Abstract
Background: Primary care clinicians (PCCs) frequently report low levels of comfort in managing seizure disorders, despite their essential role in caring for patients with seizures from systematically excluded groups. This study explores seizure care practices among PCCs at a federally qualified health center and examines possible associations between provider documentation of seizure care and proxy indicators for control of seizure disorder.
Methods: This is a retrospective observational cohort study of patients presenting with a seizure disorder between 2015 and 2019. Logistic regression models examined the associations between patient demographics, clinical processes and outcomes, and electronic medical record documentation.
Results: PCCs adhered to quality metrics for documentation in 41.5% of the 446 cases. PCCs documented seizure type 94.3% of the time but only recorded seizure frequency or time since last seizure 44.3% of the time. Most patients (52.3%) were referred to a neurologist. Better clinical documentation was significantly associated with referral to a neurologist (OR 1.73; 95%CI 1.14, 2.6; P = .009), visiting a neurologist (OR 1.7; 95%CI 1.13, 2.56; P = .01), and receiving a depression, anxiety, or quality of life screening (OR 2.97; 95%CI 1.84, 4.79; P < .001).
Discussion: Our findings suggest the need for improvements in the documentation of seizure management in the primary care setting. PCCs may benefit from additional training or tools to improve their comfort with seizure treatment and understanding of the role of primary care for these disorders.
- Cohort Studies
- Documentation
- Electronic Health Records
- Epilepsy
- Logistic Regression
- Primary Health Care
- Quality Improvement
- Quality of Care
- Seizures
Introduction
In the United States in 2021, an estimated 2.9 million individuals, or 1.1% of the adult population, had active epilepsy.1 Epilepsy* is characterized by recurrent seizures resulting from abnormal disruptions in the brain’s electric activity.3 Due to the unpredictability of when seizures occur, these disorders greatly impact quality of life. Among adults with active epilepsy and taking an antiepileptic drug (AED), 62% still reported interference with work, school, or social activities.4
Uncontrolled seizures account for up to 1 million emergency department (ED) visits annually.5 Several factors, including socioeconomic disadvantage and social isolation, contribute to poor seizure control by limiting access to specialty care and decreasing adherence to self-management.6 In addition, anxiety and depression, 2 common psychiatric comorbidities in epilepsy, often go undertreated, further diminishing quality of life.7
Primary care clinicians (PCCs) often serve as the first point of contact for patients with new onset seizures and provide routine care for approximately 40% of epilepsy patients.8 However, PCCs report a low level of comfort in managing seizure disorders, preferring instead to refer to a specialist.9,10 In the United States, seizure care in primary care settings for systematically and structurally excluded groups remains under-researched. International studies, however, suggest that PCCs can take on a more central role in caring for patients with limited access to specialty care clinicians.11 Thus, this study aims to support the provision of high-quality care by PCCs for seizure patients in underserved settings by 1) applying modified quality metrics to current practices at a large federally qualified health center (FQHC) network to explore current care and 2) examining possible association between documentation and (proxies for) control of seizure disorder.
Methods
Study Design and Approval
We conducted a retrospective chart review using electronic medical record (EMR) data. We generated a list of patients who received care at the FQHC for seizure or epilepsy between 2015 to 2019. The list included 1,892 charts from which we randomly selected 446. After excluding patients under 18 years old, those with seizures due to substance use/withdrawal, and those with febrile seizures, the final dataset included 388 charts. Data were stored in a secure electronic database.12 This study was approved by the FQHC’s Institutional Review Board.
Measures
Findings from an assessment of seizure care guidelines suggest that quality of care can be largely assessed by abstracting medical record data, and that appropriate documentation is the critical underlying component.13,14 In the absence of standard guidelines specifically for primary care, we utilized the American Academy of Neurology Institute’s (AANI) guidelines as a framework to establish the components of proper clinical documentation.13 AANI quality metrics include reports of seizure type, frequency of occurrence, and time since last seizure. Prior research suggests that collection of these metrics can result in better outcomes.15 Unlike the AANI visit-level numerators, we examined provider EMR documentation aggregated over a 4-year period. The resource-intensive chart review process required made this approach a more pragmatic choice given that patients averaged 19 office visits over the study period. Recognizing that neurology specialty-specific documentation standards may not directly translate to primary care, we modified the AANI criteria to more closely align with expected PCC documentation practices. We created a dichotomous variable indicating whether a provider documented seizure type, and a second variable indicating documentation of either seizure frequency or time since last seizure at any visit during the study period. Charts were considered ‘better documented’ if they contained both factors. Data on seizure freedom and tonic-clonic seizure reduction were not collected during the initial chart review.
Variables of Interest
Using the AANI quality measure framework and consultation with PCCs, we identified potential variables of interest including patient demographics, seizure-related diagnostic information, current and historic use of AEDs, engagement with neurology, and completion of routine screenings/preventive care services.
Lastly, we examined control of seizure disorders via proxy indicators, including patient ED visits with an associated diagnosis of seizure disorder or epilepsy during the study period. All other and unknown reasons for ED use were classified as nonseizure related. We also collected data on relevant AEDs included on a patient’s medication list. The quantity of AEDs started and ended during the study period served as another proxy for control of seizure disorder, with medication changes representing potentially uncontrolled seizures.16 We coded patients with no prescribed AED, one AED, or only one discontinued AED as having a stable medication regimen.
Data Analysis
We examined descriptive statistics of demographics and variables of interest for our sample. Two logistic regression models examined the variables of interest and clinician documentation. The first unadjusted for confounders and the second adjusted for age-group, race and ethnicity, and insurance status, with an α = 0.05 to establish statistical significance. Analyses were conducted using STATA/BE 17.0.17
Results
Patient Characteristics
Patient demographic characteristics are shown in Table 1. Approximately 69% of the patients were between the ages of 30 to 64 at the time of their first encounter, 50.5% identified as male, 57.2% as non-White, and 85% were publicly insured. Patient comorbidities included anxiety or post-traumatic stress disorder (PTSD) (44%), depression (44%), other mental health disorders (39%), tobacco use (21%), and intellectual or developmental disabilities (23%).
Demographic Characteristics of Patients (n = 388)
Provider Documentation, Processes, and Outcomes
Table 2 presents the results related to the first 2 AANI process and outcome measures, along with other potentially clinically significant metrics. Although clinicians documented seizure type 94.3% of the time, seizure frequency/time since last seizure were only recorded 44.3%, and both seizure type and frequency 41.5% of the time. The majority of patients received a referral to a neurologist (52.3%), though we could only confirm visit completion for fewer than half of referred patients. Among the patients who visited a neurologist, 7.7% of charts included a discrepancy between the PCC and neurologist’s documentation. Depression screening was conducted for 63.9% of patients, but only 29.1% and 15.7% received a quality-of-life or anxiety screening, respectively. Over the study period, 80% of patients visited the ED, with 26.8% of these visits attributed to seizure disorders or epilepsy. In addition, 31.4% of patients had unstable AEDs throughout the study.
Provider Adherence to Clinical Guidelines
Logistic Regression
Univariate logistic regression analysis examined associations between provider documentation and key variables of interest (Table 3). We identified significant associations between better documentation and referral to a neurologist (OR 1.73; 95% CI 1.14, 2.6; P = .009), visiting a neurologist (OR 1.7; 95% CI 1.13, 2.56; P = .01), and receiving a relevant screening (OR 2.97; 95% CI 1.84, 4.79; P < .001). Results remained the same when adjusted for other possible covariates. We found no clinically significant associations between better documentation and stable medication (OR 0.98; 95% CI 0.63, 1.51), using the ED for a seizure-related disorder (OR 0.72; 95% CI 0.44, 1.18), or being overdue for preventative services (OR 1.15; 95% CI 0.75, 1.75).
Logistic Regression Analysis Results
Discussion
Our findings demonstrate a critical need for enhanced documentation practices among PCCs managing seizure disorders. In particular, data capturing frequency/time since last seizure were inconsistently recorded, affecting the ability to monitor and compare outcomes over time.13 Clinicians also did not consistently document key elements of care such as safety counseling, anxiety screenings, and quality-of-life assessments, which are essential to providing comprehensive care and should be within the PCCs' existing scope.
The statistically significant associations between more thorough clinical documentation and positive outcomes including referral/visit to neurology and administration of relevant screenings suggest pathways for future study. In general, these data suggest that PCCs may benefit from additional training and tools to improve comfort in caring for patients with seizure disorders. These findings, including discrepancies between PCC and neurology notes, also suggest the need for improved integration of specialty care documentation within primary care settings. PCCs should also be trained to address other disruptions in care due to changes in insurance plans or formulary restrictions, which may require changes to an otherwise stable AED regimen or transfer to a different specialist.
Strengths and Limitations
Our study has several limitations. The generalizability of findings may be limited due to sampling from a single network, and the retrospective nature/use of EMR data, which relies on the information being available and correctly identified by reviewers. These data do not allow us to assess causal pathways, or to account for other factors, including the FQHC’s robust family medicine training program. Given the high number of encounters where seizure frequency/time since last seizure was not documented, it is possible that in some cases breakthrough seizures were not identified, influencing clinical decision making. Systems-level issues with medical record sharing and documentation practices also limited our ability to assess possible associations between seizure control and neurology referrals/visits. A strength of this study is that the FQHC network provides care for a diverse patient population. This research within the primary care context is novel and provides important data and implications for improving practice. Ultimately, the team reviewed over 20% of randomly selected charts, with manual review providing access to data not accessible via standard reporting.
Implications for Practice
Future research should consider applying, or appropriately modifying and applying, AANI or other quality metrics to clinical practice and developing system-based solutions (including SmartSets, Order Sets, and EMR-embedded prompts) to assist clinicians and clinical staff in capturing and using critical data relevant to the primary care setting. Primary care training programs may also consider exploring opportunities to increase provider comfort with these disorders, and ways to strengthen the exchange of information with specialty care practices.
Acknowledgments
The authors thank Nicole Nurse, Maxine Golub, Noreen Singh, Debra Chan, Hector Dominguez, Lexie Martocci, Talia Baurer, and Tiffany Lukenda for their assistance with this project.
Notes
This article was externally peer reviewed.
*Although a seizure is single occurrence and epilepsy is a neurological condition characterized by two or more unprovoked seizures, much of the literature uses the terms “seizure disorder” and “epilepsy” somewhat interchangeably when referring to any ongoing neurological condition with ongoing unprovoked seizures.2 We follow this convention and use these terms similarly throughout.
Funding: This research was funded by the Sergievsky Trust.
Conflict of interest: The authors have no relevant financial or nonfinancial interests to disclose.
- Received for publication August 26, 2024.
- Revision received February 14, 2025.
- Accepted for publication March 3, 2025.
