Screening and Management of Depression and Anxiety in People With Epilepsy: A Quality Improvement Study

Abstract

Background:

Due to the high prevalence of depression and anxiety in people with epilepsy, the International League Against Epilepsy Commission on the Neuropsychiatric Aspects of Epilepsy recommends implementing routine screening for depression and anxiety symptoms. Our epilepsy group began administering three screening questionnaires to all clinic patients in 2016: the Neurological Disorders Depression Inventory for Epilepsy (NDDI-E), the Beck Anxiety Inventory (BAI) and the Generalized Anxiety Disorder-7 (GAD-7).

Objective:

We aim to review our experience with this screening approach.

Methods:

We reviewed 2253 sets of questionnaires completed from January 2018 to March 2020 and studied the actions taken by epileptologists in response to a positive screening.

Results:

Thirty-six percent of all assessed patients screened positive on at least one questionnaire: 13.6% screened positive for depression symptoms (NDDI-E ≥ 16), 12.3% for anxiety symptoms (BAI ≥ 22) and 30.3% for GAD symptoms (GAD-7 > 7). Among patients with a positive screening, 36% received a care intervention, 59% did not and 5% declined the neurologist’s recommendation. Among patients for whom an intervention was implemented, 58% were referred to a mental health professional (generally a neuropsychiatrist), 29% had their antiseizure medication adjusted to alleviate their symptoms and 13% received another intervention.

Conclusion:

In our clinic, an important proportion of patients screened positive for depression and/or anxiety symptoms. Fewer than half received a management option to alleviate their symptoms. We conclude that while routine screening increases the detection of depression and anxiety among epilepsy patients, it must be accompanied by effective interventions and access to mental-health professionals.

Résumé

La dépression et l’anxiété sont fréquentes chez les personnes atteintes d’épilepsie.

Dans ce contexte, la Ligue internationale contre l’épilepsie recommande le dépistage systématique de ces symptômes. En 2016, nous avons instauré dans notre clinique d’épilepsie un dépistage de routine à l’aide de trois questionnaires : le Neurological Disorders Depression Inventory for Epilepsy (NDDI-E), le Beck Anxiety Inventory (BAI) et le Generalized Anxiety Disorder-7 (GAD-7).

Décrire notre expérience clinique avec ce dépistage systématique.

Nous avons analysé 2253 questionnaires complétés entre janvier 2018 et mars 2020, ainsi que les interventions des épileptologues après un résultat de dépistage positif.

Trente-six pourcent des patients ont obtenu un résultat positif à au moins un questionnaire. Des symptômes de dépression ont été identifiés chez 13,6 % des patients (NDDI-E ≥ 16), des symptômes d’anxiété chez 12,3 % (BAI ≥ 22) et des symptômes d’anxiété généralisée chez 30,3 % (GAD-7 > 7). Parmi les patients ayant un dépistage positif, 36 % ont bénéficié d’une prise en charge clinique, tandis que 59 % n’en ont pas reçu et que 5 % ont refusé la recommandation du neurologue. L’intervention la plus fréquente est la référence à un professionnel de santé mentale (58 %). La médication anticrise a été modifiée chez 29 % des patients, et 13 % ont bénéficié d’une autre forme d’intervention.

Bien que le dépistage systématique permette une meilleure identification des symptômes anxiodépressifs chez les personnes atteintes d’épilepsie, son impact clinique demeure limité sans interventions adaptées et un accès adéquat aux ressources en santé mentale.

Highlights

• Screening enabled the detection and management of depression and anxiety in many patients with epilepsy.

• Without additional interventions, screening alone is unlikely to ensure comprehensive management of depression and anxiety symptoms.

• The infrequent prescription of psychiatric medications reflects discomfort in managing such symptoms, highlighting the need for additional training.

Introduction

People with epilepsy (PWE) experience depression and anxiety symptoms at a higher rate than the general population. ¹ A recent meta-analysis revealed that the prevalence of depression and anxiety disorders in PWE was 22.9 and 20.2%, respectively. ² Although depression and anxiety are highly prevalent among PWE, these psychiatric disorders are underrecognized and undertreated. ^2,3 This oversight may be explained by the perception of neurologists that they have limited time for psychiatric evaluation in the context of busy epilepsy clinics, and that seizure control is the primary measure of treatment success. ⁴ Studies have shown, however, that depression and anxiety, like seizure control, greatly impact quality of life. ⁵ Furthermore, addressing these conditions can improve seizure control. ^6–8 Undertreated depression and anxiety in epilepsy are associated with work absenteeism, increased healthcare utilization and higher direct medical costs. ^5,9 These findings highlight the need for a more comprehensive approach to epilepsy treatment, one that includes the identification and management of psychiatric comorbidities to improve overall patient outcomes.

In 2011, an international consensus group of epilepsy experts published the International Consensus Clinical Practice Statements for the Treatment of Neuropsychiatric Conditions Associated with Epilepsy, ³ an evidence- and practice-based document developed to address the impact of neuropsychiatric disorders on epilepsy management and quality of life.

Among the statements in the document was a recommendation for routine screening of depression and anxiety, specifically using the Neurological Disorders Depression Inventory for Epilepsy (NDDI-E), Patient Health Questionnaire, or equivalent. Furthermore, the document stressed the importance of prompt management of neuropsychiatric symptoms when detected, with multiple statements offering general guidelines for their management.

Following the recommendation of the International Consensus Clinical Practice Statements, our epilepsy group at The University of Montreal hospital center (CHUM) epilepsy clinic implemented routine screening for depression and anxiety in 2016, which was maintained under March 2020, when the COVID-19 pandemic led to dramatic changes in clinic functioning. In this study, we assess the impact of implementing these screening questionnaires in our clinic by examining whether the screening identified symptoms among patients and how neurologists subsequently responded to these findings.

Materials and methods

Study context

The study took place at the CHUM epilepsy clinic, which is the outpatient arm of our Level Four Epilepsy Center. All neurologists participating in the clinic are fellowship-trained epileptologists. The clinic serves a wide range of patients, including patients undergoing diagnostic workup, patients with well-controlled epilepsy, and patients with complex epilepsy, including those undergoing surgical evaluation. General neurology patients may also occasionally be seen during epilepsy clinics. Routine screening was initiated in 2016; however, only paper questionnaires from January 2018 to March 2020 were stored in the clinic. Earlier forms were not conserved; analyses were therefore limited to this period.

This study was conducted as a quality improvement (QI) project and was exempt from formal Research Ethics Board approval in accordance with institutional policy.

Data collection

At each patient visit, as part of routine clinical follow-up, receptionists provided three paper-based screening questionnaires to patients: the NDDI-E, the Generalized Anxiety Disorder-7 (GAD-7) and the Beck Anxiety Inventory (BAI). Patients were instructed to complete these questionnaires in the waiting room and give them to their epileptologist at the start of their visit. The neurologist reviewed the form and calculated the score to determine whether the screen was positive. The dataset included both patients with a single visit and patients with multiple visits during the study period. Each visit with a completed questionnaire was treated as an individual encounter and included in the analysis.

Testing methods

The NDDI-E, GAD-7 and BAI questionnaires were used to screen for symptoms of depression and anxiety. These tools were selected based on several key considerations: their reliance on self-reported symptoms, their validation in PWE and their simplicity, making them easy for patients to complete within a short time.

The NDDI-E is a screening tool specifically designed to detect depression symptoms in PWE ¹⁰ (Supplementary Tables 1 and 2). It consists of questions that assess depressive symptoms (e.g., sadness, anhedonia, hopelessness…) while minimizing overlap with cognitive deficits of epilepsy and adverse effects of antiseizure medications. ¹⁰ For example, symptoms such as fatigue, poor concentration and memory problems may arise from epilepsy itself or from side effects of antiseizure medications, rather than from comorbid depression. In contrast, sadness, anhedonia and hopelessness are core features of the NDDI-E that more specifically reflect symptoms of depressive disorder. The NDDI-E consists of six items, each of which is scored on a scale from 1 to 4, with the total score ranging from 6 to 24. We used the original English version ¹⁰ and a validated French version. ¹¹ The cut-off score of ≥ 16 suggested by the NDDI-E creators was used, as it demonstrated a sensitivity of 81% and specificity of 90%. ¹⁰ The NDDI-E is recommended by the International League Against Epilepsy (ILAE) for screening for depression in PWE. ³

The GAD-7 is a self-report questionnaire designed to screen for and assess the severity of GAD symptoms ¹² (Supplementary Tables 5 and 6). It includes seven items, each rated on a scale from 0 to 3, reflecting the frequency of symptoms experienced over the preceding two weeks. The total score ranges from 0 to 21, with higher scores indicating greater severity. In the clinic, the original English version ¹² and a validated French version of the tool were used, ^13,14 with a cut-off score of 7 to screen for GAD symptoms. At this threshold, the GAD-7 demonstrates a sensitivity of 95.9% and a specificity of 76%. ¹⁴

The BAI is a 21-item self-report questionnaire designed to evaluate the severity of anxiety symptoms experienced in the preceding week ¹⁵ (Supplementary Tables 3 and 4). Each item is scored on a 4-point scale ranging from 0 (“not at all”) to 3 (“severely”), resulting in a total score range of 0–63. The BAI evaluates a broad spectrum of anxiety symptoms, including physiological, cognitive and emotional dimensions. However, some studies suggest that the BAI may demonstrate greater sensitivity to panic-related symptoms compared to other forms of anxiety. ¹⁶ We used the original English version ¹⁵ and a validated French version of the instrument, ^17,18 with 22 as a cut-off score. At this threshold, the BAI demonstrates a sensitivity of 85.0% and a specificity of 88.1% for identifying anxiety symptoms. ¹⁹

Both the GAD-7 and BAI were used for anxiety screening because they assess complementary dimensions of anxiety, with the GAD-7 focusing on cognitive and emotional symptoms and the BAI on somatic and autonomic symptoms. This approach allows for a more comprehensive evaluation in patients with epilepsy.

Data analysis

All NDDI-E, BAI and GAD-7 questionnaires completed during the study period were reviewed by the investigators to identify patients who screened positive for depression and anxiety symptoms. A screening was considered positive if at least one of the NDDI-E, BAI or GAD-7 scores met the established cut-off. Medical records of patients with positive screening results were then reviewed to determine whether the neurologist had documented the score or recorded depression and anxiety symptoms in the medical note. Subsequently, we assessed the epileptologist’s approach to managing these symptoms (Figure 1).

Figure 1. Summary of data analysis method.

First, we verified whether patients were directed to mental health professionals (neuropsychiatrist, psychiatrist or psychologist) or other healthcare professionals (e.g., family physician or social worker). Then, we evaluated adjustments made to antiseizure medications, as well as any alternative management options suggested to alleviate depression and anxiety symptoms. Neurologists’ responses to positive screening results were categorized into groups (e.g., referral, antiseizure medication modification and other management options). When no management action was explicitly documented, cases were classified as “no management option offered.” In these instances, the full medical chart was reviewed to determine whether symptoms had been addressed in prior visits or whether the patient was already receiving care for depression and anxiety (e.g., ongoing follow-up with a psychologist or psychiatrist). These prior actions were not counted as outcomes of the screening itself but were noted to provide context for the absence of new interventions. We acknowledge that some management decisions may have been communicated orally but not documented.

Results

Screening questionnaires

Between January 2018 and March 2020, 2253 sets of NDDI-E, BAI and GAD-7 questionnaires were collected. Each questionnaire set was treated as an independent entry, and patients who were seen more than once during the study period were included. Due to inaccurate and/or incomplete responses from patients, 106 NDDI-E, 196 BAI and 21 GAD-7 scores were omitted from the dataset. We excluded a further 215 NDDI-E, 209 BAI and 231 GAD-7 scores from the analysis, as these were completed by patients without epilepsy. Ultimately, 1932 NDDI-E, 1848 BAI and 2001 GAD-7 scores were included in the analysis (Figure 2).

Figure 2. Flow chart of number of NDDI-E, BAI and GAD-7 scores. NDDI-E = Neurological Disorders Depression Inventory for Epilepsy; BAI = Beck Anxiety Inventory; GAD-7 = Generalized Anxiety Disorder-7.

Among the assessed questionnaires, 36% were positive on at least one screening tool: 13.6% screened positive for depression symptoms (NDDI-E ≥ 16), 12.3% for anxiety symptoms (BAI ≥ 22) and 30.3% for GAD symptoms (GAD-7 > 7) (Figure 3). Thirty-six percent of these patients were offered an intervention for their symptoms, 59% were not and 5% declined the epileptologist’s suggestion (Figure 4A).

Figure 3. Distribution of score results. A) NDDI-E. B) BAI. C) GAD-7. NDDI-E = Neurological Disorders Depression Inventory for Epilepsy; BAI = Beck Anxiety Inventory; GAD-7 = Generalized Anxiety Disorder-7.

Interventions

Seventy-four percent of scores were duly documented in the clinic note (and not only on the questionnaires themselves). In 41% of patients with a positive screening for depressive and/or anxiety symptoms, the chart review identified one or more actions taken by the treating epileptologist to address these concerns, as detailed below (Figure 4B):

1. A. Referral to a mental health professional: More than half (58%) were referred to another healthcare professional: 43% were directed to a neuropsychiatrist, 7% to a general psychiatrist, 21% to a psychologist, 15% to both a psychologist and neuropsychiatrist and 14% were advised to seek further assistance from their family physician (Figure 4C).

2. B. Medication adjustment: Twenty-nine percent of patients had their antiseizure medication regimen adjusted. For 18% of these patients, an antiseizure medication was stopped because it was believed to contribute to depression or anxiety symptoms (e.g. levetiracetam was stopped for 16 patients; see Table 1 for details on other medications). Additionally, 41% had their antiseizure medication dosage adjusted to manage neuropsychiatric symptoms and achieve seizure control (Table 1). In 41% of cases, a different antiseizure medication was introduced to address these dual concerns (Table 1). For only two patients did the neurologist newly prescribe an antidepressant medication. A third patient had their existing antidepressant prescription renewed.

3. C. Other management options: For 13% of patients, alternative options were provided to address their symptoms. The majority (57.5%) received reassurance, counseling on healthy lifestyle habits and follow-up appointments to monitor symptom progression. Nine patients were found to have medical conditions unrelated to epilepsy as the primary cause of their depression and anxiety. These individuals were either treated directly or referred to appropriate healthcare professionals for further management. Additionally, three patients were admitted to the epilepsy monitoring unit on the basis that inadequate control of seizures was primarily responsible for the depression symptoms. One patient was referred to a sleep clinic due to sleep-related disturbances, while another was directed to a psychiatry service specializing in substance-use disorders.

Figure 4. Management following a positive screening. (A) Response to a positive screening. (B) Type of management options among patients who received an intervention. (C) Referrals to healthcare professionals following a positive screening.

Table 1. Adjustments in antiseizure medication to achieve seizure control and manage depression and anxiety symptoms

Patients with no management option offered

Among patients who screened positive for depression and/or anxiety (Supplementary Figure 1 ), more than a half (59%) had no specific management option or intervention documented in the clinic note. Of this subgroup, 4% had previously been referred to a psychiatrist or psychologist. For 62% of these patients, scores were recorded in the clinic note but without documentation of a specific intervention plan.

Discussion

In this study, we examined our experience of routine screening for depression and anxiety symptoms in our epilepsy clinic, as recommended by the ILAE Commission on the Neuropsychiatric Aspects of Epilepsy. ² Our audit revealed that most screening scores (74%) were duly documented in the clinic note, indicating that epileptologists generally reviewed the answers provided by patients. Screening questionnaires identified an important number of patients with depression or anxiety symptoms that could benefit from psychiatric evaluation and possibly treatment. Overall, 36% screened positive on at least one questionnaire: 13.6% on the NDDI-E, 12.3% on the BAI and 30.3% on the GAD-7.

We compare our results with those of two recent studies conducted in epilepsy clinics within tertiary hospital centers. These studies similarly reviewed the charts of adults with a confirmed diagnosis of epilepsy who were followed by an epileptologist. The first was carried out in an adult epilepsy clinic in North Carolina, ²⁰ analyzing NDDI-E and GAD-7 scores in 422 patients in 2018–19. This study employed the same cut-off score for depression screening (NDDI-E ≥ 16), but a different cut-off for anxiety (GAD-7 ≥ 10). Their results showed slightly higher rates of positive screening compared to ours: 20% for NDDI-E ≥ 16 and 25% for GAD-7 ≥ 10, versus 13.6% and 21.5% in our cohort. The second study was conducted in a tertiary teaching hospital in Kuala Lumpur in 2022, and included 585 patients, ²¹ using lower cut-off scores for depression (NDDI-E ≥ 15 vs. ≥ 16) and anxiety (GAD-7 ≥ 10). Their reported depression screening rate was slightly lower than ours, with 15.5% of patients screening positive for depression (NDDI-E ≥ 15) compared to 16.3% in our study. Similarly, they observed a lower rate of positive anxiety screening (GAD-7 ≥ 10), with 17% of patients screening positive, compared to 21.5% in our cohort.

Variations in findings across studies may be attributed to differences in socio-demographic and clinical characteristics of the study populations. For example, Munger et al. found that the prevalence of positive screening for anxiety and depression was significantly higher among patients with persistent seizures, while no notable differences were observed across sex or race/ethnicity. In contrast, Lim et al. reported that multiple factors contributed to higher symptom scores, including younger age, female sex, Indian ethnicity, student status and the use of specific antiseizure medications such as levetiracetam and lamotrigine. Furthermore, Lim et al. highlighted that seizure remission and longer epilepsy duration were associated with lower depression and anxiety scores, suggesting that improved seizure control and chronic disease adaptation may play a protective role. Unlike these studies, our QI project was not designed to evaluate predictors of positive screening results.

Concerning BAI screening, we were not able to find any studies with similar inclusion and exclusion criteria to compare score results. However, it is interesting to note the difference in positive screening rates for anxiety screening questionnaires observed in our own study, with 30.3% for the GAD-7 compared to only 12.3% for the BAI. This may be partially attributed to a higher rate of incomplete or inaccurately completed BAI questionnaires, with 196 BAI forms excluded compared to only 21 GAD-7 forms. This difference could be explained by the relative length of the two tools; the BAI consists of 21 items and typically takes 5–10 min to complete, whereas the GAD-7 contains only seven items and requires only 1–2 min. ^12,15 Additionally, in the sequence of questionnaires administered to patients, the BAI was positioned last. This placement may have contributed to incomplete responses, as patients may have experienced fatigue or time constraints by the time they reached the BAI. Beyond these procedural issues, the distinct focus and characteristics of the two tools could further explain the variation in results. The GAD-7 primarily assesses cognitive and worry-related symptoms, such as excessive worry and restlessness, which may be more prevalent in patients with epilepsy due to the chronic nature of their condition, including fears related to seizures, social stigma and treatment outcomes. ^5,8,22 In contrast, the BAI is more sensitive to panic attack symptoms and emphasizes somatic symptoms, such as trembling, dizziness and numbness, which may overlap with physical manifestations of epilepsy or side effects of antiseizure medications. ^19,23 This overlap may lead to underreporting of anxiety symptoms on the BAI, as patients or clinicians might attribute these physical symptoms to epilepsy rather than anxiety. These differences underscore the importance of selecting appropriate tools for specific populations and suggest that combining the GAD-7 and BAI may provide a more comprehensive understanding of anxiety in patients with epilepsy.

More recently, the Epilepsy Anxiety Screening Interview (EASI) and its brief version (brEASI) have been developed specifically for PWE, analogous to the NDDI-E for depression. ²⁴ Unlike the GAD-7 and BAI, these instruments are designed to detect a broader spectrum of anxiety disorders (not only GAD and panic disorder) and to capture anxiety related to seizure control. ²⁴ They have been translated and validated in multiple languages (e.g., French, Russian, Chinese). In comparative studies, they have outperformed the GAD-7 in identifying anxiety disorders in PWE. ^25–28 Although these scales were not available during our study period, they represent promising alternatives for future clinical practice and research.

Our audit also revealed that positive screening results led to a change in management in less than half of the patients. The most frequent management option chosen was referral to mental health professionals (neuropsychiatrists, psychiatrists and psychologists). In our institution, our epilepsy group has access to a small team of mental health professionals, who shared their time with other clinical services. It is noteworthy that only two patients were prescribed an antidepressant medication by their epileptologist, suggesting that epileptologists are uncomfortable initiating such treatments. This stance is at odds with the International Consensus Clinical Practice Statements for the Treatment of Neuropsychiatric Conditions Associated with Epilepsy, ³ which provide detailed pharmacologic recommendations for treatment of depression by neurologists. The 2021 ILAE clinical practice recommendations ²⁹ further emphasize psychological interventions for mild depression and SSRIs as first-line pharmacologic agents, with venlafaxine as a reasonable alternative in cases of inadequate response. Although our study predates these guidelines (2018–2020), our findings highlight the very challenges in practice that such recommendations aim to address.

Not surprisingly, however, epileptologists recognized that some antiseizure medications could be responsible for or exacerbate depression or anxiety symptoms (e.g., levetiracetam) or conversely may be beneficial (e.g., lamotrigine, carbamazepine, valproic acid, benzodiazepines), as several medication changes were documented. While discontinuations clearly reflected a psychiatric rationale (i.e., medications believed to contribute to symptoms either directly or indirectly), introductions and dosage modifications of “beneficial” medications were more difficult to interpret, as they could have been undertaken either to improve seizure control, alleviate psychiatric symptoms or both.

Our findings are consistent with those of the ILAE Psychology Task Force survey of epilepsy health professionals, ³⁰ which revealed similar patterns of practice and barriers in the management of depression and anxiety in PWE. In both settings, there was recognition of the importance of addressing psychiatric comorbidities; however, fewer than half of providers felt adequately resourced to manage these conditions, echoing the gaps in care and reluctance to prescribe psychiatric medications observed in our cohort. Together, these results underscore the need for updated protocols, greater integration of mental health professionals in epilepsy care and enhanced training for epileptologists in psychiatric management.

Our study has limitations inherent to all retrospective chart studies. Because our analysis was limited to information available in the medical chart, neurologists may have verbally discussed score results and treatment options with certain patients without formal documentation. This approach may have underestimated the frequency of clinical management following positive screening. Some questionnaires were incompletely filled, possibly from lack of time or instructions. As data prior to 2018 were not retrievable, we were unable to compare referral and management rates before and after the implementation of routine screening. We did not assess whether the actions taken by epileptologists were appropriately carried out – for example, whether neuropsychiatry appointments were scheduled promptly or if patients attended these appointments. Furthermore, we did not evaluate whether these actions ultimately provided benefit to patients. As noted by the AAN 2017 QI work group, ³¹ the development of outcome measures is challenging, as psychiatric treatment is often managed outside of neurology (e.g., by psychiatrists, psychologists, primary care physicians or social workers). Future studies with continuous data collection could address these limitations, for instance, by tracking appointment attendance, measuring time from referral to consultation and evaluating clinical outcomes following intervention using standardized methods and an interdisciplinary approach.

Conclusion and recommendations

The implementation of screening questionnaires enabled the detection and management of depression and anxiety in a substantial proportion of epilepsy patients. However, our audit also identified areas for improvement that could enhance patient care and well-being. Although epileptologists obtained additional information about their patients’ neuropsychiatric symptoms, the questionnaires were not accompanied by further guidance, more clinic time or an increase in mental-health resources to assist them in managing patients with positive screening tests.

Certainly, epileptologists should be expected at minimum to systematically document the results of such screening tests in the patient chart and to have a discussion with patients when a screening is positive. The content of the discussion, as well as any potential interventions reviewed, should likewise be noted in the chart.

Epileptologists were aware of the availability of a neuropsychiatrist or psychologist for referral; however, they may also have been mindful that these are limited resources and that it was not practicable to refer every patient with a positive screening test to a mental-health professional for further evaluation. This highlights the fact that the implementation of screening questionnaires, without other interventions to address the findings in the questionnaires, is unlikely to lead to comprehensive management of neuropsychiatric symptoms in the patients of an epilepsy clinic.

The extreme rarity with which epileptologists prescribed psychiatric medications suggests a discomfort with the medical management of such symptoms. Our QI study therefore suggests that epileptologists may benefit from additional training or self-directed learning in the assessment and management of depression and anxiety.

It is notable that our epileptologists each completed their fellowships at different institutions, perhaps highlighting a broader gap in epilepsy education and underscoring the need to further integrate psychiatric training into epilepsy fellowship programs.

In conclusion, while routine screening increases the detection of depression and anxiety among epilepsy patients, the full benefit to patients will only be realized when screening is followed by timely, effective interventions and access to appropriate mental-health professionals. Epilepsy clinics must not only record positive screens but also assist epileptologists in initiating or coordinating treatment, whether pharmacologic or psychosocial.