 2018). As ADHD can negatively impact development, education, occupational success and socialisation, effective identification and management of this condition is essential. The purpose of this study was to develop and validate an algorithm to identify individuals with an ADHD diagnosis in an electronic medical record (EMR) in Canada and report on the epidemiology of this disorder in primary care.
Data were obtained from the Canadian Primary Care Sentinel Surveillance Network (CPCSSN), a Canadian organisation comprised of 11 practice-based research networks that extract primary care data from EMRs. Data from six of 11 research networks were evaluated in this study conducted between 2008 and 2015. A 2-year contact group was used to define the yearly practice populations, which included any individual with an encounter in the year of study or the year prior.
An individual was classified as having ADHD if he/she was aged ≥4 years and the medical record included ICD-9 code 314 in one or more visits and one or more prescriptions for ADHD-related medications, or the medical record listed ICD-9 code 314 in two or more visits. The algorithm was validated via a manual electronic chart review of 492 individuals from one clinic: 246 individuals identified by the algorithm as having ADHD and 246 identified as not having ADHD by the algorithm. The positive predictive value of the algorithm was calculated based on the number of patients classified as having ADHD by both the algorithm and the chart reviewer out of the total number of patients identified as having ADHD by the algorithm.
Epidemiology of ADHD
Due to wider use of EMRs and increases in the number of practices providing patient data to CPCSSN, the number of individuals with ADHD in the database increased from 345,173 to 624,419 between 2008 and 2015.
Manual chart review found that 236 of the 246 individuals identified as having ADHD by the algorithm had an ADHD diagnosis (true positives); and 237 of the 246 patients that the algorithm identified as not having ADHD had no ADHD diagnosis (true negatives). The majority of the 10 individuals misclassified by the algorithm as having ADHD had been evaluated for ADHD, but a diagnosis or treatment plan was not provided. Most of the nine ADHD cases missed by the algorithm had been managed by a specialist, and as such minimal data were recorded in the primary care EMR. The EMR algorithm was found to have a 95.9% positive predictive value (95% confidence interval [CI] 92.6–98.0) and a 96.3% negative predictive value (95% CI 93.2–98.3).
The prevalence of ADHD was highest in children aged 4 to 17 years in each year of the study, but the prevalence also rose in young adults aged 18 to 34 years. A small increase in ADHD prevalence was observed in those aged 35 to 64 years. As seen in the literature, the prevalence of ADHD was found to be higher in males than females in the younger age groups, but this gap was notably smaller in the older age groups. In patients aged 4 to 17 years the ratio of male to female ADHD diagnoses increased slightly from 2008 (1.3:1) to 2015 (1.4:1) and in young adults the gap widened.
Limitations of the study include that the algorithm was validated by manual chart review from only one clinic using one EMR product, and data were limited to a sample of primary care clinics from five Canadian provinces. Therefore, results may not be applicable to the wider population.
The ADHD case-finding algorithm is a reliable tool for assessing the prevalence of ADHD in primary care practices in Canada. The algorithm revealed that the prevalence of individuals with a diagnosis of ADHD increased between 2008 and 2015 likely due to improved identification and treatment of ADHD in primary care, however, a significant gap in diagnosing ADHD in females remains. The algorithm may be a useful tool in estimating the burden of ADHD, so that resources and interventions can be allocated appropriately to improve outcomes in individuals with ADHD managed in primary care.
Read more about identifying ADHD in primary care here
Morkem R, Handelman K, Queenan JA, et al. Validation of an EMR algorithm to measure the prevalence of ADHD in the Canadian Primary Care Sentinel Surveillance Network (CPCSSN). BMC Med Inform Decis Mak 2020; 20: 166.
Sultan MA, Pastrana CS, Pajer KA. Shared care models in the treatment of pediatric attention-deficit/hyperactivity disorder (ADHD). Are they effective? Health Serv Res Manag Epidemiol 2018; 5: 2333392818762886.