Variations in Prescribing Practices for Early Life Epilepsy, a Pediatric Epilepsy Learning Healthcare System Study
Abstract number :
1.304
Submission category :
7. Anti-seizure Medications / 7E. Other
Year :
2022
Submission ID :
2204923
Source :
www.aesnet.org
Presentation date :
12/3/2022 12:00:00 PM
Published date :
Nov 22, 2022, 05:27 AM
Authors :
Alan Wu, MS – Weill Cornell Medicine; Erika Axeen, MD – University of Virginia; Sonam Bhalla, MD – Emory University/ Children’s Healthcare of Atlanta; Jason Coryell, MD – Oregon Health & Sciences University, Portland, OR; Scott Demarest, MD – Children's Hospital Colorado; William Gaillard, MD – Children's National Hospital; Howard Goodkin, MD – University of VIrginia; Ingo Helbig, MD – Children's Hospital of Philadelphia; Shaun Hussain, MD – University of California, Los Angeles; Tobias Loddenkemper, MD – Childrens Harvard; Juma Mbwana, MD – Children's National Hospital; Lindsey Morgan, MD – Seattle Childrens; Anup Patel, MD – Nationwide Childrens; Scott Perry, MD – Cook Childrens; Tristan Sands, MD – Columbia University; Renee Shellhaas, MD – University of Michigan; Nilika Singhal, MD – University of California San Francisco; Joyce Wu, MD – Lurie Childrens; Elissa Yozawitz, MD – Montefiore Medical Center; Muhammad Zafar, MD – Duke University; Zachary Grinspan, MD – Weill Cornell Medicine
Rationale: Understanding variations in healthcare delivered to children with epilepsy is foundational to design comparative effectiveness studies and to develop initiatives to improve health equity. Here we analyze variations in prescribing patterns for the first anti-seizure-medication (ASM) in young children.
Methods: We used electronic health record data from 17 centers in the Pediatric Epilepsy Learning Healthcare System from 2017 to 2020 to examine factors associated with the first ASM prescribed for children with epilepsy, 0 to 5 years old. We reviewed four commonly prescribed ASMs: levetiracetam (LEV), clobazam (CLB), oxcarbazepine (OXC), and phenobarbital (PB). We used bivariate and multivariable analyses. In the multivariable analyses, for each ASM, we used logistic regression (prescribed vs not prescribed) to look for drivers of medication selection, including age, sex, center, race/ethnicity category, insurance type, and a zip-code level measure of social determinants of health called the Child Opportunity Index (COI). We excluded individuals with no zip code. We referred to children with public insurance plus another type of insurance as having “mixed” insurance.
Results: The first prescribed medication was LEV for 5358 children (median age, 2.6 [IQR = 1.4, 3.8] years; 43% female (F)), 496 CLB (2.9 [1.7, 3.9] years; 47% F), 1458 OXC (3.3 [2.2, 4.2] years; 44% F), 498 PB (0.45 [0.1, 1.2] years; 42% F), and 1663 an other ASM (3.4 [1.8,4.3] years; 43% female). In bivariate analyses, age, center, race/ethnicity category, insurance, and COI were associated with ASM selection. There were large variations across centers: the rate of prescription for LEV varied from 43% to 72%, CLB 1.7% to 10%, OXC 4.1% to 28%, and PB 0.7% to 12% (Table 1).
In the multivariable analyses, center and age were associated with ASM selection for each medication. Race/ethnicity category was associated with the selection of LEV (more commonly prescribed for Black compared to White Non-Hispanic children; OR 1.26 [95% CI: 1.1-1.44]) and CLB (less commonly prescribed for Black compared to White Non-Hispanic children; OR 0.63 [0.45, 0.86]).
Compared to children with private insurance, LEV was more commonly prescribed for children with public only insurance (OR 1.16 [1.01, 1.34]) and less commonly prescribed for children with mixed insurance (OR 0.88 [0.77, 1.0]); CLB more commonly prescribed for children with mixed insurance (OR 1.59 [1.19, 2.14]), and PB more commonly prescribed for children with mixed insurance (OR 1.75 [1.23, 2.51]). Sex and COI were not associated with ASM selection (Table 2).
Conclusions: There are substantial variations in first ASM selection for young children with epilepsy. Drivers may include age, physician preference (center-to-center variations), ability to pay (insurance status), and race/ethnicity. Additional work is necessary to incorporate clinical drivers of ASM selection into our models (i.e., seizure type and burden, epilepsy etiology, and comorbidity profile). These data highlight the need for comparative effectiveness research to understand if variations in prescribing practices are linked to variations in outcomes.
Funding: Pediatric Epilepsy Research Foundation, Morris and Alma Shapiro Fund
Anti-seizure Medications