30-day Readmission Rates in Pediatric Patients with Functional Seizures

Abstract

Purpose:

To ascertain the rates of 30-day readmissions and emergency department presentations among pediatric patients with an index admission for functional seizures.

Method:

A retrospective chart review of pediatric patients with an index discharge from the pediatric epilepsy monitoring unit (EMU) or general neurology service for functional seizures. Data collected included demographics, comorbidities, risk factors, and treatment during the index admission.

Results:

A total of 112 patients were included, of which nearly one-in-five had a 30-day readmission or emergency department presentation. Index admission to the general neurology service was independently associated with more re-presentations to the hospital (t=3.26, p<0.0015). The univariate analysis indicated that cognitive impairment and autism were associated with a lower likelihood of readmission while a neurology referral and being started on an anti-seizure medication were associated with a greater likelihood of readmission.

Conclusion:

A substantial proportion of pediatric patients with FS return to the hospital within 30 days of discharge. Our data suggest that patients admitted to the EMU service have a lower likelihood of readmission. We speculate that this may be due to differences in patient clinical characteristics as well comprehensiveness of the diagnostic evaluation and management in the EMU compared to the general neurology service.

1. Introduction

Functional seizures (FS) are paroxysmal events of behavioral change and/or altered consciousness that resemble epileptic seizures, but are not associated with epileptiform activity. Several models have been proposed which suggest FS are caused by a complex interaction of biological, psychological, and social factors [1]. It is a common diagnosis in the Epilepsy Monitoring Unit (EMU) but also frequently encountered in other clinical settings [2].

The prevalence of FS in the pediatric population has not been adequately studied, however it has been estimated that approximately 1-9% of children with suspected epilepsy have FS.[3] FS in pediatric patients are associated with substantial inpatient and emergency department (ED) medical expenditures. For instance, a previous study in 2012 reported the total inpatient medical expenditures for pediatric functional neurological disorders (FND) in the United States was $75 million, with a mean charge of $28,000 per patient [4]. The same study showed costs increased significantly from $26 million in 2003.

In 2012, a national program was created that penalized hospitals for excessive readmissions as they are a significant contributor to healthcare expenditures and are also associated with negative outcomes, thus efforts to reduce avoidable readmissions has become an area of focus in most areas of medicine, including neurology. [5] Identification of risk factors that are associated with increased risk for readmission is an integral component of these efforts. To our knowledge, there have been no previous attempts to evaluate readmissions in pediatric patients with FS. A study in adults found that FS-related readmissions or ED visits were seen in approximately 6% of FS patients within 30 days of discharge and 23% of FS patients beyond 30 days of discharge [6]. Therefore, we investigated 30-day readmissions among pediatric patients with FS to better understand the incidence and potential factors associated with readmission rates.

2. Methods

2.1. Sample and Data Description

This study was performed at a large regional tertiary hospital with a wide catchment area including middle Tennessee, north Alabama, and south Kentucky. A retrospective chart review was performed on pediatric patients who had an index admission for FS. Patient charts were identified using a keyword search in a research database that is maintained by the institution. The query was performed during December of 2021 on all available data in the system. Charts were then manually reviewed by a board-certified physician (JF). Inclusion criteria consisted of probable or video-EEG confirmed FS and an index admission for FS. The criteria for probable FS required (1) an absence of epileptiform activity on a routine or sleep-deprived EEG and (2) an experienced clinician witnessing a typical event in person or on video that had a semiology consistent with FS [7]. Twelve patients with comorbid epileptic seizures (ES) were included, given a striking difference in semiology between the epileptic and non-epileptic events, accurately distinguished by family members and providers. Two patients were excluded because records indicated that there were some difficulties distinguishing ES from FS events. If a patient had multiple readmissions after the index discharge, only the first was included for the purpose of this analysis. This research study was reviewed and approved by the Vanderbilt University Medical Center Institutional Review Board.

2.2. Chart Review

The following variables were collected on all participants: age, gender, psychiatric diagnoses, neurological diagnoses, date of index admission, date of readmission (if applicable), presence of psychiatric or anti-seizure medications at time of index admission, EEG results, brain MRI results (if available), treatment at time of index admission, and presence of readmission within 30 days or 90 days.

In Tennessee there is a significant shortage of mental health providers, especially for the pediatric population, thus creating a barrier for patients referrals [8]. At time of discharge, if a mental health referral is appropriate, patients are provided information or resources to help find a local counselor, therapist, psychologist, or psychiatrist. For the purposes of this study, we refer to this treatment option as “psychiatric resources.” In addition, we evaluated for the presence of potential risk factors for FS including psychosocial stressors and other comorbid conditions which have been previously associated with FS, such as migraine and sleep disturbances (e.g., insomnia) [9]. The service of the index admission was documented as either the pediatric EMU or general neurology. Presentations to the ED were counted as readmissions for the purpose of this analysis, as they account for a large proportion of post-hospitalization acute care encounters and patient costs [10].

2.3. Statistical analysis

Statistical analyses were performed using GraphPad Prism 9 (GraphPad Software, San Diego, California). Descriptive statistics were analyzed followed by direct statistical comparisons between each group. Analysis of nominal variables was performed employing the chi-square test, and analysis of linearly scaled variables was done with an unpaired t-test. Multivariable logistic regression analysis was performed to assess for factors independently associated with risk of EMU readmission. Potential factors regarding the work-up and results of the index admission were treated as variables, including type of initial admission to EMU or general neurology service, prescription of psychiatric medications, referral to psychiatry or psychiatric resources, referral to neurology, and prescription of anti-seizure medications. Additional multivariate analysis was performed evaluating the potential influence of risk factors (Table 1) on readmission risk. Results with p<0.05 were considered significant and p<0.01 highly significant.

Table 1.

Pre-admission characteristics among pediatric patients with functional seizures

	Patients with 30-day readmission (N=22) (%)	Patients without 30-day readmission (N=90) (%)
Female gender	77.3	74.4
Any psychiatric history	86.6	86.6
Depression	40.9	52.2
Anxiety	81.8	77.8
Bipolar	4.5	1.1
Post-traumatic stress disorder	31.8	22.2
Attention deficit hyperactivity disorder	18.2	35.6
Obsessive compulsive disorder	9.1	7.8
Autism	0*	15.6*
History of abuse	27.2	24.4
Physical abuse	18.1	12.2
Sexual abuse	22.7	16.7
Death of family member or close friend	9.1	23.3
Recent life altering medical diagnosis	0	7.8
Traumatic brain injury	4.5	4.4
Cognitive impairment	0*	17.8*
Sleep disturbance	54.5	48.9
Chronic pain	22.7	22.2
Migraine	18.2	31.1
Asthma	22.7	28.9
Bullying	36.4	33.3
Family instability (e.g., divorce, prison)	50	44.4
Reported family history of seizure	18.2	25.6
Taking anti-seizure medication	13.6	27.8
Taking psychiatric medication	36.4	57.8

^*Bold indicates a statistically significant difference with a p < 0.05

3. Results

3.1. Pre-admission baseline clinical characteristics

A total of 112 patients were identified and included, 22 of which had a 30-day readmission. Approximately three-quarters of the patients were female and 86.6% had a prior psychiatric diagnosis (Table 1). A prior history of abuse (physical or sexual) or sleep disturbances were seen in approximately one-quarter and one-half of patients, respectively. Bullying was reported by nearly one-third of patients and other stressors such as familial instability due to divorce or imprisonment were also common. A family history of seizures was reported in approximately one-fifth of those who were readmitted and one-quarter of those who were not readmitted. The univariate analysis indicated that autism (p=0.048) and cognitive impairment (p=0.038) were more common in those who were not readmitted within 30 days.

3.2. Index admission characteristics and reasons for readmission

Of the patients who were readmitted within 30 days of discharge, 90.9% had an index admission to the general neurology service, compared to 54.4% of patients who were not readmitted (p=0.001). EEG and brain imaging results were normal in the vast majority of patients except as indicated in Table 2. A neurology referral (p=0.02) or being treated with an anti-seizure medication (p=0.02) was more common among those with a 30-day readmission as demonstrated in the univariate analysis. The most common cause for readmission within 30 days was recurrent spells (86.4%) (Table 3). Among those who had a readmission between 30-90 days of discharge, there was a greater diversity of reasons for readmission. For instance, 46.7% of patients presented for psychiatric issues which were not directly related to their FS, 20% presented due to unrelated medical concerns and another 20% presented with a new FND presentation (e.g., functional movement disorder).

Table 2.

Index admission characteristics among pediatric patients with functional seizures

	Patients with 30-day readmission (N=22) (%)	Patients without 30-day readmission (N=90) (%)
Admitted to EMU	9.1**	45.6**
Admitted to floor or emergency department	90.9**	54.4**
Readmitted between 30-90 days of index admission	4.5	16.7
EEG results
Normal	86.4	85.6
Focal epileptiform discharges	9.1	4.4
Generalized epileptiform discharges	0	6.7
Focal slow activity	0	2.2
Generalized slow activity	4.5	6.7
Brain imaging resultsa
Normal (N)	16/16	41/45
Migration abnormalities (N)	0	1/45
Encephalomalacia (N)	0	0
Stroke (N)	0	2/45
Treatment at index admission
None	22.7	18.9
Psychiatric medications	13.6	11.1
Psychiatric referral or follow-up	40.9	51.1
Psychiatric resources	18.2	25.6
Neurology referral	27.2*	7.8*
Anti-seizure medication	9.1*	1.1*
Transfer to psychiatric hospital	0	4.4

^aBrain imaging was not exclusively obtained during the index admission

^*Bold indicates a statistically significant difference with a p < 0.05

^**Indicates a highly statistically significant difference with a p < 0.01

Table 3.

Reasons for readmission among pediatric patients with functional seizures

	Patients with 30-day readmission* (N=22) (%)	Patients with 90-day readmission* (N=15) (%)
Recurrent spells	86.4	40
Psychiatric issue	9.1	46.7
Medical issue	9.1	20
Other functional presentation	4.5	20

^*Reasons for readmission were not mutually exclusive.

There was a total of 10 patients with more than one readmission within 90 days of the index admission. Six of the patients had one additional readmission, three had two additional readmissions and one had three additional readmissions. Overall, 11 and 4 of the additional readmissions occurred within 30 days and between 30-90 days of the index admission, respectively.

3.3. Multivariate logistic regression

The multivariate logistic regression model showed that patients with admission to the EMU service were less likely to have a 30-day readmission (t=3.26, p<0.0015). No other management decisions during the index admission appeared to correlate with readmission rates. Furthermore, with exception to admitting service, no other variable was independently associated with readmission within 30 days, including psychiatric diagnoses, psychosocial stressors, or other risk factors.

4. Discussion

The results of our study suggest that nearly one in five pediatric patients with an index admission for FS are readmitted or represent to the ED within 30 days. The vast majority of the patients who were readmitted had an index admission to the general neurology service compared to the EMU. We have several hypotheses that may account for this finding. First, patients who are initially admitted to the general neurology service may have different clinical characteristics compared to those who are admitted to the EMU; they may have more frequent events or have not yet established with a neurologist or epileptologist suggesting they are more likely to return to the hospital for management. Second, patients on the general neurology service may be less likely to be seen by an attending physician with extensive experience diagnosing, counseling, and managing FS compared to the EMU, which is staffed exclusively by pediatric epileptologists. Therefore, when an additional event occurs, patient families may experience uncertainty about how to manage events in the short-term at home. According to the univariate analysis, patients with readmissions within 30 days were more likely to be started on anti-seizure medications or referred to neurology, supporting the notion that a definitive diagnosis of FS was not always obtained in this cohort. In addition, a larger number of patients without readmissions were either referred to psychiatry or given psychiatric resources, suggesting these patients were more likely given a definitive FS diagnosis and an appropriate outpatient management plan. These findings are concordant with prior studies showing that both adults and pediatric patients with FS who receive proper education about their diagnosis had a subsequent reduction in healthcare utilization [11, 12].

We found that our cohort had similar clinical characteristics as previously reported for pediatric patients with FS. In contrast to adult cohorts where physical and sexual trauma are the most common risk factors, we found that stressors related to the family as well as bullying were more common [13-15]. A few previous studies have shown a variable frequency of bullying among pediatric patients with FS that ranged from 5-38% [16, 17]. Our results indicated that bullying was present in approximately one-third of patients and was second in frequency after those relating to familial stressors. Nearly one-quarter of patients reported a family history of seizure. This finding has been previously reported and has been speculated as related to either biological factors or a form of behavioral modeling [18, 19].

Psychiatric disorders were found to be highly prevalent in our cohort. Nearly nine out of ten patients had a psychiatric diagnosis, with depression and anxiety being the most common. Prior studies have found a variable range in the prevalence of comorbid psychiatric disorders in pediatric FS from 39.8 to 100% [9, 14, 20, 21]. Some of the variability between these studies, including ours, may relate to differences with regards to which diagnoses were included, the sample population and other methodological differences. The high prevalence of psychiatric diagnoses in our cohort contributed to the overall readmission frequency as it was a common cause for readmission within 90 days.

The univariate analysis indicated that autism and cognitive impairment were associated with a reduced risk of readmission. Though the multivariate analysis did not confirm this association, we speculate that there could be several factors which may account for this finding. Prior studies suggested that pediatric FS patients with comorbid autism or cognitive impairment may have different demographic characteristics and comorbidities.[22-24] It has also been speculated that FS in patients with autism or cognitive impairment have a partially different pathophysiology.[22, 25] It is possible that some of these differences contributed to our finding. In addition, it may be that patients in our cohort with comorbid autism or cognitive impairment had already established care with a provider who could give guidance for recurrent events.

The reason for readmission within 30 days were predominantly due to recurrent spells. However, the reasons for readmission within 30-90 days were more diverse. It is possible that with time, patients are more likely to receive proper diagnosis and treatment, thus lowering their chance of representing to the hospital with FS, relative to other reasons.

There are several limitations to this study. For instance, our sample size was relatively small, which may have resulted in insufficient power. If our sample size was larger, additional factors associated with likelihood of readmission may have been detected. In addition, this study was performed at a single center, limiting the ability to generalize our findings. On the other hand, to our knowledge this is the first study to evaluate readmission rates for FS in pediatric patients and may serve as a relative baseline for potential future studies. Our finding that a substantial proportion of FS patients discharged from the general neurology service represent to the hospital within 30 days suggests that improvements could be made as to how they are managed. It is critical that all efforts are made to ensure efficient and accurate diagnosis, appropriate counseling about diagnosis, and referrals to outpatient mental health providers are made prior to discharge. Future, larger studies could evaluate the impact of a standardized diagnosis and treatment protocol for patients who are admitted with concern for FS to assess how it may impact outcomes as well as resource utilization.

5. Conclusion

Nearly one-fifth of pediatric patients who are admitted with FS are readmitted within 30 days, of which the majority had an index admission to the general neurology service compared to the EMU. We hypothesize this disparity is due to differences in clinical characteristics of the patients, comprehensiveness of diagnostic evaluation and management, and quality and quantity of patient education regarding their diagnosis. Future studies could evaluate the impact of a standardized diagnostic and management protocol for patients admitted to the general neurology service with concern for FS.

• Nearly one-fifth of functional seizure patients are readmitted within 30-days.

• Patients admitted to the EMU were less likely to be readmitted.

• Readmissions after 30-days were more likely to be for other problems.