Authors :
Presenting Author: Colin Josephson, MSc, MSc, MD – University of Calgary
Marta Berglund, MSc – University College London; Spiros Denaxas, PhD – University College London; Arturo Gonzalez-Izquierdo, PhD – University College London; Tolulope Sajobi, PhD – University of Calgary; Samuel Wiebe, MD MSc – University of Calgary
Rationale:
The incidence of late onset epilepsy (LOE) continues to rise leading to significant utilization of health care resources by these patients. We sought to delineate the trajectories of health care use (HCU) before and after diagnosis and compare them to those of people without LOE.
Methods:
We performed a retrospective observational cohort study linking primary care and hospital UK electronic health records to vital statistics data covering years 1998 through 2019. These patients were matched 1:10 to controls by age, sex, and primary care physician. The primary outcome was all health care visits (weighted by the inverse proportions of outpatient care, admitted patient care, and Accident & Emergency [A&E] to primary care) in one year intervals for the four years before and four years after diagnosis of LOE. An interrupted time series analysis was used to assess whether was a significant change in the trajectory of HCU between LOE and controls. We then calculated the incidence rate ratio (IRR) for HCU following diagnosis/matching date using a mixed effects negative binomial regression adjusted for sociodemographic and clinical variables. Secondary analyses repeated the above steps but concentrated on specific forms of HCU (A&E, admitted patient care, and outpatient care).
Results:
The linked cohort contained 18,410,930 people, of whom 2,569,874 were 65+ years old and, of those, 1,048 (4%) developed incident LOE. Median age of onset was 69 (IQR 6) and 474 (45%) were female. Primary care and outpatient visits were the main drivers of HCU. For LOE, HCU increased by a median of five visits per year (95% confidence interval [95%CI] 1-9; p=0.026) up to the point of diagnosis. During the year of diagnosis, HCU increased by a median of 10 HCU interactions per patient (95%CI 5-14; p=0.003). Annual per person HCU interactions dropped by a median of eight visits per person per year (95%CI 4-11; p=0.001) following the diagnosis of LOE. Conversely, there was no pre-post matching date change in median HCU interactions for the control population (median increase of 1 HCU interaction, 95%CI -1 to 3; p=0.254; Figure 1). Weighted HCU remained higher for people with LOE (adjusted IRR 1.91; 95%CI 1.80-2.04; p< 0.001) compared to those without epilepsy over the four years following diagnosis. Notably, those with LOE have higher rates of A&E attendances per person-year by 195% (adjusted IRR 2.95, 95%CI 2.40-3.63; p< 0.001) and admitted patient care by 78% (adjusted IRR 1.78, 95%CI 1.26-2.52; p=0.001) compared to controls over the four years following diagnosis.