• Epilepsy syndrome;
  • Pediatric epilepsy;
  • Syndrome severity


  1. Top of page
  2. Abstract
  6. Acknowledgments

Summary: Purpose: A severity rating for each of the pediatric epilepsy syndromes can contribute to a more comprehensive understanding of seizure condition severity. We describe the development of the Epilepsy Syndrome Severity Scores–Child (ESSS–C).

Methods: The Delphi Technique was used to establish severity scores for pediatric epileptic syndromes as defined by the International League Against Epilepsy (ILAE). Pediatric neurologists with expertise in epilepsy were asked to rate the severity of each syndrome on a scale of 1 to 10, considering: (a) response to medical treatment, (b) seizure severity, and (c) long-term prognosis. Syndrome severity ratings took place in four different rounds. Experts provided initial scores in rounds 1 and 2.

Results: Of the 18 experts who agreed to participate in the development of the scale, 12 completed all four rounds. The median served as the syndrome severity score if the mean and median agreed within 0.5. Otherwise, whichever of these two numbers was closest to the mode was selected. Syndromes that were rated with high severity scores (9 or 10) or low severity scores (1 or 2) had unanimous or near unanimous ratings. The syndromes with scores in the middle range (4 to 8) had a wider range of ratings.

Conclusions: If further studies provide empirical support for the validity of these epileptic syndrome severity scores, then the ESSS-C has potential for use both clinically and in future research in the prediction of outcomes of treatments.

In children with epilepsy, the severity of the seizure condition is important for documenting outcomes of various forms of therapy and for predicting psychosocial problems. A number of factors contribute to the broad concept of condition severity, including seizure severity, epileptic syndrome severity, the number and side effects of antiepileptic drugs (AEDs), and the impact on the children and families. Although condition severity is important both clinically and in research, its measurement has been problematic. Many different methods have been used because no agreed-upon way exists to measure condition severity. A major limitation is that most methods are limited in scope and fail to reflect the true multidimensional nature of condition severity. For example, some authors have selected a single variable such as active or inactive epilepsy or seizure frequency to reflect condition severity (1,2). Other authors have used a combination of variables, including seizure type, age at onset and duration of seizures, seizure frequency, and treatment with medications (3–8).

Although the International League Against Epilepsy (ILAE) has separate classifications for seizure type and epileptic syndrome (9,10), the current measures of severity of the seizure condition address only the severity of the seizures (often represented, at least in part, by seizure type) but not the syndrome. These measures of seizure severity can be divided into two main groups. One group starts with the clinician's classification of seizure type plus seizure frequency. For example, the Veterans Administration (VA) seizure frequency and severity scale, developed initially for clinical AED trials at the VA, starts with three seizure types as a major component of the measure and then determines seizure frequency for each seizure type as well as patterns of seizures, auras, clusters of seizures, and impaired function (11). Hermann et al. (4) categorized seizure types as mild, moderate, or severe and added an assessment of seizure control for each general seizure type. Austin et al. (12) used a similar classification, with the addition of scores for number of AEDs and side effects of AEDs. The second group of seizure severity measures uses parent or patient descriptions of how often and how much the child's life is disrupted by seizures. As examples, both the Liverpool (13) and Hague (14) scales emphasize impairment during and after seizures.

It was the purpose of this study to address this gap in measurement of condition severity by developing a method to rate the severity of pediatric epileptic syndromes. We anticipate that this new score developed by systematic appraisals of the severity of each epileptic syndrome by experts would be combined with other variables such as seizure frequency within seizure type to obtain a comprehensive measure of seizure condition severity.


  1. Top of page
  2. Abstract
  6. Acknowledgments

The Delphi technique served as the model for development and testing of the Epileptic Syndrome Severity Scores–Child (ESSS-C). This method, which has been used in a number of medical studies (15–17), was developed by the Rand Corporation to use the expertise of specific panels by issuing a series of questionnaires and giving feedback on a specific topic (18). The strength of this method lies in the anonymous reporting of individual experts on a selected topic, thus avoiding the problems associated with group dynamics (19). By using the Delphi method, we were able to pool the opinion of a number of experts in pediatric epilepsy. One weakness of the Delphi method is the difficulty in obtaining consensus (19). Definitions of consensus can mean “true consensus” to “majority rules,” and the definition of consensus must be clear when using the Delphi method (19).

To reach a consensus on the severity of epileptic syndromes as identified by the ILAE (10), we asked 25 child neurologists to participate in the development of the scale. All raters had a major interest in epilepsy. Seven of the 25 raters were academic neurologists focusing exclusively on epilepsy; two were private practitioners with medical school faculty appointments; and 16 were academic general child neurologists. Seven women and 19 men were included. The raters had an average of 24 years of clinical experience when the ratings were completed. The process consisted of four rounds to establish a severity score for each of 36 epilepsy syndromes. Three rounds allow appropriate feedback without diluting answers and are recommended to obtain consensus (18). Four rounds were used in this study because the first round did not include all of the syndromes. A second round was identical to the first, except that it included all syndromes. Each round took ∼3 months including distribution, return of scores, and analysis.

Round 1

The purpose of Round 1 was to establish an initial severity score for the most commonly diagnosed pediatric epilepsy syndromes. A list of the 26 most frequently seen pediatric epilepsy syndromes, as defined by the ILAE, was sent to the panel of expert child neurologists with instructions to rate each syndrome, considering: (a) response to medical treatment, (b) seizure severity, and (c) long-term prognosis. Each syndrome was rated on a 10-point scale, with 1 being least severe and 10 being most severe. The rating scale was sent to 25 experts, and 18 responded. The responders were similar to the original group. They had an average of 22 years of clinical experience, and included five women and 12 men, plus one anonymous responder. After completing the ratings, some experts provided feedback that the list of syndromes was incomplete and should include all classified syndromes. In addition, some experts reported that certain syndromes were too variable to be easily rated.

Round 2

In response to the feedback from Round 1, the research team expanded the list of syndromes to include all 36 of the ILAE-identified epilepsy syndromes. With these additions, the validity of the scale should increase, because the rating of each syndrome would be in the context of all syndromes. Moreover, the final scores should have broader applicability.

The purpose of Round 2 was to establish initial severity scores for the newly included syndromes. The expanded list was returned to the 18 experts with their original severity ratings. They were asked to rate the added syndromes by using the Round 1 severity rating rules. A total of 16 experts completed the second round. Using descriptive statistics (mean, median, and mode) to analyze their ratings of the syndromes, the authors assigned a severity score to each syndrome. The following procedures were used. The median served as the severity score if the mean and median agreed within 0.5. The score was rounded to the nearest whole number. For example, if the mean was 5.75 and the median was 6.00, then the severity score assigned was 6. If the mean and median were disparate by >0.5, whichever number was closer to the mode was used (rounded to the nearest whole number). For example, if the mean was 5.40, the median was 6.00, and the mode was also 6.00, a score of 6 was assigned to that syndrome. The median supported by the mean was the basis for a severity score in 25 syndromes. Of those 25 syndromes, only one had a unanimous severity score. The experts unanimously agreed that “early infantile epileptic encephalopathy with suppression bursts” should receive a score of 10. In 11 syndromes, the mode also was used to arrive at the severity score.

Round 3

The syndrome rating scale was sent to the 16 experts, with the initial severity score identified for each syndrome. The purpose of Round 3 was to have the experts consider the syndrome severity score and either agree or disagree with the score. If they disagreed with the severity score, they were asked to give a rationale for their disagreement and provide an alternative score. Of the original 16 experts, 13 responded. A unanimous severity score was reached for 11 syndromes. For the other 25 syndromes, the mean and median were all within 0.5, and the median served as the basis for the severity score.

The main reasons for disagreements with the severity scores were the rarity of certain syndromes and the variability of response to medical treatment, seizure severity, and long-term prognosis in other seizure syndromes. For example, four experts stated that “Primary Reading Epilepsy” was too rare to rate. In addition, two experts reported that the syndromes “seizures characterized by specific modes of precipitation” and “other epilepsy syndromes based on etiology or location” were too variable to rate. Finally, one expert stated that the severity of each syndrome was dependent on etiology and location. Because of the experts' comments, the research team came to the conclusion that a unanimous severity score could not be achieved for some syndromes.

Round 4

After evaluating the severity scores from Round 3, we initiated a fourth round to address two concerns. In the preceding rounds, the syndromes had been presented in four categories: (a) localization-related epilepsies and syndromes, (b) generalized epilepsies and syndromes, (c) epilepsies without unequivocal generalized or focal features, and (d) special syndromes. We were concerned that the experts might have arrived at severity scores by comparing the syndromes within these categories and not across all categories. The second concern was the need to be sure that equal scores meant equal syndrome severity. For example, eight syndromes had a score of 5, and we wanted to make sure that the experts believed that they were all equivalent in severity.

To address these issues, we asked the experts to reconsider each severity score to ensure that the score reflected their perceptions of the severity of each syndrome within the context of all of the syndromes. We also asked them to rate no more than four and no fewer than three syndromes with the same number. For this round, the syndromes were arranged in order of severity, starting with the highest score.

The syndrome severity scale was sent to the same 13 experts who responded in Round 3. In this round, 12 experts responded. There were no significant differences in the average ratings given at Round 1 by the 12 experts who completed all rounds compared with those who did not participate in each round. One expert was unable to rate six syndromes, stating that the outcomes were too variable. Two other experts did not rate one syndrome each, so in total, eight syndromes were rated by 11 rather than 12 experts.


  1. Top of page
  2. Abstract
  6. Acknowledgments

In this final round, three unanimous severity scores were found. For two syndromes, the mode was used to help establish a score. In one syndrome, the mean and median fell equally between two scores, and the distribution was bimodal. Specifically, West syndrome had a mean and median of 9.5 and modes of 9 and 10. For this syndrome, the median was selected as the severity score. An evaluation of all syndromes revealed a pattern. Syndromes that were rated as having high severity (scores of 9 or 10) or low severity (scores of 1 or 2) had unanimous or near unanimous ratings. The syndrome scores in the middle range (4 to 8) had a wider range of ratings. For example, neonatal seizures had a range of scores from 4 to 8. Table 1 shows the range of scores for each syndrome [with mean and standard deviation (SD)], with the syndromes ordered according to decreasing mean score. Table 2 lists the final syndrome severity scores by category.

Table 1. Distribution of Round 4 Scoresa,bThumbnail image of
Table 2. Final syndrome severity scores in the context of all syndromes
  1. aUnanimous severity score agreement.

1. Localization related 
 1.1 Idiopathic epilepsy (with age-related onset) 
       Benign childhood epilepsy with centrotemporal spikes2
       Childhood epilepsy with occipital paroxysms3
       Primary reading epilepsy4
 1.2 Symptomatic 
       Chronic progressive epilepsia partialis continua of childhood9
       Seizures characterized by seizures with specific modes of precipitation5
       Other epilepsy syndromes based on etiology or location 
       By virtue of etiology7
       By virtue of localization (otherwise normal child)5
 1.3 Cryptogenic (localization related)7
2. Generalized epilepsies and syndromes 
 2.1 Idiopathic (with age-related onset) 
       Benign neonatal familial convulsions2
       Benign neonatal convulsions2
       Benign myoclonic epilepsy in infancy2
       Childhood absence epilepsy (pyknolepsy)3
       Juvenile absence epilepsy5
       Juvenile myoclonic epilepsy5
       Epilepsy with generalized tonic–clonic seizures on awakening5
       Other generalized idiopathic epilepsies not defined above5
       Epilepsy with seizures characterized by specific modes of precipitation (e.g., photosensitive epilepsy)6
       Random grand mal4
 2.2 Cryptogenic and/or symptomatic 
       West syndrome (infantile spasms)9.5
       Lennox–Gastaut syndrome10
       Epilepsy with myoclonic–astatic seizures9
       Epilepsy with myoclonic absences8
 2.3 Symptomatic 
       2.3.1 Nonspecific etiology 
           Early myoclonic encephalopathy10a
           Early infantile epileptic encephalopathy with suppression burst10a
           Other symptomatic generalized epilepsies not defined above8
3. Epilepsy and syndromes undetermined whether focal or generalized. 
 3.1 With both generalized and focal seizures 
       Neonatal seizures6
       Severe myoclonic epilepsy in infancy10
       Epilepsy with csws in slow-wave sleep8
       Acquired epileptic aphasia (Landau-Kleffner syndrome)8
       Other undetermined epilepsy syndromes not defined above7
 3.2 Without unequivocal generalized or focal features6
4. Special syndromes 
 4.1 Situation-related seizures 
       Febrile convulsions—simple1a
       Febrile convulsions—complex3
       Seizures occurring only in the context of acute metabolic or toxic events1
 4.2 Isolated events 
       Isolated seizure (or cluster of seizures in a single day as the first unprovoked seizures, not status)2
       Isolated status epilepticus4


  1. Top of page
  2. Abstract
  6. Acknowledgments

This study, by using the Delphi technique, is a first attempt to obtain ratings of pediatric epileptic syndrome severity. It has the advantage of providing a score for each epileptic syndrome listed by the ILAE. For the majority of syndromes, the experts were quite consistent in assignment of a severity score. However, in the middle range of severity, there was somewhat more variability, although within a fairly narrow range.

A number of reasons can be postulated for the variability of some scores, although this study did not fully explore the rationale for an expert's score. These include the variability of the syndrome, the expert's familiarity with it, and the different weights assigned to different outcomes. Those syndromes assigned either the highest (worse) or lowest (best) scores had a very uniform unfavorable or favorable outcome. This included even rare syndromes. The variability in scores reflected more the variability of the syndromes such as reflex epilepsy or cryptogenic partial seizures than their frequency. An additional reason for variability in the scores of even common syndromes may be the raters' being more strongly influenced by either the prognosis for complete seizure control or the likely cognitive outcome for children with each seizure syndrome. For example, childhood absence has a favorable prognosis for remission but is associated with a high frequency of behavioral and academic problems (20). Juvenile myoclonic epilepsy has a favorable prognosis for being controlled with medications but usually requires treatment for many years without full spontaneous remission. The range of scores does allow distinctions of severity among various seizure syndromes. Clearly in terms of guidance to the clinician, the most useful scores are those at the extremes, but scores in the middle also contain some useful information. Note that a score in the middle does not necessarily reflect uncertainty. Raters were asked to spread their scores uniformly, and thus no more than four syndromes could be assigned to any score.

Concerns of experts were addressed through several rounds in the development of this scale. Consensus could not be reached for certain rare seizure syndromes. For example, several raters commented that certain disorders such as reading epilepsy were so uncommon that they could not use their own experience as a guide to assigning a severity score. In addition, several raters believed that certain syndromes were so variable that assigning a single score was difficult. Therefore our scores reflect the average of severity ratings for each syndrome, and for some of these syndromes, a wider range of severity was assigned than for others. A limitation of this study is that only 12 of the 18 experts who agreed to participate in the study provided data for all rounds; however, there did not appear to be differences in the initial rating of those who finished the study compared with those who did not complete all rounds. A final limitation is that these syndrome severity scores are limited to the current ILAE classification, and modifications will be required should the classification scheme change.

Despite these limitations, these epileptic syndrome severity scores address weaknesses of past methods for rating severity. We argue that measurement of seizure condition severity with indicators such as age at onset, seizure duration, total number of seizures, or seizure severity is not adequate to capture the severity of a complex condition such as epilepsy in children. We propose that a valid measure of condition severity should be based on multiple indicators and that epileptic syndrome severity should be included as one of the indicators. As one example of why syndrome is important, there seems to be a significant difference between the complex partial seizures of benign focal epilepsy of childhood and a similar seizure type in the child with localization-related symptomatic epilepsy of frontal or temporal lobe origin.

We do not advocate that only an epileptic syndrome severity score be used to reflect the severity of a child's epilepsy condition. Minimally, this syndrome score should be coupled with an estimate of seizure frequency for it to reflect the multidimensionality of condition severity and for it to be responsive enough to be useful in studies of epilepsy therapies or evaluations of change over time. Further assessment of the validity of these syndrome severity scores is needed. For example, these severity scores could be coupled with other seizure variables such as seizure frequency and the combination evaluated in relation to predicting behavior or quality-of-life outcomes. If these further studies provide empirical support for the validity of these epileptic syndrome severity scores, they could then be used in research as one part of a more comprehensive assessment. We expect that syndrome severity would best be used along with seizure frequency, seizure type, age at onset and duration, severity of seizure social impact, AED adverse effects, and illness-specific quality-of-life measures in the multidimensional evaluations of both current and ongoing effects of epilepsy on children.

Although these severity scores were designed primarily for research on groups of children with epilepsy, they might be of use clinically. For the individual child, they could be used for educating parents about the relative severity of the child's condition, acknowledging individual variability.


  1. Top of page
  2. Abstract
  6. Acknowledgments

Acknowledgment:  This research was supported by grant NS22416 from the National Institute of Neurological Disorders and Stroke to J.K.A. and by grant NR04536 from the National Institute of Nursing Research to J.K.A. We thank C. Benson and J. Critchfield for help with data collection, P. Dexter for editorial comments, and C. Benson for editorial assistance. Dr. Jan Buelow's work was supported by Indiana University School of Nursing, T32 Training Grant, NR 07066, National Institute of Nursing Research, PI: Joan K. Austin, DNS, RN, FAAN. We also thank the following child neurologists who participated in rating the severity of each syndrome: Carol Camfield, Harry Chugani, Patricia Crumrine, Edwin Dodson, Michael Duchowny, Roy Elterman, Tracy Glauser, May Griebel, Brad Hale, Gregory Holmes, Prakash Kotagal, Wendy Mitchell, Solomon Moshe, Doug Nordli, Carter Snead, James Wheless, Mary Zupanc, and anonymous.


  1. Top of page
  2. Abstract
  6. Acknowledgments
  • 1
    Westbrook LE, Silver EJ, Coupey SM, et al. Social characteristics of adolescents with idiopathic epilepsy: a comparison to chronically ill and nonchronically ill peers. J Epilepsy 1991;4: 8794.
  • 2
    Carlton-Ford S, Miller R, Brown M, et al. Epilepsy and children's social and psychological adjustment. J Health Social Behav 1995;36: 285301.
  • 3
    Hoare P. The development of psychiatric disorder among schoolchildren with epilepsy. Dev Med Child Neurol 1984;26: 313.
  • 4
    Hermann BP, Whitman S, Dell J. Correlates of behavior problems and social competence in children with epilepsy, aged 6–11. In: HermannBP, SeidenbergM, eds. Childhood epilepsies: neuropsychological, psychosocial and intervention aspects. New York : John Wiley & Sons, 1989: 14357.
  • 5
    Austin JK, Risinger MW, Beckett LA. Correlates of behavior problems in children with epilepsy. Epilepsia 1992;33: 111522.
  • 6
    Mitchell WG, Scheier LM, Baker SA. Psychosocial, behavioral, and medical outcomes in children with epilepsy: a developmental risk factor model using longitudinal data. Pediatrics 1994;94: 4717.
  • 7
    Caplan R, Arbelle S, Magharious W, et al. Psychopathology in pediatric complex partial and primary generalized epilepsy. Dev Med Child Neurol 1998;40: 80511.
  • 8
    Schoenfeld J, Seidenberg M, Woodard A, et al. Neuropsychological and behavioral status of children with complex partial seizures. Dev Med Child Neurol 1999;41: 72431.
  • 9
    Commission on Classification and Terminology of the International League Against Epilepsy. Proposal for revised clinical and electroencephalographic classification of epileptic seizures. Epilepsia 1981;22: 489501.
  • 10
    Commission on Classification and Terminology of the International League Against Epilepsy. Proposal for revised classification of epilepsies and epileptic syndromes. Epilepsia 1989;30: 38999.
  • 11
    Cramer JA, Smith DB, Mattson RH, et al. A method of quantification for the evaluation of antiepileptic drug therapy. Neurology 1983;33(suppl 1):2637.
  • 12
    Austin JK, Dunn DW, Huster GA. Childhood epilepsy and asthma: changes in behavior problems related to gender and change in severity. Epilepsia 2000;41: 61523.
  • 13
    Baker GA, Smith DF, Dewey M, et al. The development of a seizure severity scale as an outcome measure in epilepsy. Epilepsy Res 1991;8: 24551.
  • 14
    Carpay JA, Vermuelen J, Stroink H, et al. Seizure severity in children with epilepsy: a parent-completed scale compared with clinical data. Epilepsia 1997;38: 34652.
  • 15
    Herdman M, Rajmil L, Ravens-Sieverer U, et al. European Kidscreen and Disabkids groups: expert consensus in the development of a European health-related quality of life measure for children and adolescents: a Delphi study. Acta Paediatr 2002;91: 138590.
  • 16
    Burns SP, Frederick P, Johanses JM, et al. Rehabilitation of traumatic injuries: use of the Delphi method to identify topics for evidence-based review. Am J Phys Med Rehab 2003;82: 4104.
  • 17
    O'Boyle C, Jackson M, Henly SJ. Staffing requirements for infection control programs in US health care facilities: Delphi project. Am J Infect Control 2002;30: 32133.
  • 18
    LinstoneHA, TuroffM, eds. The Delphi method: techniques and applications. Reading , MA : Addison-Wesley, 1975.
  • 19
    Crisp J, Pelletier D, Duffield S, et al. The Dephi method? Nurs Res 1997;46: 1168.
  • 20
    Wirrell EC, Camfield CS, Camfield PR, et al. Long-term psychosocial outcome in typical absence epilepsy. Arch Pediatr Adolesc Med 1997;151: 1528.