TABLE 1: KEY VARIABLES:

If these variables could be extracted from the electronic health record automatically, an antiepileptic drug efficacy study could be performed.

1) Exposure:

a) Prescribed medication extraction (S+U)

2) Covariables:

a) Age, gender, & ethnicity (S)

b) Socioeconomic status by zip code, type of insurance (S)

c) Epilepsy type (U)

d) Seizure type (U)

e) Underlying cause of the epilepsy (S+U):

i) genetic syndrome

ii) structural lesions

(1) strokes

(2) tumors

(3) surgery

(4) bleeding

iii) autoimmune disease

iv) neurodegenerative disease

v) traumatic brain injury

f) Age of onset (S)

g) Medication allergies (S)

h) Co-morbid diseases, aka past medical history (S+U)

i) Refractory epilepsy: as defined by number of other medications tried, or surgery tried

j) Epilepsy risk factors (S+U)

i) Abnormal birth and development (e.g. prematurity, developmental delay, autism)

ii) History of brain infection

iii) Substance abuse

iv) Family history of epilepsy

v) History of febrile seizure

k) Prior EEG abnormalities (U)

3)  Outcome: seizure frequency and surrogates thereof

a) Seizure frequency (U)

b) Use of rescue medications (S+U)

c) Need for increase in therapy: e.g. doses increased, new meds added (S+U)

d) Use of the nursing telephone help-line (S+U)

e) Use of the emergency room (S+U)

f) Hospital admissions (S+U)

g) Total health-care cost (calculated based on the above)

(S) = structured fields; (U) = Unstructured data; (S+U) = structured fields and unstructured data.