Readmission Following Surgical Resection for Intractable Epilepsy: Nationwide Rates, Causes, Predictors, and Outcomes

Readmission Following Surgical Resection for Intractable Epilepsy: Nationwide Rates, Causes,... Abstract BACKGROUND Hospital readmissions can be detrimental to patients and may interfere with the potential benefits of the therapeutic procedure. Government agencies have begun to focus on reducing readmissions; however, the etiology of readmissions is lacking. OBJECTIVE To report the national rates, risk factors, and outcomes associated with 30- and 90-d readmissions following surgery for intractable epilepsy. METHODS We queried the Nationwide Readmissions Database from January to September 2013 using International Classification of Diseases, Ninth Edition, Clinical Modification codes to identify all patients with intractable epilepsy, who underwent hemispherectomy (01.52), brain lobectomy (01.53), amydalohippocampectomy, or partial lobectomy (01.59). Predictor variables included epilepsy type, presurgical diagnostic testing, surgery type, medical complications, surgical complications, and discharge disposition. RESULTS In 1587 patients, the 30- and 90-d readmission rates were 11.5% and 16.8%, respectively. The most common reasons for readmission were persistent epilepsy, video electroencephalography monitoring, postoperative infection, and postoperative central nervous system complication. In multivariable analysis, risk factors associated with both 30- and 90-d readmission were Medicare payer status, lowest quartile of median income, depression, hemispherectomy, and postoperative complications (P < .05). The only unique predictor of 30-d readmission was small bedsize hospital (P = .001). Readmissions within 30 d were associated with longer length of stay (6.8 vs 5.8 d), greater costs ($18 660 vs $15 515), and increased adverse discharges (26.4% vs 21.8%). CONCLUSION Following epilepsy surgery, most readmissions that occurred within 30 d can be attributed to management of persistent epilepsy and predicted by Medicare payer status, depression, and complications. These data can assist the clinician in preventing readmissions and assist policy makers determine which admissions are potentially avoidable. Complications, Epilepsy, Readmissions, Seizure ABBREVIATIONS ABBREVIATIONS AHC amygdalohippocampectomy AHRQ Agency for Healthcare Research and Quality CI confidence interval DVT deep venous thrombosis HCUP Healthcare Cost and Utilization Project ICD-9-CM International Classification of Diseases, Ninth Edition, Clinical Modification IEEG intracranial electroencephalography NIS Nationwide Inpatient Sample NRD Nationwide Readmissions Database OR odds ratio SD standard deviation SID State Inpatient Databases VEEG video electroencephalography Hospital readmissions rate is an important metric for evaluating quality of patient care, and the advent of the Patient Protection and Affordable Care Act has required agencies such as the Centers for Medicare and Medicaid Services to penalize hospitals with excessive 30-d readmission rates. Furthermore, readmissions after a longer period (eg, 90 d) may reveal unique problems. Kim et al1 found that 30-d readmission indices underestimated true readmission rates. A few studies have reported readmissions in patients undergoing surgery for intractable epilepsy using the Nationwide Inpatient Sample (NIS), which can evaluate in-hospital outcomes.2,3 Furthermore, other studies are limited to a handful of single-center retrospective cohorts with small sample sizes.4 In 2013, under the Healthcare Cost and Utilization Project (HCUP), the Agency for Healthcare Research and Quality (AHRQ) released the Nationwide Readmissions Database (NRD), which combines the strengths of the NIS and the State Inpatient Databases (SID) to enable researchers to perform analyses at the national level. Thus, we sought to report the national rates, risk factors, reasons, and outcomes associated with 30- and 90-d readmission. METHODS Data Source The NRD was released in 2013 as a new addition to the HCUP family and was designed to enable researchers to effectively analyze readmissions.1 The NRD contains approximately 36 million weighted discharge records5 and is equipped with the strengths of NIS, namely sample size, with the addition of data elements from the SID, including the patient-linkage number, days-to-event variable, and hospital length of stay. The AHRQ’s Clinical Classification Software is a helpful tool that collapses the multitude of International Classification of Diseases, Ninth Edition, Clinical Modification (ICD-9-CM) codes into clinically meaningful categories.6 The required HCUP Data Use Agreement Training prohibits users from reporting any cell size representing less than 11 patients.5 Neither Institutional Review Board (IRB) approval nor patient consent was required. Study Design The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline checklist for retrospective cohort study was utilized. To define the index visit, we queried the NRD to identify all patients in the 2013 NRD who underwent epilepsy surgery between January and September. We excluded the last 3 mo to enable the evaluation of 90-d readmissions. Participants We searched all NRD diagnosis fields to identify patients with intractable epilepsy-related ICD-9-CM diagnoses (345.x1, exclude 345.2, 345.3, 345.61, 345.71).7 Among this cohort, only patients with ICD-9-CM procedure codes for epilepsy surgery (01.52, hemispherectomy; 01.53, brain lobectomy; 01.59, partial brain lobectomy or amygdalohippocampectomy [AHC]) were included.3,7,8 Setting A patient's first readmission following surgery was considered a readmission, but all preadmissions and subsequent readmissions were excluded. The NRD treats same-day readmissions and transfers as a single discharge record. Predictor Variables We abstracted and defined variables pertaining to patient demographics, pre-existing comorbidities, admission characteristics, hospital characteristics, presurgical diagnostic testing, and type of procedure (hemispherectomy, lobectomy, AHC, partial lobectomy/lesionectomy). The patient demographic variables included categorical age, gender, payer status, and median household income. Pre-existing comorbidities were scored using the Elixhauser Comorbidity Index as computed by AHRQ.9,10 Admission characteristics included admission type and day of admission. Hospital characteristics included bedsize, hospital type, and hospital location. We used ICD-9-CM codes to define more specific types of epilepsy, including generalized epilepsy (345.01, 345.11), localization-related complex partial epilepsy (345.41), localization-related simple partial epilepsy (345.51), and other intractable epilepsy (345.81, 345.91). The presurgical diagnostic tests evaluated were video electroencephalography (VEEG; 89.19) and intracranial electroencephalography (IEEG; 02.93). Outcome Variables The primary outcomes of interest were hospital readmission within 30 and 90 d. However, we defined additional variables for complications, length of stay, total hospital costs, discharge disposition, and in-hospital mortality. We defined surgical complication using secondary ICD-9-CM codes including stroke and hematoma (432.1, 997.00, 997.01, 997.02, 997.09, 998.11, 998.12), postoperative intracranial infections (320.0, 320.7, 320.81-320.9, 322.9, 998.51, 998.59), status epilepticus (345.3), hydrocephalus (331.3, 331.4), ventriculoperitoneal shunt placement (02.34), blood transfusion (99.01-99.09), postoperative mechanical ventilation (96.70-96.79), and meningitis (320-322, 326).11 Common in-hospital complication variables included cardiac (997.1, 410), respiratory (518.4, 518.5, 518.81-518.84), renal/urinary (584 and 997.5), and thromboembolic (415, 415.11-415.19, 451.0-451.9).12 Discharge disposition was dichotomized into “routine” and “adverse discharge,” meaning any discharge other than routine (transfer to skilled nursing, intermediate care facility, short-term acute-care hospital, or home health care).5 The NRD contains a total hospital charges variable; however, HCUP provides users with cost-to-charge ratio files to convert “charges” to “cost,” which reflects how much hospitals actually received in payment.13 Statistical Methods The SPSS v.24 statistical software (IBM Inc, Armonk, New York) was used for analyses (alpha 0.05). In SPSS, we used the complex samples function, which requires the stratum, cluster, and discharge weights to produce national estimates. Subgroup analysis was performed to compare causes of admission by type of surgery. Risk factors for readmission were identified using univariate and multivariable analysis (Chi-square test for categorical variables, independent samples t-test for continuous variables). Due to limited sample size for individual comorbidity and complication groups, these variables were treated as composite variables in analyses. Variables with statistical significance in the univariate analysis and/or strong clinical justification were entered in multivariable logistic regression model. RESULTS Participants A total of 1587 patients who underwent elective AHC/partial lobectomy (77.9%), lobectomy (23.2%), or hemispherectomy (4.2%) for intractable epilepsy were identified in the NRD between January and September of 2013. The intractable epilepsy types included complex partial (57.0%), simple partial (17.6%), generalized (4.8%), and other (24.7%). The rate of IEEG and VEEG presurgical diagnostic testing during the index hospitalization was 35.8% and 27.4%, respectively. Descriptive Data The mean age of the surgical cohort (±standard deviation [SD]) was 29.9 ± 1.4 yr, 49.2% were female, and 56.6% had at least 1 pre-existing comorbidity. The most common comorbidities were fluid/electrolyte disorder (14.0%), paralysis (12.6%), depression (11.9%), obesity (9.6%), and hypothyroidism (6.1%). The vast majority of surgeries were performed in teaching hospitals, so this predictor variable was not considered in further analyses. A total of 456 surgical complications and 78 medical complications were reported during the index surgical visits, but fewer than 11 patients suffered in-hospital mortality. The most common surgical complications were blood transfusion, stroke or hematoma, and intracranial infection (Table 1). Furthermore, the most common medical complications were thromboembolic and respiratory (Table 1). The mean cost of a surgical visit was $49 779, and the average length of stay was 8 ± 0.6 d. TABLE 1. Hospitalization-Associated Outcomes of 1587 Patients who Underwent Surgery for Intractable Epilepsy, NRD 2013 Variable Total patients (n) 1587 Surgical (index) visit outcomes  Mean LOS in days (SD) 8.3 (0.6)  Mean total costs, $USD (SD) $49 779 (5125)  Medical complications (n) 78  Surgical complications (n) 456  Adverse discharge disposition (%) 15.6 30-d reasons for readmission, n 183  Diagnoses (ICD-9-CM diagnosis codes), n   Epilepsy (345.3, 345.41, 345.51, 345.90) 58   Postoperative infection (998.59) 28   Postoperative CNS complication (997.01) 18   Headache (784.0) 14   Other (individual codes <11 cases) 65  Procedures (ICD-9-CM procedure codes), n   V-EEG monitoring (89.19) 26   Venous catherization (38.93) 25   Other (individual codes <11 cases) 14 30-d readmission visit outcomes, n 183  Mean length of stay, days (SD) 6.8 (0.7)  Mean total costs, $USD (SD) $18 660 (2921)  Adverse discharge disposition (%) 26.4 90-d reasons for readmission, n 267  Diagnoses (ICD-9-CM diagnosis codes), n   Epilepsy (345.41, 345.50, 345.51, 345.60, 345.81, 345.90, 345.91) 103   Postoperative infection (998.59) 34   Postoperative CNS complication (997.01) 18   Headache (784.0 346.90) 16  Procedures (ICD-9-CM procedure codes), n   V-EEG monitoring (89.19) 47   Venous catherization (38.93) 25   Other (individual codes < 11 cases) 75 90-d readmission visit outcomes, n  Mean length of stay, days (SD) 5.8 (0.3)  Mean total costs, $USD (SD) $15 515 (1360)  Adverse discharge disposition (%) 21.8 Variable Total patients (n) 1587 Surgical (index) visit outcomes  Mean LOS in days (SD) 8.3 (0.6)  Mean total costs, $USD (SD) $49 779 (5125)  Medical complications (n) 78  Surgical complications (n) 456  Adverse discharge disposition (%) 15.6 30-d reasons for readmission, n 183  Diagnoses (ICD-9-CM diagnosis codes), n   Epilepsy (345.3, 345.41, 345.51, 345.90) 58   Postoperative infection (998.59) 28   Postoperative CNS complication (997.01) 18   Headache (784.0) 14   Other (individual codes <11 cases) 65  Procedures (ICD-9-CM procedure codes), n   V-EEG monitoring (89.19) 26   Venous catherization (38.93) 25   Other (individual codes <11 cases) 14 30-d readmission visit outcomes, n 183  Mean length of stay, days (SD) 6.8 (0.7)  Mean total costs, $USD (SD) $18 660 (2921)  Adverse discharge disposition (%) 26.4 90-d reasons for readmission, n 267  Diagnoses (ICD-9-CM diagnosis codes), n   Epilepsy (345.41, 345.50, 345.51, 345.60, 345.81, 345.90, 345.91) 103   Postoperative infection (998.59) 34   Postoperative CNS complication (997.01) 18   Headache (784.0 346.90) 16  Procedures (ICD-9-CM procedure codes), n   V-EEG monitoring (89.19) 47   Venous catherization (38.93) 25   Other (individual codes < 11 cases) 75 90-d readmission visit outcomes, n  Mean length of stay, days (SD) 5.8 (0.3)  Mean total costs, $USD (SD) $15 515 (1360)  Adverse discharge disposition (%) 21.8 Discharge disposition was dichotomized into the variable “adverse discharge” in which any discharge other than routine (transfer to skilled nursing, intermediate care facility, short-term acute-care hospital, home health care). View Large TABLE 1. Hospitalization-Associated Outcomes of 1587 Patients who Underwent Surgery for Intractable Epilepsy, NRD 2013 Variable Total patients (n) 1587 Surgical (index) visit outcomes  Mean LOS in days (SD) 8.3 (0.6)  Mean total costs, $USD (SD) $49 779 (5125)  Medical complications (n) 78  Surgical complications (n) 456  Adverse discharge disposition (%) 15.6 30-d reasons for readmission, n 183  Diagnoses (ICD-9-CM diagnosis codes), n   Epilepsy (345.3, 345.41, 345.51, 345.90) 58   Postoperative infection (998.59) 28   Postoperative CNS complication (997.01) 18   Headache (784.0) 14   Other (individual codes <11 cases) 65  Procedures (ICD-9-CM procedure codes), n   V-EEG monitoring (89.19) 26   Venous catherization (38.93) 25   Other (individual codes <11 cases) 14 30-d readmission visit outcomes, n 183  Mean length of stay, days (SD) 6.8 (0.7)  Mean total costs, $USD (SD) $18 660 (2921)  Adverse discharge disposition (%) 26.4 90-d reasons for readmission, n 267  Diagnoses (ICD-9-CM diagnosis codes), n   Epilepsy (345.41, 345.50, 345.51, 345.60, 345.81, 345.90, 345.91) 103   Postoperative infection (998.59) 34   Postoperative CNS complication (997.01) 18   Headache (784.0 346.90) 16  Procedures (ICD-9-CM procedure codes), n   V-EEG monitoring (89.19) 47   Venous catherization (38.93) 25   Other (individual codes < 11 cases) 75 90-d readmission visit outcomes, n  Mean length of stay, days (SD) 5.8 (0.3)  Mean total costs, $USD (SD) $15 515 (1360)  Adverse discharge disposition (%) 21.8 Variable Total patients (n) 1587 Surgical (index) visit outcomes  Mean LOS in days (SD) 8.3 (0.6)  Mean total costs, $USD (SD) $49 779 (5125)  Medical complications (n) 78  Surgical complications (n) 456  Adverse discharge disposition (%) 15.6 30-d reasons for readmission, n 183  Diagnoses (ICD-9-CM diagnosis codes), n   Epilepsy (345.3, 345.41, 345.51, 345.90) 58   Postoperative infection (998.59) 28   Postoperative CNS complication (997.01) 18   Headache (784.0) 14   Other (individual codes <11 cases) 65  Procedures (ICD-9-CM procedure codes), n   V-EEG monitoring (89.19) 26   Venous catherization (38.93) 25   Other (individual codes <11 cases) 14 30-d readmission visit outcomes, n 183  Mean length of stay, days (SD) 6.8 (0.7)  Mean total costs, $USD (SD) $18 660 (2921)  Adverse discharge disposition (%) 26.4 90-d reasons for readmission, n 267  Diagnoses (ICD-9-CM diagnosis codes), n   Epilepsy (345.41, 345.50, 345.51, 345.60, 345.81, 345.90, 345.91) 103   Postoperative infection (998.59) 34   Postoperative CNS complication (997.01) 18   Headache (784.0 346.90) 16  Procedures (ICD-9-CM procedure codes), n   V-EEG monitoring (89.19) 47   Venous catherization (38.93) 25   Other (individual codes < 11 cases) 75 90-d readmission visit outcomes, n  Mean length of stay, days (SD) 5.8 (0.3)  Mean total costs, $USD (SD) $15 515 (1360)  Adverse discharge disposition (%) 21.8 Discharge disposition was dichotomized into the variable “adverse discharge” in which any discharge other than routine (transfer to skilled nursing, intermediate care facility, short-term acute-care hospital, home health care). View Large National Rates, Causes of Readmission, and Discharge Outcomes A total of 183 (11.5%) patients from the surgical cohort were readmitted within 30 d, and 267 (16.8%) patients were readmitted within 90 d. The most common reasons for both 30- and 90-d readmission were persistent epilepsy, VEEG monitoring, postoperative infection, and postoperative CNS complication (997.01; Table 1). The rate of persistent epilepsy varied by surgery type: partial lobectomy/AHC (6.0%), total lobectomy (6.8%), and hemispherectomy (3.0%). The rate of postoperative infection also varied by surgery type: partial lobectomy/AHC (2.0%), total lobectomy (1.3%), and hemispherectomy (7.2%). VEEG did not vary by type of surgery. Readmission visits within 30 d were associated with longer length of stay during readmission (6.8 vs 5.8 d), greater costs ($18 660 vs $15 515), and increased percentage of adverse discharges (26.4% vs 21.8%) than readmissions within 90 d (Table 1). Main Results The complete univariate analysis of various variables on 30- and 90-d readmissions is shown in Tables 2 and 3. The rate of 30-d (P = .007) readmission, but not 90-d readmission (P = .078) was significantly higher among Medicaid and Medicare patients compared to privately insured. Overall, the rate of both 30-d (P = .024) and 90-d (P = .031) readmissions increased with the number of pre-existing comorbidities. Upon examination of 31 individual Elixhauser comorbidities, depression (11.9%) was the only one associated with significantly higher rates of 30-d (27.9% vs 11.5%, P < .001) and 90-d readmission (33.9% vs 16.8%, P < .001). Compared to the average medical complication rate among surgical patients of 5%, patients readmitted at 30 or 90 d had rates of 33% (P = .002) and 39% (P = .005), respectively, during their initial surgical visit. Descriptively, the rates of readmission among patients who suffered from surgical complications during the index visit were higher, but were not statistically significant. Lastly, prolonged length of stay and increased total hospital costs during the index visit were associated with 30- and 90-d readmission (P < .0001). We performed a multivariable analysis to identify predictors of prolonged length of stay, and the significant variables were 3+ pre-existing comorbidities (odds ratio [OR]: 7.8, 95% confidence interval [CI]: 2.8-22.0), IEEG (OR: 20.1, 95% CI: 7.9-54.7), VEEG (OR: 2.8, 95% CI: 1.1-7.6), hemispherectomy (OR: 4.2, 95% CI: 1.0-18.7), respiratory complications (OR: 11.4, 95% CI: 2.1-61.7), perioperative stroke/hematoma (OR: 9.7, 95% CI: 3.2-28.8), and status epilepticus (OR: 7.5, 95% CI: 1.2-46.0). TABLE 2. The Univariate (Unadjusted) Effect of Patient Demographics, Hospital Characteristics on 30- and 90-d Readmission Rates in 1587 Patients who Underwent Surgery for Intractable Epilepsy, NRD 2013 Variable Patients N (%) 30-d readmission rate N (%) 90-d readmission rate N (%)d Total patients (n) 1587 183 (11.5) 267 (16.8) Age group NS NS  0-20 562 (35.4) 58 (10.3) 111 (19.8)  21-35 418 (26.4) 49 (11.8) 59 (14.2)  36-50 332 (20.9) 44 (13.2) 58 (17.4)  51+ 275 (17.3) 32 (11.6) 38 (13.7) Gender NS NS  Male 806 (50.8) 90 (11.1) 140 (17.4)  Female 781 (49.2) 93 (12.0) 126 (16.2) Payera b NS  Private 888 (56.0 66 (7.5) 121 (13.7)  Medicare 257 (16.2) 57 (22.1) 68 (26.6)  Medicaid 345 (22.4) 53 (15.0) 64 (18.1)  Other 84 (5.3) c c Median annual household incomea NS NS  0-25th percentile 313 (20.2) 50 (15.9) 68 (21.8)  26th-50th percentile 382 (24.7) 37 (9.7) 58 (15.2)  51st-75th percentile 425 (27.4) 34 (8.0) 44 (10.3)  76th-100th percentile 428 (27.7) 63 (14.6) 94 (22.0) Bed size of hospital NS NS  Small 36 (2.3) c c  Medium 212 (13.3) 18 (8.4) 24 (11.1)  Large 1339 (84.4) 156 (11.7) 234 (17.4) Weekend admissiona NS NS  Weekday 1569 (98.9) 183 (11.7) 263 (16.7)  Weekend 17 (1.1) c c Hospital location NS NS  Small metropolitan 507 128 (11.9) 179 (16.5)  Large metropolitan 1,080 55 (10.8) 88 (17.4) Pre-existing comorbidity score (%)a b b  0 689 (43.4) 55 (7.9) 82 (11.8)  1 471 (29.7) 56 (11.9) 98 (20.7)  2 249 (15.7) 26 (10.2) 32 (12.7)  3+ 177 (11.2) 47 (26.6) 56 (31.4) Significant comorbidities  Depression 189 (11.9) 53 (27.9) 64 (33.9) Variable Patients N (%) 30-d readmission rate N (%) 90-d readmission rate N (%)d Total patients (n) 1587 183 (11.5) 267 (16.8) Age group NS NS  0-20 562 (35.4) 58 (10.3) 111 (19.8)  21-35 418 (26.4) 49 (11.8) 59 (14.2)  36-50 332 (20.9) 44 (13.2) 58 (17.4)  51+ 275 (17.3) 32 (11.6) 38 (13.7) Gender NS NS  Male 806 (50.8) 90 (11.1) 140 (17.4)  Female 781 (49.2) 93 (12.0) 126 (16.2) Payera b NS  Private 888 (56.0 66 (7.5) 121 (13.7)  Medicare 257 (16.2) 57 (22.1) 68 (26.6)  Medicaid 345 (22.4) 53 (15.0) 64 (18.1)  Other 84 (5.3) c c Median annual household incomea NS NS  0-25th percentile 313 (20.2) 50 (15.9) 68 (21.8)  26th-50th percentile 382 (24.7) 37 (9.7) 58 (15.2)  51st-75th percentile 425 (27.4) 34 (8.0) 44 (10.3)  76th-100th percentile 428 (27.7) 63 (14.6) 94 (22.0) Bed size of hospital NS NS  Small 36 (2.3) c c  Medium 212 (13.3) 18 (8.4) 24 (11.1)  Large 1339 (84.4) 156 (11.7) 234 (17.4) Weekend admissiona NS NS  Weekday 1569 (98.9) 183 (11.7) 263 (16.7)  Weekend 17 (1.1) c c Hospital location NS NS  Small metropolitan 507 128 (11.9) 179 (16.5)  Large metropolitan 1,080 55 (10.8) 88 (17.4) Pre-existing comorbidity score (%)a b b  0 689 (43.4) 55 (7.9) 82 (11.8)  1 471 (29.7) 56 (11.9) 98 (20.7)  2 249 (15.7) 26 (10.2) 32 (12.7)  3+ 177 (11.2) 47 (26.6) 56 (31.4) Significant comorbidities  Depression 189 (11.9) 53 (27.9) 64 (33.9) RR, readmission rate, NS, not significant. aExplanations provided to the reviewer about the apparent numerical discrepancies based on data present in the NRD for analysis. bStatistical significance (P < .05) via Pearson chi-square test. cLess than the HCUP reporting minimum of 11 cases. dThe 30 and 90-d readmission cohorts were not mutually exclusive, as the 90-d readmission cohort includes any patient that was readmitted within 90 d, including readmissions within 30 d. View Large TABLE 2. The Univariate (Unadjusted) Effect of Patient Demographics, Hospital Characteristics on 30- and 90-d Readmission Rates in 1587 Patients who Underwent Surgery for Intractable Epilepsy, NRD 2013 Variable Patients N (%) 30-d readmission rate N (%) 90-d readmission rate N (%)d Total patients (n) 1587 183 (11.5) 267 (16.8) Age group NS NS  0-20 562 (35.4) 58 (10.3) 111 (19.8)  21-35 418 (26.4) 49 (11.8) 59 (14.2)  36-50 332 (20.9) 44 (13.2) 58 (17.4)  51+ 275 (17.3) 32 (11.6) 38 (13.7) Gender NS NS  Male 806 (50.8) 90 (11.1) 140 (17.4)  Female 781 (49.2) 93 (12.0) 126 (16.2) Payera b NS  Private 888 (56.0 66 (7.5) 121 (13.7)  Medicare 257 (16.2) 57 (22.1) 68 (26.6)  Medicaid 345 (22.4) 53 (15.0) 64 (18.1)  Other 84 (5.3) c c Median annual household incomea NS NS  0-25th percentile 313 (20.2) 50 (15.9) 68 (21.8)  26th-50th percentile 382 (24.7) 37 (9.7) 58 (15.2)  51st-75th percentile 425 (27.4) 34 (8.0) 44 (10.3)  76th-100th percentile 428 (27.7) 63 (14.6) 94 (22.0) Bed size of hospital NS NS  Small 36 (2.3) c c  Medium 212 (13.3) 18 (8.4) 24 (11.1)  Large 1339 (84.4) 156 (11.7) 234 (17.4) Weekend admissiona NS NS  Weekday 1569 (98.9) 183 (11.7) 263 (16.7)  Weekend 17 (1.1) c c Hospital location NS NS  Small metropolitan 507 128 (11.9) 179 (16.5)  Large metropolitan 1,080 55 (10.8) 88 (17.4) Pre-existing comorbidity score (%)a b b  0 689 (43.4) 55 (7.9) 82 (11.8)  1 471 (29.7) 56 (11.9) 98 (20.7)  2 249 (15.7) 26 (10.2) 32 (12.7)  3+ 177 (11.2) 47 (26.6) 56 (31.4) Significant comorbidities  Depression 189 (11.9) 53 (27.9) 64 (33.9) Variable Patients N (%) 30-d readmission rate N (%) 90-d readmission rate N (%)d Total patients (n) 1587 183 (11.5) 267 (16.8) Age group NS NS  0-20 562 (35.4) 58 (10.3) 111 (19.8)  21-35 418 (26.4) 49 (11.8) 59 (14.2)  36-50 332 (20.9) 44 (13.2) 58 (17.4)  51+ 275 (17.3) 32 (11.6) 38 (13.7) Gender NS NS  Male 806 (50.8) 90 (11.1) 140 (17.4)  Female 781 (49.2) 93 (12.0) 126 (16.2) Payera b NS  Private 888 (56.0 66 (7.5) 121 (13.7)  Medicare 257 (16.2) 57 (22.1) 68 (26.6)  Medicaid 345 (22.4) 53 (15.0) 64 (18.1)  Other 84 (5.3) c c Median annual household incomea NS NS  0-25th percentile 313 (20.2) 50 (15.9) 68 (21.8)  26th-50th percentile 382 (24.7) 37 (9.7) 58 (15.2)  51st-75th percentile 425 (27.4) 34 (8.0) 44 (10.3)  76th-100th percentile 428 (27.7) 63 (14.6) 94 (22.0) Bed size of hospital NS NS  Small 36 (2.3) c c  Medium 212 (13.3) 18 (8.4) 24 (11.1)  Large 1339 (84.4) 156 (11.7) 234 (17.4) Weekend admissiona NS NS  Weekday 1569 (98.9) 183 (11.7) 263 (16.7)  Weekend 17 (1.1) c c Hospital location NS NS  Small metropolitan 507 128 (11.9) 179 (16.5)  Large metropolitan 1,080 55 (10.8) 88 (17.4) Pre-existing comorbidity score (%)a b b  0 689 (43.4) 55 (7.9) 82 (11.8)  1 471 (29.7) 56 (11.9) 98 (20.7)  2 249 (15.7) 26 (10.2) 32 (12.7)  3+ 177 (11.2) 47 (26.6) 56 (31.4) Significant comorbidities  Depression 189 (11.9) 53 (27.9) 64 (33.9) RR, readmission rate, NS, not significant. aExplanations provided to the reviewer about the apparent numerical discrepancies based on data present in the NRD for analysis. bStatistical significance (P < .05) via Pearson chi-square test. cLess than the HCUP reporting minimum of 11 cases. dThe 30 and 90-d readmission cohorts were not mutually exclusive, as the 90-d readmission cohort includes any patient that was readmitted within 90 d, including readmissions within 30 d. View Large TABLE 3. The Univariate (Unadjusted) Effect of Epilepsy Diagnosis, Presurgical Diagnostic Testing, Type of Surgery, Medical Complications, and Surgical Complications on 30-d and 90-d Readmission Rates in 1587 who Underwent Surgery for Intractable Epilepsy, NRD 2013 Patients N (%) 30-d RR (%) 90-d RR (%) d Total Patients (n) 1587 183 (11.5) 267 (16.8) Epilepsy diagnosisa NS NS  Generalized 76 (4.8) b 17 (22.4)  Complex Partial 904 (57.0) 111 (12.3) 148 (16.4)  Simple Partial 279 (17.6) 33 (11.7) 43 (15.5)  Other 392 (24.7) 42 (10.6) 64 (16.4) Presurgical diagnostic testing NS NS  VEEG 435 (27.2) 56 (12.8) 79 (18.2)  IEEG 568 (35.8) 66 (11.7) 88 (15.5) Type of surgerya NS NS  Hemispherectomy 66 (4.2) 15 (22.8) 24 (36.5)  Lobectomy 368 (23.2) 44 (11.9) 55 (14.9)  Amygdalohippocampectomy or partial lobectomy 1236 (77.9) 129 (10.4) 197 (16.0) Medical complications composite c c  0 1508 (95.0%) 157 (10.4) 236 (15.6)  1 79 (5%) 26 (33.0) 31 (39.0)  2 b b b  3+ b b b Medical complications individual  Cardiac b b b  Respiratory 24 (1.5) b 11 (45.4c)  Renal and urinary 14 (0.9) b b  Thromboembolic 32 (2.0) 18 (57.9) 18 (57.9) Surgical complications composite NS NS  0 1356 (85.4) 145 (10.7) 216 (16.0)  1 197 (12.4) 32 (16.2) 40 (20.1)  2 28 (1.8) b 11 (37.1)  3+ b b b Surgical complications individual NS NS  Stroke/hematoma 78 (4.9) b 14 (18.6)  Postoperative intracranial infections 34 (2.2) b b  Status epilepticus b b b  Hydrocephalus 11 (0.7) b b  VP shunt b b b  Blood transfusion 122 (7.7) 26 (21.4) 32 (26.5)  Postoperative mechanical ventilation 16 (1.0) b b Surgical visit discharge disposition  Routine 1479 (93.2) 162 (11.0)  Adverse 107 (6.8) 21 (19.4) 29 (27.1)  Mean surgical visit total costs (SD) 49 779 (5125) 58 780 (9666) 55 158 (8882)  Mean surgical visit length of stay (SD) 8.3 (0.6) 9.8 (1.1) 9.0 (1.0) Patients N (%) 30-d RR (%) 90-d RR (%) d Total Patients (n) 1587 183 (11.5) 267 (16.8) Epilepsy diagnosisa NS NS  Generalized 76 (4.8) b 17 (22.4)  Complex Partial 904 (57.0) 111 (12.3) 148 (16.4)  Simple Partial 279 (17.6) 33 (11.7) 43 (15.5)  Other 392 (24.7) 42 (10.6) 64 (16.4) Presurgical diagnostic testing NS NS  VEEG 435 (27.2) 56 (12.8) 79 (18.2)  IEEG 568 (35.8) 66 (11.7) 88 (15.5) Type of surgerya NS NS  Hemispherectomy 66 (4.2) 15 (22.8) 24 (36.5)  Lobectomy 368 (23.2) 44 (11.9) 55 (14.9)  Amygdalohippocampectomy or partial lobectomy 1236 (77.9) 129 (10.4) 197 (16.0) Medical complications composite c c  0 1508 (95.0%) 157 (10.4) 236 (15.6)  1 79 (5%) 26 (33.0) 31 (39.0)  2 b b b  3+ b b b Medical complications individual  Cardiac b b b  Respiratory 24 (1.5) b 11 (45.4c)  Renal and urinary 14 (0.9) b b  Thromboembolic 32 (2.0) 18 (57.9) 18 (57.9) Surgical complications composite NS NS  0 1356 (85.4) 145 (10.7) 216 (16.0)  1 197 (12.4) 32 (16.2) 40 (20.1)  2 28 (1.8) b 11 (37.1)  3+ b b b Surgical complications individual NS NS  Stroke/hematoma 78 (4.9) b 14 (18.6)  Postoperative intracranial infections 34 (2.2) b b  Status epilepticus b b b  Hydrocephalus 11 (0.7) b b  VP shunt b b b  Blood transfusion 122 (7.7) 26 (21.4) 32 (26.5)  Postoperative mechanical ventilation 16 (1.0) b b Surgical visit discharge disposition  Routine 1479 (93.2) 162 (11.0)  Adverse 107 (6.8) 21 (19.4) 29 (27.1)  Mean surgical visit total costs (SD) 49 779 (5125) 58 780 (9666) 55 158 (8882)  Mean surgical visit length of stay (SD) 8.3 (0.6) 9.8 (1.1) 9.0 (1.0) IEEG, intracranial electroencephalography; NS, not significant; RR, readmission rate; VEEG, video electroencephalography. aExplanations provided to the reviewer about the apparent numerical discrepancies based on data present in the NRD for analysis. bLess than the HCUP reporting minimum of 11 cases cStatistical significance (P < .05) via Pearson chi-square test. dThe 30- and 90-d readmission cohorts were not mutually exclusive, as the 90-d readmission cohort includes any patient that was readmitted within 90 d, including readmissions within 30 d. Discharge disposition was dichotomized into the variable “adverse discharge” in which any discharge other than routine (transfer to skilled nursing, intermediate care facility, short-term acute-care hospital, home health care). View Large TABLE 3. The Univariate (Unadjusted) Effect of Epilepsy Diagnosis, Presurgical Diagnostic Testing, Type of Surgery, Medical Complications, and Surgical Complications on 30-d and 90-d Readmission Rates in 1587 who Underwent Surgery for Intractable Epilepsy, NRD 2013 Patients N (%) 30-d RR (%) 90-d RR (%) d Total Patients (n) 1587 183 (11.5) 267 (16.8) Epilepsy diagnosisa NS NS  Generalized 76 (4.8) b 17 (22.4)  Complex Partial 904 (57.0) 111 (12.3) 148 (16.4)  Simple Partial 279 (17.6) 33 (11.7) 43 (15.5)  Other 392 (24.7) 42 (10.6) 64 (16.4) Presurgical diagnostic testing NS NS  VEEG 435 (27.2) 56 (12.8) 79 (18.2)  IEEG 568 (35.8) 66 (11.7) 88 (15.5) Type of surgerya NS NS  Hemispherectomy 66 (4.2) 15 (22.8) 24 (36.5)  Lobectomy 368 (23.2) 44 (11.9) 55 (14.9)  Amygdalohippocampectomy or partial lobectomy 1236 (77.9) 129 (10.4) 197 (16.0) Medical complications composite c c  0 1508 (95.0%) 157 (10.4) 236 (15.6)  1 79 (5%) 26 (33.0) 31 (39.0)  2 b b b  3+ b b b Medical complications individual  Cardiac b b b  Respiratory 24 (1.5) b 11 (45.4c)  Renal and urinary 14 (0.9) b b  Thromboembolic 32 (2.0) 18 (57.9) 18 (57.9) Surgical complications composite NS NS  0 1356 (85.4) 145 (10.7) 216 (16.0)  1 197 (12.4) 32 (16.2) 40 (20.1)  2 28 (1.8) b 11 (37.1)  3+ b b b Surgical complications individual NS NS  Stroke/hematoma 78 (4.9) b 14 (18.6)  Postoperative intracranial infections 34 (2.2) b b  Status epilepticus b b b  Hydrocephalus 11 (0.7) b b  VP shunt b b b  Blood transfusion 122 (7.7) 26 (21.4) 32 (26.5)  Postoperative mechanical ventilation 16 (1.0) b b Surgical visit discharge disposition  Routine 1479 (93.2) 162 (11.0)  Adverse 107 (6.8) 21 (19.4) 29 (27.1)  Mean surgical visit total costs (SD) 49 779 (5125) 58 780 (9666) 55 158 (8882)  Mean surgical visit length of stay (SD) 8.3 (0.6) 9.8 (1.1) 9.0 (1.0) Patients N (%) 30-d RR (%) 90-d RR (%) d Total Patients (n) 1587 183 (11.5) 267 (16.8) Epilepsy diagnosisa NS NS  Generalized 76 (4.8) b 17 (22.4)  Complex Partial 904 (57.0) 111 (12.3) 148 (16.4)  Simple Partial 279 (17.6) 33 (11.7) 43 (15.5)  Other 392 (24.7) 42 (10.6) 64 (16.4) Presurgical diagnostic testing NS NS  VEEG 435 (27.2) 56 (12.8) 79 (18.2)  IEEG 568 (35.8) 66 (11.7) 88 (15.5) Type of surgerya NS NS  Hemispherectomy 66 (4.2) 15 (22.8) 24 (36.5)  Lobectomy 368 (23.2) 44 (11.9) 55 (14.9)  Amygdalohippocampectomy or partial lobectomy 1236 (77.9) 129 (10.4) 197 (16.0) Medical complications composite c c  0 1508 (95.0%) 157 (10.4) 236 (15.6)  1 79 (5%) 26 (33.0) 31 (39.0)  2 b b b  3+ b b b Medical complications individual  Cardiac b b b  Respiratory 24 (1.5) b 11 (45.4c)  Renal and urinary 14 (0.9) b b  Thromboembolic 32 (2.0) 18 (57.9) 18 (57.9) Surgical complications composite NS NS  0 1356 (85.4) 145 (10.7) 216 (16.0)  1 197 (12.4) 32 (16.2) 40 (20.1)  2 28 (1.8) b 11 (37.1)  3+ b b b Surgical complications individual NS NS  Stroke/hematoma 78 (4.9) b 14 (18.6)  Postoperative intracranial infections 34 (2.2) b b  Status epilepticus b b b  Hydrocephalus 11 (0.7) b b  VP shunt b b b  Blood transfusion 122 (7.7) 26 (21.4) 32 (26.5)  Postoperative mechanical ventilation 16 (1.0) b b Surgical visit discharge disposition  Routine 1479 (93.2) 162 (11.0)  Adverse 107 (6.8) 21 (19.4) 29 (27.1)  Mean surgical visit total costs (SD) 49 779 (5125) 58 780 (9666) 55 158 (8882)  Mean surgical visit length of stay (SD) 8.3 (0.6) 9.8 (1.1) 9.0 (1.0) IEEG, intracranial electroencephalography; NS, not significant; RR, readmission rate; VEEG, video electroencephalography. aExplanations provided to the reviewer about the apparent numerical discrepancies based on data present in the NRD for analysis. bLess than the HCUP reporting minimum of 11 cases cStatistical significance (P < .05) via Pearson chi-square test. dThe 30- and 90-d readmission cohorts were not mutually exclusive, as the 90-d readmission cohort includes any patient that was readmitted within 90 d, including readmissions within 30 d. Discharge disposition was dichotomized into the variable “adverse discharge” in which any discharge other than routine (transfer to skilled nursing, intermediate care facility, short-term acute-care hospital, home health care). View Large In multivariable analysis analyzing risk factors for readmission, we found that Medicare, but not Medicaid, patients were more likely to suffer 30-d readmission and 90-d readmission (Table 4). In subgroup analysis, Medicaid patients had a higher rate of 3 or more comorbidities (19.3% vs 13.2%, P = .03) than Medicare patients. Depression was significantly associated with increased likelihood of both 30-d (OR: 4.2, 95% CI: 1.9-9.4, P < .001) and 90-d readmission (OR: 3.7, 95% CI: 1.7-7.7, P = .001). Patients receiving hemispherectomy compared to other surgical approaches were more likely to be readmitted at both 30 d (OR: 4.5, 95% CI: 1.4-14.7, P = .014) and 90 d (OR: 5.0, 95% CI: 1.7-14.4, P = .003). In a subgroup analysis, hemispherectomy was associated with decreased rate of persistent epilepsy-related readmission, but increased rate of postoperative infection-related readmission. If a medical complication occurred during the index visit, a patient's likelihood of both 30-d (OR: 5.2, 95% CI: 2.1-13.2, P = .001) and 90-d readmission (OR: 4.3, 95% CI: 1.8-10.4, P = .012) increased significantly. TABLE 4. Multivariable Predictors of 30- and 90-d Readmission in 1587 Patients who Underwent Surgery for Intractable Epilepsy Variable 30-d readmission 90-d readmission Total Patients (n) OR (95% CI) P value OR (95% CI) P value Payer .020a .043a  Private Reference Reference  Medicare 4.0 (1.6-10.1) 2.7 (1.3-5.8)  Medicaid 2.3 (1.0-5.4) 1.4 (0.7-2.6)  Other 0.9 (0.2-4.5) 0.8 (0.3-2.2) Median annual household income .038a .006a  0-25th percentile Reference Reference  26th-50th percentile 0.6 (0.2-1.6) 0.7 (0.3-1.6)  51st-75th percentile 0.4 (0.3-0.9) 0.4 (0.2-0.8)  76th-100th percentile 1.3 (0.5-3.8) 1.4 (0.7-2.7) Bed size of hospital .001a –  Small Reference –  Medium 0.2 (0.08-0.6) –  Large 0.3 (0.2-0.6) – Pre-existing comorbidityb  Depression 4.2 (1.9-9.4) <.001a 3.7 (1.7-7.7) .001a Type of surgery .014a .003a  Other surgery Reference Reference  Hemispherectomy 4.5 (1.4-14.7) 5.0 (1.7-14.4) Medical complication scorec 5.2 (2.08-13.2) .001a 4.3 (1.8-10.4) .012a Variable 30-d readmission 90-d readmission Total Patients (n) OR (95% CI) P value OR (95% CI) P value Payer .020a .043a  Private Reference Reference  Medicare 4.0 (1.6-10.1) 2.7 (1.3-5.8)  Medicaid 2.3 (1.0-5.4) 1.4 (0.7-2.6)  Other 0.9 (0.2-4.5) 0.8 (0.3-2.2) Median annual household income .038a .006a  0-25th percentile Reference Reference  26th-50th percentile 0.6 (0.2-1.6) 0.7 (0.3-1.6)  51st-75th percentile 0.4 (0.3-0.9) 0.4 (0.2-0.8)  76th-100th percentile 1.3 (0.5-3.8) 1.4 (0.7-2.7) Bed size of hospital .001a –  Small Reference –  Medium 0.2 (0.08-0.6) –  Large 0.3 (0.2-0.6) – Pre-existing comorbidityb  Depression 4.2 (1.9-9.4) <.001a 3.7 (1.7-7.7) .001a Type of surgery .014a .003a  Other surgery Reference Reference  Hemispherectomy 4.5 (1.4-14.7) 5.0 (1.7-14.4) Medical complication scorec 5.2 (2.08-13.2) .001a 4.3 (1.8-10.4) .012a aStatistical significance (P < .05) via Pearson chi-square test. bOdds ratio represents increased likelihood of readmission per each additional comorbidity. cOdds ratio represents increased likelihood of readmission per each additional complication. dIndicates variables not included in the final multivariable model. View Large TABLE 4. Multivariable Predictors of 30- and 90-d Readmission in 1587 Patients who Underwent Surgery for Intractable Epilepsy Variable 30-d readmission 90-d readmission Total Patients (n) OR (95% CI) P value OR (95% CI) P value Payer .020a .043a  Private Reference Reference  Medicare 4.0 (1.6-10.1) 2.7 (1.3-5.8)  Medicaid 2.3 (1.0-5.4) 1.4 (0.7-2.6)  Other 0.9 (0.2-4.5) 0.8 (0.3-2.2) Median annual household income .038a .006a  0-25th percentile Reference Reference  26th-50th percentile 0.6 (0.2-1.6) 0.7 (0.3-1.6)  51st-75th percentile 0.4 (0.3-0.9) 0.4 (0.2-0.8)  76th-100th percentile 1.3 (0.5-3.8) 1.4 (0.7-2.7) Bed size of hospital .001a –  Small Reference –  Medium 0.2 (0.08-0.6) –  Large 0.3 (0.2-0.6) – Pre-existing comorbidityb  Depression 4.2 (1.9-9.4) <.001a 3.7 (1.7-7.7) .001a Type of surgery .014a .003a  Other surgery Reference Reference  Hemispherectomy 4.5 (1.4-14.7) 5.0 (1.7-14.4) Medical complication scorec 5.2 (2.08-13.2) .001a 4.3 (1.8-10.4) .012a Variable 30-d readmission 90-d readmission Total Patients (n) OR (95% CI) P value OR (95% CI) P value Payer .020a .043a  Private Reference Reference  Medicare 4.0 (1.6-10.1) 2.7 (1.3-5.8)  Medicaid 2.3 (1.0-5.4) 1.4 (0.7-2.6)  Other 0.9 (0.2-4.5) 0.8 (0.3-2.2) Median annual household income .038a .006a  0-25th percentile Reference Reference  26th-50th percentile 0.6 (0.2-1.6) 0.7 (0.3-1.6)  51st-75th percentile 0.4 (0.3-0.9) 0.4 (0.2-0.8)  76th-100th percentile 1.3 (0.5-3.8) 1.4 (0.7-2.7) Bed size of hospital .001a –  Small Reference –  Medium 0.2 (0.08-0.6) –  Large 0.3 (0.2-0.6) – Pre-existing comorbidityb  Depression 4.2 (1.9-9.4) <.001a 3.7 (1.7-7.7) .001a Type of surgery .014a .003a  Other surgery Reference Reference  Hemispherectomy 4.5 (1.4-14.7) 5.0 (1.7-14.4) Medical complication scorec 5.2 (2.08-13.2) .001a 4.3 (1.8-10.4) .012a aStatistical significance (P < .05) via Pearson chi-square test. bOdds ratio represents increased likelihood of readmission per each additional comorbidity. cOdds ratio represents increased likelihood of readmission per each additional complication. dIndicates variables not included in the final multivariable model. View Large DISCUSSION Cause of Readmissions The major cause of readmissions is persistent epilepsy and VEEG monitoring. It is well known that epilepsy surgery success rates range from 50% to 80%.14,15 Hence, it is not surprising that patients might require readmission for additional monitoring or management of persistent seizures. Similar findings in a smaller study were reported by Wilson et al16 who examined the records of 100 consecutive epilepsy patients undergoing anterior temporal lobectomy. Twenty-one (21%) of patients required readmission.16 The most common cause of readmission were psychiatric reasons (anxiety, depression; 53%), and the second most common cause was persistent epilepsy (28%).16 In our multivariable analysis of risk factors, depression was the only pre-existing comorbidity that was significantly associated with increased likelihood of readmission. Similarly, Kanner et al,17 in a retrospective review of 95 TLE patients, demonstrated that a life-time history of depression preoperatively was a predictor of worse postsurgical seizure outcome, specifically, less likely to be seizure free. Patient demographics and hospital type were found to influence the rate of readmissions. We found that Medicare but not Medicaid or privately insured were significantly more likely to be readmitted within both 30 and 90 d. Medicare patients are typically much older than Medicaid patients and may be more likely to have medical comorbidities. Upon further subgroup analysis, we found, however, that Medicaid patients had a higher rate of 3 or more comorbidities (19.3% vs 13.2%, P = .03) than Medicare patients. Medicare patients are less likely to have socioeconomic barriers to medical care compared to Medicaid patients. However, this could not be further evaluated in our dataset. Aggressive maintenance of antiepileptic medications after surgery might reduce readmissions for persistent epilepsy or monitoring. Likewise, scheduled visits with psychiatrists, therapists, and social workers might prevent psychiatric issues from arising as frequently after surgery. The categorical variable for median annual household income was statistically significant, but we were unable to identify clinically significant trends. For example, the analysis found that patients in the 51th to 75th percentile were less likely to be readmitted compared to patients in the 0 to 25th percentile. However, this trend did not apply to the 76th to 100th percentile in terms of income. Smaller hospitals (fewer beds) were significantly more likely to have 30-d readmissions compared to larger hospitals. Larger hospitals are known to have more resources, including advanced technology, and are more likely to be teaching facilities. Furthermore, smaller hospitals generally represent lower volume centers with less experience and are more likely to reach capacity, potentially prompting premature discharges. However, the precise characteristics of smaller hospitals that lead to more readmissions were not able to be queried from our database. A similar argument can be made for hemispherectomy surgery. These surgeries are often performed on young children and are long, tedious surgeries, which often require blood transfusions. Vadera et al11 utilized the NIS to identify 304 pediatric patients who underwent hemispherectomy between 1988 and 2010. They found that 56% of patients encountered a complication (42% related to surgery).11 In the pediatric population, undergoing such a long, complicated surgery prone to blood loss likely necessitates a transfusion as part of the procedure, and it is unclear if this is truly a preventable complication. The length and complexity of the surgery likely also increased the risk of postoperative infection, as borne out by our data. Frequency and Timing of Readmissions Prior studies of readmission for epilepsy have been examined as a subgroup of all craniotomies and not thoroughly analyzed. Moghavem et al18 investigated unplanned 30-d readmissions utilizing the SID for California, Florida, and New York. They identified 43 356 patients, of which 2.80% underwent cranial neurosurgery for seizure.18 Thirty-day readmissions were 13.89% for the seizure cohort.18 Significant predictors for readmission included male gender (OR 1.74; 95% CI 1.17-2.60) and initial admission via emergency department (OR 2.22; 95% CI 1.45-3.43).18 Taylor et al19 investigated unplanned readmissions following neurosurgery utilizing a statewide database for New York. One hundred fifty-four patients had epilepsy surgery, and the 30-d readmission rate was 7.14%.19 The overall most common reasons for unplanned readmission included infection (29.52%), medical complications (19.22%), stroke (6.09%), and venous thromboembolism (5.71%). Postoperative seizures accounted for less than 5%.19 Our study presents quite different conclusions for top reasons for readmission, but postoperative seizures accounted for approximately 5.88% of patients, similar to the aforementioned results. It is possible that results from our study vary due to the nationwide nature of the population as well as aspects of coding database queries. Our readmission rates for 30 and 90 d (11.5% and 16.8%, respectively) are of similar magnitude to previously reported studies: 7.14% to 21%. Almost two-thirds of the readmissions occurred within 30 d. Rates vary, however, due to study design including single-institution,16 single state,19 or select states18 from the SID. Our study adds data for readmissions outside the initial 30-d window and tracks patients longitudinally, perhaps more accurately reflecting readmissions. Reasons for admission are not qualitatively different at 30 vs 90 d, and this distinction will not alter prevention strategies. Predictors of 90-d readmission, in addition to Medicare payer status and depression, included increasing number of comorbidities and medical complications. Although pre-existing comorbidities were only significant in univariate analysis, each additional comorbidity and/or complication increased the likelihood of readmission. The medical complication score represents the OR for increased likelihood of readmission for each additional comorbidity or complication within 30 and 90 d. The most common medical complications likely resulted from mechanical ventilation and immobility or deep venous thrombosis (DVT). Therefore, additional attention is deserved for preventing these medical complications, ie, DVT prophylaxis, early ambulation, and incentive spirometry.20 Our results indicate that medical, rather than surgical, complications are a stronger contributor to the likelihood of readmission. However, no patient had more than one reported medical complication during index hospitalization. Furthermore, upon descriptive analysis, patients with surgical complications appeared to have higher rates of 30-d (16.2% vs 10.7%) and 90-d (37.1% vs 16%) readmission. Thus, these comparisons likely failed to reach statistical significance due to lack of statistical power resulting from the relatively low number of surgical complications. Limitations The NRD maintains patient-linkage numbers to track patients longitudinally, which is a limitation of other databases. The NRD does have limitations including its dependence on ICD-9-CM coding, with varying sensitivities and specificities, and small percentage of missing data for some variables.21-27 It is possible that additional causes of readmission listed in secondary diagnosis fields were missed. Since the NRD was only available for 1 calendar year when this study was conducted, our follow-up time was limited. In order to have a substantial population size for analysis as well as a reasonable follow-up period, we included surgeries performed during the first 9 mo of the year and left the remaining 3 mo (90 d) as a window to capture readmissions. Since the diagnoses in the dataset are compiled by medical coders, not clinicians, the diagnosis of stroke and hematoma is likely overestimated since most routine postoperative imaging will show some degree of blood and/or clinically insignificant ischemia. However, if this complication was not specifically entered into the patient chart by the clinical team, it is unlikely to have been coded or billed. Additionally, the NRD, like other databases, lacks functional status, clinical exams, radiology impressions, surgeon details, and surgical technique. Quality of life is not recorded or tracked in the NRD. Certain univariate variables could not be considered in the multivariate analysis due to small sample size. CONCLUSION Minimizing factors that contribute to readmission in various patient populations and procedures becomes important for patient care, resource utilization, and physician reimbursement. In epilepsy surgery, the majority of readmissions occurred within 30 d, and the causes of readmission were similar within 90 d: persistent epilepsy, postoperative infection, and postoperative CNS complications. Awareness of the reasons for readmission is important for patient counseling and surgical decision-making. In epilepsy surgery, readmissions might be reduced with optimization of pre-existing comorbidities particularly in the elderly, minimization of complications, close follow-up with hemispherectomy patients, and adequate control of epileptic seizures during the index visit. However, persistent seizures after a procedure with a known efficacy of only 50% to 80% might not be avoidable. Disclosures Dr Arnold has commercial relationships with Z-Plasty (stock, stock options, or other ownership interest); Medtronic Sofamor Danek (salary and any payment for services not otherwise identified as salary such as consulting, fees, honoraria, paid authorship, or other payments for services); Stryker Spine (sponsored or reimbursed travel, salary and any payment for services no otherwise identified as salary such as consulting, fees, honoraria, paid authorship, or other payments for services); AO Spine North America (sponsored or reimbursed travel); Invivo (salary and any payment for series not otherwise identified as salary such as consulting, fees, honoraria, paid authorship, or other payments for services). Dr Schwartz has stock options in Visionsense. The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article. Notes The content of this article was presented at the Neurosurgical Society of America meeting on April 3, 2017 in Jacksonville, Florida, as an Oral Presentation. REFERENCES 1. Kim Y , Gani F , Lucas DJ et al. Early versus late readmission after surgery among patients with employer-provided health insurance . Ann Surg . 2015 ; 262 ( 3 ): 502 – 511 . Google Scholar CrossRef Search ADS PubMed 2. Koubeissi MZ , Puwanant A , Jehi L , Alshekhlee A . In-hospital complications of epilepsy surgery: a six-year nationwide experience . Br J Neurosurg . 2009 ; 23 ( 5 ): 524 – 529 . Google Scholar CrossRef Search ADS PubMed 3. McClelland S 3rd , Guo H , Okuyemi KS . Racial disparities in the surgical management of intractable temporal lobe epilepsy in the United States . Arch Neurol . 2010 ; 67 ( 5 ): 577 – 583 . Google Scholar CrossRef Search ADS PubMed 4. Vale FL , Reintjes S , Garcia HG . Complications after mesial temporal lobe surgery via inferiortemporal gyrus approach . Neurosurg Focus . 2013 ; 34 ( 6 ): E2 . Google Scholar CrossRef Search ADS PubMed 5. NRD Overview . Healthcare Cost and Utilization Project (HCUP) . 2015 . Available at: www.hcup-us.ahrq.gov/nrdoverview.jsp. Accessed June 1, 2016 . 6. HCUP . Clinical Classifications Software (CCS) for ICD-9-CM . 2011 . Available at: www.hcup-us.ahrq.gov/toolssoftware/ccs/ccs.jsp. Accessed June 1, 2016 . 7. Schiltz NK , Koroukian SM , Lhatoo SD , Kaiboriboon K . Temporal trends in pre-surgical evaluations and epilepsy surgery in the U.S. from 1998–2009 . Epilepsy Res . 2013 ; 103 ( 2-3 ): 270 – 278 . Google Scholar CrossRef Search ADS PubMed 8. McClelland S 3rd , Curran CC , Davey CS , Okuyemi KS . Intractable pediatric temporal lobe epilepsy in the United States: examination of race, age, sex, and insurance status as factors predicting receipt of resective treatment . J Neurosurg . 2007 ; 107 ( 6 Suppl ): 469 – 473 . Google Scholar PubMed 9. HCUP . Comorbidity Software . 2012 . Available at: www.hcup-us.ahrq.gov/toolssoftware/comorbidity/comorbidity.jsp. Accessed June 1, 2016 . 10. Elixhauser A , Steiner C , Harris DR , Coffey RM . Comorbidity measures for use with administrative data . Med Care . 1998 ; 36 ( 1 ): 8 – 27 . Google Scholar CrossRef Search ADS PubMed 11. Vadera S , Griffith SD , Rosenbaum BP et al. National trends and In-hospital complication rates in more than 1600 hemispherectomies from 1988 to 2010 . Neurosurgery . 2015 ; 77 ( 2 ): 185 – 191 . Google Scholar CrossRef Search ADS PubMed 12. Sharma M , Sonig A , Ambekar S , Nanda A . Discharge dispositions, complications, and costs of hospitalization in spinal cord tumor surgery: analysis of data from the United States Nationwide Inpatient Sample, 2003–2010 . J Neurosurg Spine . 2014 ; 20 ( 2 ): 125 – 141 . Google Scholar CrossRef Search ADS PubMed 13. HCUP . Cost-to-Charge Ratio Files . 2012 . Available at: www.hcup-us.ahrq.gov/db/state/costtocharge.jsp. Accessed June 1, 2016 . 14. Spencer S , Huh L . Outcomes of epilepsy surgery in adults and children . Lancet Neurol . 2008 ; 7 ( 6 ): 525 – 537 . Google Scholar CrossRef Search ADS PubMed 15. Wiebe S , Blume WT , Girvin JP , Eliasziw M , Effectiveness, Efficiency of Surgery for Temporal Lobe Epilepsy Study Group. A randomized, controlled trial of surgery for temporal-lobe epilepsy . N Engl J Med . 2001 ; 345 ( 5 ): 311 – 318 . Google Scholar CrossRef Search ADS PubMed 16. Wilson SJ , Kincade P , Saling MM , Bladin PF . Patient readmission and support utilization following anterior temporal lobectomy . Seizure . 1999 ; 8 ( 1 ): 20 – 25 . Google Scholar CrossRef Search ADS PubMed 17. Kanner AM BR , Smith MC , Balabanov AJ , Frey M . Does a life-time history of depression predict a worse post-surgical seizure outcome following temporal lobectomy? Ann Neurol . 2006 ; 60 ( Suppl 3 ): S19 . 18. Moghavem N , Morrison D , Ratliff JK , Hernandez-Boussard T . Cranial neurosurgical 30-day readmissions by clinical indication . J Neurosurg . 2015 ; 123 ( 1 ): 189 – 197 . Google Scholar CrossRef Search ADS PubMed 19. Taylor BE , Youngerman BE , Goldstein H et al. Causes and timing of unplanned early readmission after neurosurgery . Neurosurgery . 2016 ; 79 ( 3 ): 356 – 369 . Google Scholar CrossRef Search ADS PubMed 20. Pavon JM , Adam SS , Razouki ZA et al. Effectiveness of intermittent pneumatic compression devices for venous thromboembolism prophylaxis in High-Risk surgical patients: A systematic review . J Arthroplasty . 2016 ; 31 ( 2 ): 524 – 532 . Google Scholar CrossRef Search ADS PubMed 21. Gologorsky Y , Knightly JJ , Chi JH , Groff MW . The Nationwide Inpatient Sample database does not accurately reflect surgical indications for fusion . J Neurosurg Spine . 2014 ; 21 ( 6 ): 984 – 993 . Google Scholar CrossRef Search ADS PubMed 22. Jalai CM , Worley N , Marascalchi BJ et al. The impact of advanced age on peri-operative outcomes in the surgical treatment of cervical spondylotic myelopathy . Spine . 2016 ; 41 ( 3 ): E139 – E147 . Google Scholar CrossRef Search ADS PubMed 23. King JT Jr , Abbed KM , Gould GC , Benzel EC , Ghogawala Z . Cervical spine reoperation rates and hospital resource utilization after initial surgery for degenerative cervical spine disease in 12 338 patients in Washington State . Neurosurgery . 2009 ; 65 ( 6 ): 1011 – 1023 . Google Scholar CrossRef Search ADS PubMed 24. Marquez-Lara A , Nandyala SV , Fineberg SJ , Singh K . Current trends in demographics, practice, and In-Hospital outcomes in cervical spine surgery . Spine . 2014 ; 39 ( 6 ): 476 – 481 . Google Scholar CrossRef Search ADS PubMed 25. Patil CG , Santarelli J , Lad SP , Ho C , Tian W , Boakye M . Inpatient complications, mortality, and discharge disposition after surgical correction of idiopathic scoliosis: a national perspective . Spine J . 2008 ; 8 ( 6 ): 904 – 910 . Google Scholar CrossRef Search ADS PubMed 26. Shamji MF , Cook C , Pietrobon R , Tackett S , Brown C , Isaacs RE . Impact of surgical approach on complications and resource utilization of cervical spine fusion: a nationwide perspective to the surgical treatment of diffuse cervical spondylosis . Spine J . 2009 ; 9 ( 1 ): 31 – 38 . Google Scholar CrossRef Search ADS PubMed 27. Wang MC , Chan L , Maiman DJ , Kreuter W , Deyo RA . Complications and mortality associated with cervical spine surgery for degenerative disease in the United States . Spine . 2007 ; 32 ( 3 ): 342 – 347 . Google Scholar CrossRef Search ADS PubMed Operative Neurosurgery Speaks! Audio abstracts available for this article at www.operativeneurosurgery-online.com. COMMENTS The authors have provided a valuable review of the NRD for data on patients undergoing open epilepsy surgery. Their findings are very much in line with experience, literature, and common sense. Failure to control epilepsy and surgical complications, especially in more complex operations were the most likely reasons for readmission. Another interesting finding is that patients covered by Medicare or coming from a lower socioeconomic background were more likely to be readmitted. This may be a proxy for more severe epilepsy, or for a lack of home resources for postoperative care. The unique predictor of 30-day readmission was a small treating hospital. It is not clear why this would be the case, but may be an indicator of a smaller epilepsy program. This data may be of use to epilepsy surgery centers as it may mark those patients who should be further supported with home healthcare services, or should be seen in the outpatient clinic soon after surgery for evaluation. Richard W. Byrne Chicago, Illinois The present study investigates rates, causes, and predictors of 30- and 90-day readmissions after epilepsy surgery, by using the Nationwide Readmission Database on a significantly large sample of readmitted cases. The authors should be congratulated for their effort, since the presented results may represent a useful basis to measure the impact of readmissions on additional costs for the healthcare system and to direct future actions aimed at limiting this phenomenon. Massimo Cossu Milano, Italy In their publication the authors report on the United States readmission rate after epilepsy surgery between January and September 2013 (roughly, an 8-month period). They queried a nationwide US database for readmission and used the ICD-9CM code to identify patients who were readmitted after being treated for intractable epilepsy (with surgery). Patients had undergone a variety of procedures (hemispherectomy, lobectomy, partial lobectomy, or amygdalohippocampectomy). A plethora of predictors were statistically analyzed to characterize readmission reasons. A total of 1587 patients were identified that fulfilled the mentioned criteria. The 30-day readmission rate was 11.5% and the 90-day readmission rate was 16.8%. Reasons for this were typically persistent epilepsy, postoperative complications, or severity of epilepsy (indicated by Video EEG). This could be - medically and politically - a very interesting paper. It probably sheds some light on the medical care given to patients with intractable epilepsy (and more so in other chronic conditions). In our experience a typical readmission happens when a patient was prematurely discharged; either a medication was not fully titrated (up to a therapeutic level) or a complication was not excluded/fully treated, or the patient was simply not fit enough to go home and take care of himself. Of course, another reason could be a high complication rate at a hospital that is not usually dealing with complicated procedures. Nevertheless, the pressure on neurosurgical units these days is high and often now financially driven. Decision making is not solely based on medical facts but often driven by financial considerations. We as physicians might draw conclusions out of information like this and try to raise pressure on politicians and hospital authorities against pure financial towards more medical reasoning when discharging a patient who otherwise might benefit from some more days under our care. Volker A. Coenen Freiburg im Breisgau, Germany Operative Neurosurgery Speaks (Audio Abstracts) Listen to audio translations of this paper's abstract into select languages by choosing from one of the selections below. Chinese: Zuowei Wang, MD. Department of Neurosurgery Beijing Hospital Beijing, China Chinese: Zuowei Wang, MD. Department of Neurosurgery Beijing Hospital Beijing, China Close French: Johan Pallud, MD, PhD. Department of Neurosurgery Medical School of Paris Descartes University Paris, France French: Johan Pallud, MD, PhD. Department of Neurosurgery Medical School of Paris Descartes University Paris, France Close English: William W. Ashley, MD, PhD, MBA. Department of Neurological Surgery Sinai Hospital and LifeBridge Health System Baltimore, Maryland English: William W. Ashley, MD, PhD, MBA. Department of Neurological Surgery Sinai Hospital and LifeBridge Health System Baltimore, Maryland Close Italian: Marco Cenzato, MD. Neurosurgical Department Ospedale Maggiore Niguarda Ca 'Grande Milan, Italy Italian: Marco Cenzato, MD. Neurosurgical Department Ospedale Maggiore Niguarda Ca 'Grande Milan, Italy Close Spanish: Alejandro Enriquez-Marulanda, MD. Department of Neurosurgery Beth Israel Deaconess Medical Center Boston, Massachusetts Spanish: Alejandro Enriquez-Marulanda, MD. Department of Neurosurgery Beth Israel Deaconess Medical Center Boston, Massachusetts Close Portuguese: José Luís Alves, MD. Department of Neurosurgery Centro Hospitalar e Universitário de Coimbra Coimbra, Portugal Portuguese: José Luís Alves, MD. Department of Neurosurgery Centro Hospitalar e Universitário de Coimbra Coimbra, Portugal Close Japanese: Masaru Aoyagi, MD. Department of Neurosurgery Tokyo Medical and Dental University Tokyo, Japan Japanese: Masaru Aoyagi, MD. Department of Neurosurgery Tokyo Medical and Dental University Tokyo, Japan Close Korean: Hye Ran Park, MD. Department of Neurosurgery Soonchunhyang University Seoul Hospital Seoul, Republic of Korea Korean: Hye Ran Park, MD. Department of Neurosurgery Soonchunhyang University Seoul Hospital Seoul, Republic of Korea Close Russian: Vsevolod Shurhkhay, MD. Burdenko Institute of Neurosurgery Moscow, Russian Federation Russian: Vsevolod Shurhkhay, MD. Burdenko Institute of Neurosurgery Moscow, Russian Federation Close Greek: Marios Themistocleous, MD. Department of Neurosurgery Aghia Sophia Children's Hospital Athens, Greece Greek: Marios Themistocleous, MD. Department of Neurosurgery Aghia Sophia Children's Hospital Athens, Greece Close Copyright © 2018 by the Congress of Neurological Surgeons This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Operative Neurosurgery Oxford University Press

Readmission Following Surgical Resection for Intractable Epilepsy: Nationwide Rates, Causes, Predictors, and Outcomes

Loading next page...
 
/lp/ou_press/readmission-following-surgical-resection-for-intractable-epilepsy-wd3FmsJWfl
Publisher
Congress of Neurological Surgeons
Copyright
Copyright © 2018 by the Congress of Neurological Surgeons
ISSN
2332-4252
eISSN
2332-4260
D.O.I.
10.1093/ons/opy099
Publisher site
See Article on Publisher Site

Abstract

Abstract BACKGROUND Hospital readmissions can be detrimental to patients and may interfere with the potential benefits of the therapeutic procedure. Government agencies have begun to focus on reducing readmissions; however, the etiology of readmissions is lacking. OBJECTIVE To report the national rates, risk factors, and outcomes associated with 30- and 90-d readmissions following surgery for intractable epilepsy. METHODS We queried the Nationwide Readmissions Database from January to September 2013 using International Classification of Diseases, Ninth Edition, Clinical Modification codes to identify all patients with intractable epilepsy, who underwent hemispherectomy (01.52), brain lobectomy (01.53), amydalohippocampectomy, or partial lobectomy (01.59). Predictor variables included epilepsy type, presurgical diagnostic testing, surgery type, medical complications, surgical complications, and discharge disposition. RESULTS In 1587 patients, the 30- and 90-d readmission rates were 11.5% and 16.8%, respectively. The most common reasons for readmission were persistent epilepsy, video electroencephalography monitoring, postoperative infection, and postoperative central nervous system complication. In multivariable analysis, risk factors associated with both 30- and 90-d readmission were Medicare payer status, lowest quartile of median income, depression, hemispherectomy, and postoperative complications (P < .05). The only unique predictor of 30-d readmission was small bedsize hospital (P = .001). Readmissions within 30 d were associated with longer length of stay (6.8 vs 5.8 d), greater costs ($18 660 vs $15 515), and increased adverse discharges (26.4% vs 21.8%). CONCLUSION Following epilepsy surgery, most readmissions that occurred within 30 d can be attributed to management of persistent epilepsy and predicted by Medicare payer status, depression, and complications. These data can assist the clinician in preventing readmissions and assist policy makers determine which admissions are potentially avoidable. Complications, Epilepsy, Readmissions, Seizure ABBREVIATIONS ABBREVIATIONS AHC amygdalohippocampectomy AHRQ Agency for Healthcare Research and Quality CI confidence interval DVT deep venous thrombosis HCUP Healthcare Cost and Utilization Project ICD-9-CM International Classification of Diseases, Ninth Edition, Clinical Modification IEEG intracranial electroencephalography NIS Nationwide Inpatient Sample NRD Nationwide Readmissions Database OR odds ratio SD standard deviation SID State Inpatient Databases VEEG video electroencephalography Hospital readmissions rate is an important metric for evaluating quality of patient care, and the advent of the Patient Protection and Affordable Care Act has required agencies such as the Centers for Medicare and Medicaid Services to penalize hospitals with excessive 30-d readmission rates. Furthermore, readmissions after a longer period (eg, 90 d) may reveal unique problems. Kim et al1 found that 30-d readmission indices underestimated true readmission rates. A few studies have reported readmissions in patients undergoing surgery for intractable epilepsy using the Nationwide Inpatient Sample (NIS), which can evaluate in-hospital outcomes.2,3 Furthermore, other studies are limited to a handful of single-center retrospective cohorts with small sample sizes.4 In 2013, under the Healthcare Cost and Utilization Project (HCUP), the Agency for Healthcare Research and Quality (AHRQ) released the Nationwide Readmissions Database (NRD), which combines the strengths of the NIS and the State Inpatient Databases (SID) to enable researchers to perform analyses at the national level. Thus, we sought to report the national rates, risk factors, reasons, and outcomes associated with 30- and 90-d readmission. METHODS Data Source The NRD was released in 2013 as a new addition to the HCUP family and was designed to enable researchers to effectively analyze readmissions.1 The NRD contains approximately 36 million weighted discharge records5 and is equipped with the strengths of NIS, namely sample size, with the addition of data elements from the SID, including the patient-linkage number, days-to-event variable, and hospital length of stay. The AHRQ’s Clinical Classification Software is a helpful tool that collapses the multitude of International Classification of Diseases, Ninth Edition, Clinical Modification (ICD-9-CM) codes into clinically meaningful categories.6 The required HCUP Data Use Agreement Training prohibits users from reporting any cell size representing less than 11 patients.5 Neither Institutional Review Board (IRB) approval nor patient consent was required. Study Design The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline checklist for retrospective cohort study was utilized. To define the index visit, we queried the NRD to identify all patients in the 2013 NRD who underwent epilepsy surgery between January and September. We excluded the last 3 mo to enable the evaluation of 90-d readmissions. Participants We searched all NRD diagnosis fields to identify patients with intractable epilepsy-related ICD-9-CM diagnoses (345.x1, exclude 345.2, 345.3, 345.61, 345.71).7 Among this cohort, only patients with ICD-9-CM procedure codes for epilepsy surgery (01.52, hemispherectomy; 01.53, brain lobectomy; 01.59, partial brain lobectomy or amygdalohippocampectomy [AHC]) were included.3,7,8 Setting A patient's first readmission following surgery was considered a readmission, but all preadmissions and subsequent readmissions were excluded. The NRD treats same-day readmissions and transfers as a single discharge record. Predictor Variables We abstracted and defined variables pertaining to patient demographics, pre-existing comorbidities, admission characteristics, hospital characteristics, presurgical diagnostic testing, and type of procedure (hemispherectomy, lobectomy, AHC, partial lobectomy/lesionectomy). The patient demographic variables included categorical age, gender, payer status, and median household income. Pre-existing comorbidities were scored using the Elixhauser Comorbidity Index as computed by AHRQ.9,10 Admission characteristics included admission type and day of admission. Hospital characteristics included bedsize, hospital type, and hospital location. We used ICD-9-CM codes to define more specific types of epilepsy, including generalized epilepsy (345.01, 345.11), localization-related complex partial epilepsy (345.41), localization-related simple partial epilepsy (345.51), and other intractable epilepsy (345.81, 345.91). The presurgical diagnostic tests evaluated were video electroencephalography (VEEG; 89.19) and intracranial electroencephalography (IEEG; 02.93). Outcome Variables The primary outcomes of interest were hospital readmission within 30 and 90 d. However, we defined additional variables for complications, length of stay, total hospital costs, discharge disposition, and in-hospital mortality. We defined surgical complication using secondary ICD-9-CM codes including stroke and hematoma (432.1, 997.00, 997.01, 997.02, 997.09, 998.11, 998.12), postoperative intracranial infections (320.0, 320.7, 320.81-320.9, 322.9, 998.51, 998.59), status epilepticus (345.3), hydrocephalus (331.3, 331.4), ventriculoperitoneal shunt placement (02.34), blood transfusion (99.01-99.09), postoperative mechanical ventilation (96.70-96.79), and meningitis (320-322, 326).11 Common in-hospital complication variables included cardiac (997.1, 410), respiratory (518.4, 518.5, 518.81-518.84), renal/urinary (584 and 997.5), and thromboembolic (415, 415.11-415.19, 451.0-451.9).12 Discharge disposition was dichotomized into “routine” and “adverse discharge,” meaning any discharge other than routine (transfer to skilled nursing, intermediate care facility, short-term acute-care hospital, or home health care).5 The NRD contains a total hospital charges variable; however, HCUP provides users with cost-to-charge ratio files to convert “charges” to “cost,” which reflects how much hospitals actually received in payment.13 Statistical Methods The SPSS v.24 statistical software (IBM Inc, Armonk, New York) was used for analyses (alpha 0.05). In SPSS, we used the complex samples function, which requires the stratum, cluster, and discharge weights to produce national estimates. Subgroup analysis was performed to compare causes of admission by type of surgery. Risk factors for readmission were identified using univariate and multivariable analysis (Chi-square test for categorical variables, independent samples t-test for continuous variables). Due to limited sample size for individual comorbidity and complication groups, these variables were treated as composite variables in analyses. Variables with statistical significance in the univariate analysis and/or strong clinical justification were entered in multivariable logistic regression model. RESULTS Participants A total of 1587 patients who underwent elective AHC/partial lobectomy (77.9%), lobectomy (23.2%), or hemispherectomy (4.2%) for intractable epilepsy were identified in the NRD between January and September of 2013. The intractable epilepsy types included complex partial (57.0%), simple partial (17.6%), generalized (4.8%), and other (24.7%). The rate of IEEG and VEEG presurgical diagnostic testing during the index hospitalization was 35.8% and 27.4%, respectively. Descriptive Data The mean age of the surgical cohort (±standard deviation [SD]) was 29.9 ± 1.4 yr, 49.2% were female, and 56.6% had at least 1 pre-existing comorbidity. The most common comorbidities were fluid/electrolyte disorder (14.0%), paralysis (12.6%), depression (11.9%), obesity (9.6%), and hypothyroidism (6.1%). The vast majority of surgeries were performed in teaching hospitals, so this predictor variable was not considered in further analyses. A total of 456 surgical complications and 78 medical complications were reported during the index surgical visits, but fewer than 11 patients suffered in-hospital mortality. The most common surgical complications were blood transfusion, stroke or hematoma, and intracranial infection (Table 1). Furthermore, the most common medical complications were thromboembolic and respiratory (Table 1). The mean cost of a surgical visit was $49 779, and the average length of stay was 8 ± 0.6 d. TABLE 1. Hospitalization-Associated Outcomes of 1587 Patients who Underwent Surgery for Intractable Epilepsy, NRD 2013 Variable Total patients (n) 1587 Surgical (index) visit outcomes  Mean LOS in days (SD) 8.3 (0.6)  Mean total costs, $USD (SD) $49 779 (5125)  Medical complications (n) 78  Surgical complications (n) 456  Adverse discharge disposition (%) 15.6 30-d reasons for readmission, n 183  Diagnoses (ICD-9-CM diagnosis codes), n   Epilepsy (345.3, 345.41, 345.51, 345.90) 58   Postoperative infection (998.59) 28   Postoperative CNS complication (997.01) 18   Headache (784.0) 14   Other (individual codes <11 cases) 65  Procedures (ICD-9-CM procedure codes), n   V-EEG monitoring (89.19) 26   Venous catherization (38.93) 25   Other (individual codes <11 cases) 14 30-d readmission visit outcomes, n 183  Mean length of stay, days (SD) 6.8 (0.7)  Mean total costs, $USD (SD) $18 660 (2921)  Adverse discharge disposition (%) 26.4 90-d reasons for readmission, n 267  Diagnoses (ICD-9-CM diagnosis codes), n   Epilepsy (345.41, 345.50, 345.51, 345.60, 345.81, 345.90, 345.91) 103   Postoperative infection (998.59) 34   Postoperative CNS complication (997.01) 18   Headache (784.0 346.90) 16  Procedures (ICD-9-CM procedure codes), n   V-EEG monitoring (89.19) 47   Venous catherization (38.93) 25   Other (individual codes < 11 cases) 75 90-d readmission visit outcomes, n  Mean length of stay, days (SD) 5.8 (0.3)  Mean total costs, $USD (SD) $15 515 (1360)  Adverse discharge disposition (%) 21.8 Variable Total patients (n) 1587 Surgical (index) visit outcomes  Mean LOS in days (SD) 8.3 (0.6)  Mean total costs, $USD (SD) $49 779 (5125)  Medical complications (n) 78  Surgical complications (n) 456  Adverse discharge disposition (%) 15.6 30-d reasons for readmission, n 183  Diagnoses (ICD-9-CM diagnosis codes), n   Epilepsy (345.3, 345.41, 345.51, 345.90) 58   Postoperative infection (998.59) 28   Postoperative CNS complication (997.01) 18   Headache (784.0) 14   Other (individual codes <11 cases) 65  Procedures (ICD-9-CM procedure codes), n   V-EEG monitoring (89.19) 26   Venous catherization (38.93) 25   Other (individual codes <11 cases) 14 30-d readmission visit outcomes, n 183  Mean length of stay, days (SD) 6.8 (0.7)  Mean total costs, $USD (SD) $18 660 (2921)  Adverse discharge disposition (%) 26.4 90-d reasons for readmission, n 267  Diagnoses (ICD-9-CM diagnosis codes), n   Epilepsy (345.41, 345.50, 345.51, 345.60, 345.81, 345.90, 345.91) 103   Postoperative infection (998.59) 34   Postoperative CNS complication (997.01) 18   Headache (784.0 346.90) 16  Procedures (ICD-9-CM procedure codes), n   V-EEG monitoring (89.19) 47   Venous catherization (38.93) 25   Other (individual codes < 11 cases) 75 90-d readmission visit outcomes, n  Mean length of stay, days (SD) 5.8 (0.3)  Mean total costs, $USD (SD) $15 515 (1360)  Adverse discharge disposition (%) 21.8 Discharge disposition was dichotomized into the variable “adverse discharge” in which any discharge other than routine (transfer to skilled nursing, intermediate care facility, short-term acute-care hospital, home health care). View Large TABLE 1. Hospitalization-Associated Outcomes of 1587 Patients who Underwent Surgery for Intractable Epilepsy, NRD 2013 Variable Total patients (n) 1587 Surgical (index) visit outcomes  Mean LOS in days (SD) 8.3 (0.6)  Mean total costs, $USD (SD) $49 779 (5125)  Medical complications (n) 78  Surgical complications (n) 456  Adverse discharge disposition (%) 15.6 30-d reasons for readmission, n 183  Diagnoses (ICD-9-CM diagnosis codes), n   Epilepsy (345.3, 345.41, 345.51, 345.90) 58   Postoperative infection (998.59) 28   Postoperative CNS complication (997.01) 18   Headache (784.0) 14   Other (individual codes <11 cases) 65  Procedures (ICD-9-CM procedure codes), n   V-EEG monitoring (89.19) 26   Venous catherization (38.93) 25   Other (individual codes <11 cases) 14 30-d readmission visit outcomes, n 183  Mean length of stay, days (SD) 6.8 (0.7)  Mean total costs, $USD (SD) $18 660 (2921)  Adverse discharge disposition (%) 26.4 90-d reasons for readmission, n 267  Diagnoses (ICD-9-CM diagnosis codes), n   Epilepsy (345.41, 345.50, 345.51, 345.60, 345.81, 345.90, 345.91) 103   Postoperative infection (998.59) 34   Postoperative CNS complication (997.01) 18   Headache (784.0 346.90) 16  Procedures (ICD-9-CM procedure codes), n   V-EEG monitoring (89.19) 47   Venous catherization (38.93) 25   Other (individual codes < 11 cases) 75 90-d readmission visit outcomes, n  Mean length of stay, days (SD) 5.8 (0.3)  Mean total costs, $USD (SD) $15 515 (1360)  Adverse discharge disposition (%) 21.8 Variable Total patients (n) 1587 Surgical (index) visit outcomes  Mean LOS in days (SD) 8.3 (0.6)  Mean total costs, $USD (SD) $49 779 (5125)  Medical complications (n) 78  Surgical complications (n) 456  Adverse discharge disposition (%) 15.6 30-d reasons for readmission, n 183  Diagnoses (ICD-9-CM diagnosis codes), n   Epilepsy (345.3, 345.41, 345.51, 345.90) 58   Postoperative infection (998.59) 28   Postoperative CNS complication (997.01) 18   Headache (784.0) 14   Other (individual codes <11 cases) 65  Procedures (ICD-9-CM procedure codes), n   V-EEG monitoring (89.19) 26   Venous catherization (38.93) 25   Other (individual codes <11 cases) 14 30-d readmission visit outcomes, n 183  Mean length of stay, days (SD) 6.8 (0.7)  Mean total costs, $USD (SD) $18 660 (2921)  Adverse discharge disposition (%) 26.4 90-d reasons for readmission, n 267  Diagnoses (ICD-9-CM diagnosis codes), n   Epilepsy (345.41, 345.50, 345.51, 345.60, 345.81, 345.90, 345.91) 103   Postoperative infection (998.59) 34   Postoperative CNS complication (997.01) 18   Headache (784.0 346.90) 16  Procedures (ICD-9-CM procedure codes), n   V-EEG monitoring (89.19) 47   Venous catherization (38.93) 25   Other (individual codes < 11 cases) 75 90-d readmission visit outcomes, n  Mean length of stay, days (SD) 5.8 (0.3)  Mean total costs, $USD (SD) $15 515 (1360)  Adverse discharge disposition (%) 21.8 Discharge disposition was dichotomized into the variable “adverse discharge” in which any discharge other than routine (transfer to skilled nursing, intermediate care facility, short-term acute-care hospital, home health care). View Large National Rates, Causes of Readmission, and Discharge Outcomes A total of 183 (11.5%) patients from the surgical cohort were readmitted within 30 d, and 267 (16.8%) patients were readmitted within 90 d. The most common reasons for both 30- and 90-d readmission were persistent epilepsy, VEEG monitoring, postoperative infection, and postoperative CNS complication (997.01; Table 1). The rate of persistent epilepsy varied by surgery type: partial lobectomy/AHC (6.0%), total lobectomy (6.8%), and hemispherectomy (3.0%). The rate of postoperative infection also varied by surgery type: partial lobectomy/AHC (2.0%), total lobectomy (1.3%), and hemispherectomy (7.2%). VEEG did not vary by type of surgery. Readmission visits within 30 d were associated with longer length of stay during readmission (6.8 vs 5.8 d), greater costs ($18 660 vs $15 515), and increased percentage of adverse discharges (26.4% vs 21.8%) than readmissions within 90 d (Table 1). Main Results The complete univariate analysis of various variables on 30- and 90-d readmissions is shown in Tables 2 and 3. The rate of 30-d (P = .007) readmission, but not 90-d readmission (P = .078) was significantly higher among Medicaid and Medicare patients compared to privately insured. Overall, the rate of both 30-d (P = .024) and 90-d (P = .031) readmissions increased with the number of pre-existing comorbidities. Upon examination of 31 individual Elixhauser comorbidities, depression (11.9%) was the only one associated with significantly higher rates of 30-d (27.9% vs 11.5%, P < .001) and 90-d readmission (33.9% vs 16.8%, P < .001). Compared to the average medical complication rate among surgical patients of 5%, patients readmitted at 30 or 90 d had rates of 33% (P = .002) and 39% (P = .005), respectively, during their initial surgical visit. Descriptively, the rates of readmission among patients who suffered from surgical complications during the index visit were higher, but were not statistically significant. Lastly, prolonged length of stay and increased total hospital costs during the index visit were associated with 30- and 90-d readmission (P < .0001). We performed a multivariable analysis to identify predictors of prolonged length of stay, and the significant variables were 3+ pre-existing comorbidities (odds ratio [OR]: 7.8, 95% confidence interval [CI]: 2.8-22.0), IEEG (OR: 20.1, 95% CI: 7.9-54.7), VEEG (OR: 2.8, 95% CI: 1.1-7.6), hemispherectomy (OR: 4.2, 95% CI: 1.0-18.7), respiratory complications (OR: 11.4, 95% CI: 2.1-61.7), perioperative stroke/hematoma (OR: 9.7, 95% CI: 3.2-28.8), and status epilepticus (OR: 7.5, 95% CI: 1.2-46.0). TABLE 2. The Univariate (Unadjusted) Effect of Patient Demographics, Hospital Characteristics on 30- and 90-d Readmission Rates in 1587 Patients who Underwent Surgery for Intractable Epilepsy, NRD 2013 Variable Patients N (%) 30-d readmission rate N (%) 90-d readmission rate N (%)d Total patients (n) 1587 183 (11.5) 267 (16.8) Age group NS NS  0-20 562 (35.4) 58 (10.3) 111 (19.8)  21-35 418 (26.4) 49 (11.8) 59 (14.2)  36-50 332 (20.9) 44 (13.2) 58 (17.4)  51+ 275 (17.3) 32 (11.6) 38 (13.7) Gender NS NS  Male 806 (50.8) 90 (11.1) 140 (17.4)  Female 781 (49.2) 93 (12.0) 126 (16.2) Payera b NS  Private 888 (56.0 66 (7.5) 121 (13.7)  Medicare 257 (16.2) 57 (22.1) 68 (26.6)  Medicaid 345 (22.4) 53 (15.0) 64 (18.1)  Other 84 (5.3) c c Median annual household incomea NS NS  0-25th percentile 313 (20.2) 50 (15.9) 68 (21.8)  26th-50th percentile 382 (24.7) 37 (9.7) 58 (15.2)  51st-75th percentile 425 (27.4) 34 (8.0) 44 (10.3)  76th-100th percentile 428 (27.7) 63 (14.6) 94 (22.0) Bed size of hospital NS NS  Small 36 (2.3) c c  Medium 212 (13.3) 18 (8.4) 24 (11.1)  Large 1339 (84.4) 156 (11.7) 234 (17.4) Weekend admissiona NS NS  Weekday 1569 (98.9) 183 (11.7) 263 (16.7)  Weekend 17 (1.1) c c Hospital location NS NS  Small metropolitan 507 128 (11.9) 179 (16.5)  Large metropolitan 1,080 55 (10.8) 88 (17.4) Pre-existing comorbidity score (%)a b b  0 689 (43.4) 55 (7.9) 82 (11.8)  1 471 (29.7) 56 (11.9) 98 (20.7)  2 249 (15.7) 26 (10.2) 32 (12.7)  3+ 177 (11.2) 47 (26.6) 56 (31.4) Significant comorbidities  Depression 189 (11.9) 53 (27.9) 64 (33.9) Variable Patients N (%) 30-d readmission rate N (%) 90-d readmission rate N (%)d Total patients (n) 1587 183 (11.5) 267 (16.8) Age group NS NS  0-20 562 (35.4) 58 (10.3) 111 (19.8)  21-35 418 (26.4) 49 (11.8) 59 (14.2)  36-50 332 (20.9) 44 (13.2) 58 (17.4)  51+ 275 (17.3) 32 (11.6) 38 (13.7) Gender NS NS  Male 806 (50.8) 90 (11.1) 140 (17.4)  Female 781 (49.2) 93 (12.0) 126 (16.2) Payera b NS  Private 888 (56.0 66 (7.5) 121 (13.7)  Medicare 257 (16.2) 57 (22.1) 68 (26.6)  Medicaid 345 (22.4) 53 (15.0) 64 (18.1)  Other 84 (5.3) c c Median annual household incomea NS NS  0-25th percentile 313 (20.2) 50 (15.9) 68 (21.8)  26th-50th percentile 382 (24.7) 37 (9.7) 58 (15.2)  51st-75th percentile 425 (27.4) 34 (8.0) 44 (10.3)  76th-100th percentile 428 (27.7) 63 (14.6) 94 (22.0) Bed size of hospital NS NS  Small 36 (2.3) c c  Medium 212 (13.3) 18 (8.4) 24 (11.1)  Large 1339 (84.4) 156 (11.7) 234 (17.4) Weekend admissiona NS NS  Weekday 1569 (98.9) 183 (11.7) 263 (16.7)  Weekend 17 (1.1) c c Hospital location NS NS  Small metropolitan 507 128 (11.9) 179 (16.5)  Large metropolitan 1,080 55 (10.8) 88 (17.4) Pre-existing comorbidity score (%)a b b  0 689 (43.4) 55 (7.9) 82 (11.8)  1 471 (29.7) 56 (11.9) 98 (20.7)  2 249 (15.7) 26 (10.2) 32 (12.7)  3+ 177 (11.2) 47 (26.6) 56 (31.4) Significant comorbidities  Depression 189 (11.9) 53 (27.9) 64 (33.9) RR, readmission rate, NS, not significant. aExplanations provided to the reviewer about the apparent numerical discrepancies based on data present in the NRD for analysis. bStatistical significance (P < .05) via Pearson chi-square test. cLess than the HCUP reporting minimum of 11 cases. dThe 30 and 90-d readmission cohorts were not mutually exclusive, as the 90-d readmission cohort includes any patient that was readmitted within 90 d, including readmissions within 30 d. View Large TABLE 2. The Univariate (Unadjusted) Effect of Patient Demographics, Hospital Characteristics on 30- and 90-d Readmission Rates in 1587 Patients who Underwent Surgery for Intractable Epilepsy, NRD 2013 Variable Patients N (%) 30-d readmission rate N (%) 90-d readmission rate N (%)d Total patients (n) 1587 183 (11.5) 267 (16.8) Age group NS NS  0-20 562 (35.4) 58 (10.3) 111 (19.8)  21-35 418 (26.4) 49 (11.8) 59 (14.2)  36-50 332 (20.9) 44 (13.2) 58 (17.4)  51+ 275 (17.3) 32 (11.6) 38 (13.7) Gender NS NS  Male 806 (50.8) 90 (11.1) 140 (17.4)  Female 781 (49.2) 93 (12.0) 126 (16.2) Payera b NS  Private 888 (56.0 66 (7.5) 121 (13.7)  Medicare 257 (16.2) 57 (22.1) 68 (26.6)  Medicaid 345 (22.4) 53 (15.0) 64 (18.1)  Other 84 (5.3) c c Median annual household incomea NS NS  0-25th percentile 313 (20.2) 50 (15.9) 68 (21.8)  26th-50th percentile 382 (24.7) 37 (9.7) 58 (15.2)  51st-75th percentile 425 (27.4) 34 (8.0) 44 (10.3)  76th-100th percentile 428 (27.7) 63 (14.6) 94 (22.0) Bed size of hospital NS NS  Small 36 (2.3) c c  Medium 212 (13.3) 18 (8.4) 24 (11.1)  Large 1339 (84.4) 156 (11.7) 234 (17.4) Weekend admissiona NS NS  Weekday 1569 (98.9) 183 (11.7) 263 (16.7)  Weekend 17 (1.1) c c Hospital location NS NS  Small metropolitan 507 128 (11.9) 179 (16.5)  Large metropolitan 1,080 55 (10.8) 88 (17.4) Pre-existing comorbidity score (%)a b b  0 689 (43.4) 55 (7.9) 82 (11.8)  1 471 (29.7) 56 (11.9) 98 (20.7)  2 249 (15.7) 26 (10.2) 32 (12.7)  3+ 177 (11.2) 47 (26.6) 56 (31.4) Significant comorbidities  Depression 189 (11.9) 53 (27.9) 64 (33.9) Variable Patients N (%) 30-d readmission rate N (%) 90-d readmission rate N (%)d Total patients (n) 1587 183 (11.5) 267 (16.8) Age group NS NS  0-20 562 (35.4) 58 (10.3) 111 (19.8)  21-35 418 (26.4) 49 (11.8) 59 (14.2)  36-50 332 (20.9) 44 (13.2) 58 (17.4)  51+ 275 (17.3) 32 (11.6) 38 (13.7) Gender NS NS  Male 806 (50.8) 90 (11.1) 140 (17.4)  Female 781 (49.2) 93 (12.0) 126 (16.2) Payera b NS  Private 888 (56.0 66 (7.5) 121 (13.7)  Medicare 257 (16.2) 57 (22.1) 68 (26.6)  Medicaid 345 (22.4) 53 (15.0) 64 (18.1)  Other 84 (5.3) c c Median annual household incomea NS NS  0-25th percentile 313 (20.2) 50 (15.9) 68 (21.8)  26th-50th percentile 382 (24.7) 37 (9.7) 58 (15.2)  51st-75th percentile 425 (27.4) 34 (8.0) 44 (10.3)  76th-100th percentile 428 (27.7) 63 (14.6) 94 (22.0) Bed size of hospital NS NS  Small 36 (2.3) c c  Medium 212 (13.3) 18 (8.4) 24 (11.1)  Large 1339 (84.4) 156 (11.7) 234 (17.4) Weekend admissiona NS NS  Weekday 1569 (98.9) 183 (11.7) 263 (16.7)  Weekend 17 (1.1) c c Hospital location NS NS  Small metropolitan 507 128 (11.9) 179 (16.5)  Large metropolitan 1,080 55 (10.8) 88 (17.4) Pre-existing comorbidity score (%)a b b  0 689 (43.4) 55 (7.9) 82 (11.8)  1 471 (29.7) 56 (11.9) 98 (20.7)  2 249 (15.7) 26 (10.2) 32 (12.7)  3+ 177 (11.2) 47 (26.6) 56 (31.4) Significant comorbidities  Depression 189 (11.9) 53 (27.9) 64 (33.9) RR, readmission rate, NS, not significant. aExplanations provided to the reviewer about the apparent numerical discrepancies based on data present in the NRD for analysis. bStatistical significance (P < .05) via Pearson chi-square test. cLess than the HCUP reporting minimum of 11 cases. dThe 30 and 90-d readmission cohorts were not mutually exclusive, as the 90-d readmission cohort includes any patient that was readmitted within 90 d, including readmissions within 30 d. View Large TABLE 3. The Univariate (Unadjusted) Effect of Epilepsy Diagnosis, Presurgical Diagnostic Testing, Type of Surgery, Medical Complications, and Surgical Complications on 30-d and 90-d Readmission Rates in 1587 who Underwent Surgery for Intractable Epilepsy, NRD 2013 Patients N (%) 30-d RR (%) 90-d RR (%) d Total Patients (n) 1587 183 (11.5) 267 (16.8) Epilepsy diagnosisa NS NS  Generalized 76 (4.8) b 17 (22.4)  Complex Partial 904 (57.0) 111 (12.3) 148 (16.4)  Simple Partial 279 (17.6) 33 (11.7) 43 (15.5)  Other 392 (24.7) 42 (10.6) 64 (16.4) Presurgical diagnostic testing NS NS  VEEG 435 (27.2) 56 (12.8) 79 (18.2)  IEEG 568 (35.8) 66 (11.7) 88 (15.5) Type of surgerya NS NS  Hemispherectomy 66 (4.2) 15 (22.8) 24 (36.5)  Lobectomy 368 (23.2) 44 (11.9) 55 (14.9)  Amygdalohippocampectomy or partial lobectomy 1236 (77.9) 129 (10.4) 197 (16.0) Medical complications composite c c  0 1508 (95.0%) 157 (10.4) 236 (15.6)  1 79 (5%) 26 (33.0) 31 (39.0)  2 b b b  3+ b b b Medical complications individual  Cardiac b b b  Respiratory 24 (1.5) b 11 (45.4c)  Renal and urinary 14 (0.9) b b  Thromboembolic 32 (2.0) 18 (57.9) 18 (57.9) Surgical complications composite NS NS  0 1356 (85.4) 145 (10.7) 216 (16.0)  1 197 (12.4) 32 (16.2) 40 (20.1)  2 28 (1.8) b 11 (37.1)  3+ b b b Surgical complications individual NS NS  Stroke/hematoma 78 (4.9) b 14 (18.6)  Postoperative intracranial infections 34 (2.2) b b  Status epilepticus b b b  Hydrocephalus 11 (0.7) b b  VP shunt b b b  Blood transfusion 122 (7.7) 26 (21.4) 32 (26.5)  Postoperative mechanical ventilation 16 (1.0) b b Surgical visit discharge disposition  Routine 1479 (93.2) 162 (11.0)  Adverse 107 (6.8) 21 (19.4) 29 (27.1)  Mean surgical visit total costs (SD) 49 779 (5125) 58 780 (9666) 55 158 (8882)  Mean surgical visit length of stay (SD) 8.3 (0.6) 9.8 (1.1) 9.0 (1.0) Patients N (%) 30-d RR (%) 90-d RR (%) d Total Patients (n) 1587 183 (11.5) 267 (16.8) Epilepsy diagnosisa NS NS  Generalized 76 (4.8) b 17 (22.4)  Complex Partial 904 (57.0) 111 (12.3) 148 (16.4)  Simple Partial 279 (17.6) 33 (11.7) 43 (15.5)  Other 392 (24.7) 42 (10.6) 64 (16.4) Presurgical diagnostic testing NS NS  VEEG 435 (27.2) 56 (12.8) 79 (18.2)  IEEG 568 (35.8) 66 (11.7) 88 (15.5) Type of surgerya NS NS  Hemispherectomy 66 (4.2) 15 (22.8) 24 (36.5)  Lobectomy 368 (23.2) 44 (11.9) 55 (14.9)  Amygdalohippocampectomy or partial lobectomy 1236 (77.9) 129 (10.4) 197 (16.0) Medical complications composite c c  0 1508 (95.0%) 157 (10.4) 236 (15.6)  1 79 (5%) 26 (33.0) 31 (39.0)  2 b b b  3+ b b b Medical complications individual  Cardiac b b b  Respiratory 24 (1.5) b 11 (45.4c)  Renal and urinary 14 (0.9) b b  Thromboembolic 32 (2.0) 18 (57.9) 18 (57.9) Surgical complications composite NS NS  0 1356 (85.4) 145 (10.7) 216 (16.0)  1 197 (12.4) 32 (16.2) 40 (20.1)  2 28 (1.8) b 11 (37.1)  3+ b b b Surgical complications individual NS NS  Stroke/hematoma 78 (4.9) b 14 (18.6)  Postoperative intracranial infections 34 (2.2) b b  Status epilepticus b b b  Hydrocephalus 11 (0.7) b b  VP shunt b b b  Blood transfusion 122 (7.7) 26 (21.4) 32 (26.5)  Postoperative mechanical ventilation 16 (1.0) b b Surgical visit discharge disposition  Routine 1479 (93.2) 162 (11.0)  Adverse 107 (6.8) 21 (19.4) 29 (27.1)  Mean surgical visit total costs (SD) 49 779 (5125) 58 780 (9666) 55 158 (8882)  Mean surgical visit length of stay (SD) 8.3 (0.6) 9.8 (1.1) 9.0 (1.0) IEEG, intracranial electroencephalography; NS, not significant; RR, readmission rate; VEEG, video electroencephalography. aExplanations provided to the reviewer about the apparent numerical discrepancies based on data present in the NRD for analysis. bLess than the HCUP reporting minimum of 11 cases cStatistical significance (P < .05) via Pearson chi-square test. dThe 30- and 90-d readmission cohorts were not mutually exclusive, as the 90-d readmission cohort includes any patient that was readmitted within 90 d, including readmissions within 30 d. Discharge disposition was dichotomized into the variable “adverse discharge” in which any discharge other than routine (transfer to skilled nursing, intermediate care facility, short-term acute-care hospital, home health care). View Large TABLE 3. The Univariate (Unadjusted) Effect of Epilepsy Diagnosis, Presurgical Diagnostic Testing, Type of Surgery, Medical Complications, and Surgical Complications on 30-d and 90-d Readmission Rates in 1587 who Underwent Surgery for Intractable Epilepsy, NRD 2013 Patients N (%) 30-d RR (%) 90-d RR (%) d Total Patients (n) 1587 183 (11.5) 267 (16.8) Epilepsy diagnosisa NS NS  Generalized 76 (4.8) b 17 (22.4)  Complex Partial 904 (57.0) 111 (12.3) 148 (16.4)  Simple Partial 279 (17.6) 33 (11.7) 43 (15.5)  Other 392 (24.7) 42 (10.6) 64 (16.4) Presurgical diagnostic testing NS NS  VEEG 435 (27.2) 56 (12.8) 79 (18.2)  IEEG 568 (35.8) 66 (11.7) 88 (15.5) Type of surgerya NS NS  Hemispherectomy 66 (4.2) 15 (22.8) 24 (36.5)  Lobectomy 368 (23.2) 44 (11.9) 55 (14.9)  Amygdalohippocampectomy or partial lobectomy 1236 (77.9) 129 (10.4) 197 (16.0) Medical complications composite c c  0 1508 (95.0%) 157 (10.4) 236 (15.6)  1 79 (5%) 26 (33.0) 31 (39.0)  2 b b b  3+ b b b Medical complications individual  Cardiac b b b  Respiratory 24 (1.5) b 11 (45.4c)  Renal and urinary 14 (0.9) b b  Thromboembolic 32 (2.0) 18 (57.9) 18 (57.9) Surgical complications composite NS NS  0 1356 (85.4) 145 (10.7) 216 (16.0)  1 197 (12.4) 32 (16.2) 40 (20.1)  2 28 (1.8) b 11 (37.1)  3+ b b b Surgical complications individual NS NS  Stroke/hematoma 78 (4.9) b 14 (18.6)  Postoperative intracranial infections 34 (2.2) b b  Status epilepticus b b b  Hydrocephalus 11 (0.7) b b  VP shunt b b b  Blood transfusion 122 (7.7) 26 (21.4) 32 (26.5)  Postoperative mechanical ventilation 16 (1.0) b b Surgical visit discharge disposition  Routine 1479 (93.2) 162 (11.0)  Adverse 107 (6.8) 21 (19.4) 29 (27.1)  Mean surgical visit total costs (SD) 49 779 (5125) 58 780 (9666) 55 158 (8882)  Mean surgical visit length of stay (SD) 8.3 (0.6) 9.8 (1.1) 9.0 (1.0) Patients N (%) 30-d RR (%) 90-d RR (%) d Total Patients (n) 1587 183 (11.5) 267 (16.8) Epilepsy diagnosisa NS NS  Generalized 76 (4.8) b 17 (22.4)  Complex Partial 904 (57.0) 111 (12.3) 148 (16.4)  Simple Partial 279 (17.6) 33 (11.7) 43 (15.5)  Other 392 (24.7) 42 (10.6) 64 (16.4) Presurgical diagnostic testing NS NS  VEEG 435 (27.2) 56 (12.8) 79 (18.2)  IEEG 568 (35.8) 66 (11.7) 88 (15.5) Type of surgerya NS NS  Hemispherectomy 66 (4.2) 15 (22.8) 24 (36.5)  Lobectomy 368 (23.2) 44 (11.9) 55 (14.9)  Amygdalohippocampectomy or partial lobectomy 1236 (77.9) 129 (10.4) 197 (16.0) Medical complications composite c c  0 1508 (95.0%) 157 (10.4) 236 (15.6)  1 79 (5%) 26 (33.0) 31 (39.0)  2 b b b  3+ b b b Medical complications individual  Cardiac b b b  Respiratory 24 (1.5) b 11 (45.4c)  Renal and urinary 14 (0.9) b b  Thromboembolic 32 (2.0) 18 (57.9) 18 (57.9) Surgical complications composite NS NS  0 1356 (85.4) 145 (10.7) 216 (16.0)  1 197 (12.4) 32 (16.2) 40 (20.1)  2 28 (1.8) b 11 (37.1)  3+ b b b Surgical complications individual NS NS  Stroke/hematoma 78 (4.9) b 14 (18.6)  Postoperative intracranial infections 34 (2.2) b b  Status epilepticus b b b  Hydrocephalus 11 (0.7) b b  VP shunt b b b  Blood transfusion 122 (7.7) 26 (21.4) 32 (26.5)  Postoperative mechanical ventilation 16 (1.0) b b Surgical visit discharge disposition  Routine 1479 (93.2) 162 (11.0)  Adverse 107 (6.8) 21 (19.4) 29 (27.1)  Mean surgical visit total costs (SD) 49 779 (5125) 58 780 (9666) 55 158 (8882)  Mean surgical visit length of stay (SD) 8.3 (0.6) 9.8 (1.1) 9.0 (1.0) IEEG, intracranial electroencephalography; NS, not significant; RR, readmission rate; VEEG, video electroencephalography. aExplanations provided to the reviewer about the apparent numerical discrepancies based on data present in the NRD for analysis. bLess than the HCUP reporting minimum of 11 cases cStatistical significance (P < .05) via Pearson chi-square test. dThe 30- and 90-d readmission cohorts were not mutually exclusive, as the 90-d readmission cohort includes any patient that was readmitted within 90 d, including readmissions within 30 d. Discharge disposition was dichotomized into the variable “adverse discharge” in which any discharge other than routine (transfer to skilled nursing, intermediate care facility, short-term acute-care hospital, home health care). View Large In multivariable analysis analyzing risk factors for readmission, we found that Medicare, but not Medicaid, patients were more likely to suffer 30-d readmission and 90-d readmission (Table 4). In subgroup analysis, Medicaid patients had a higher rate of 3 or more comorbidities (19.3% vs 13.2%, P = .03) than Medicare patients. Depression was significantly associated with increased likelihood of both 30-d (OR: 4.2, 95% CI: 1.9-9.4, P < .001) and 90-d readmission (OR: 3.7, 95% CI: 1.7-7.7, P = .001). Patients receiving hemispherectomy compared to other surgical approaches were more likely to be readmitted at both 30 d (OR: 4.5, 95% CI: 1.4-14.7, P = .014) and 90 d (OR: 5.0, 95% CI: 1.7-14.4, P = .003). In a subgroup analysis, hemispherectomy was associated with decreased rate of persistent epilepsy-related readmission, but increased rate of postoperative infection-related readmission. If a medical complication occurred during the index visit, a patient's likelihood of both 30-d (OR: 5.2, 95% CI: 2.1-13.2, P = .001) and 90-d readmission (OR: 4.3, 95% CI: 1.8-10.4, P = .012) increased significantly. TABLE 4. Multivariable Predictors of 30- and 90-d Readmission in 1587 Patients who Underwent Surgery for Intractable Epilepsy Variable 30-d readmission 90-d readmission Total Patients (n) OR (95% CI) P value OR (95% CI) P value Payer .020a .043a  Private Reference Reference  Medicare 4.0 (1.6-10.1) 2.7 (1.3-5.8)  Medicaid 2.3 (1.0-5.4) 1.4 (0.7-2.6)  Other 0.9 (0.2-4.5) 0.8 (0.3-2.2) Median annual household income .038a .006a  0-25th percentile Reference Reference  26th-50th percentile 0.6 (0.2-1.6) 0.7 (0.3-1.6)  51st-75th percentile 0.4 (0.3-0.9) 0.4 (0.2-0.8)  76th-100th percentile 1.3 (0.5-3.8) 1.4 (0.7-2.7) Bed size of hospital .001a –  Small Reference –  Medium 0.2 (0.08-0.6) –  Large 0.3 (0.2-0.6) – Pre-existing comorbidityb  Depression 4.2 (1.9-9.4) <.001a 3.7 (1.7-7.7) .001a Type of surgery .014a .003a  Other surgery Reference Reference  Hemispherectomy 4.5 (1.4-14.7) 5.0 (1.7-14.4) Medical complication scorec 5.2 (2.08-13.2) .001a 4.3 (1.8-10.4) .012a Variable 30-d readmission 90-d readmission Total Patients (n) OR (95% CI) P value OR (95% CI) P value Payer .020a .043a  Private Reference Reference  Medicare 4.0 (1.6-10.1) 2.7 (1.3-5.8)  Medicaid 2.3 (1.0-5.4) 1.4 (0.7-2.6)  Other 0.9 (0.2-4.5) 0.8 (0.3-2.2) Median annual household income .038a .006a  0-25th percentile Reference Reference  26th-50th percentile 0.6 (0.2-1.6) 0.7 (0.3-1.6)  51st-75th percentile 0.4 (0.3-0.9) 0.4 (0.2-0.8)  76th-100th percentile 1.3 (0.5-3.8) 1.4 (0.7-2.7) Bed size of hospital .001a –  Small Reference –  Medium 0.2 (0.08-0.6) –  Large 0.3 (0.2-0.6) – Pre-existing comorbidityb  Depression 4.2 (1.9-9.4) <.001a 3.7 (1.7-7.7) .001a Type of surgery .014a .003a  Other surgery Reference Reference  Hemispherectomy 4.5 (1.4-14.7) 5.0 (1.7-14.4) Medical complication scorec 5.2 (2.08-13.2) .001a 4.3 (1.8-10.4) .012a aStatistical significance (P < .05) via Pearson chi-square test. bOdds ratio represents increased likelihood of readmission per each additional comorbidity. cOdds ratio represents increased likelihood of readmission per each additional complication. dIndicates variables not included in the final multivariable model. View Large TABLE 4. Multivariable Predictors of 30- and 90-d Readmission in 1587 Patients who Underwent Surgery for Intractable Epilepsy Variable 30-d readmission 90-d readmission Total Patients (n) OR (95% CI) P value OR (95% CI) P value Payer .020a .043a  Private Reference Reference  Medicare 4.0 (1.6-10.1) 2.7 (1.3-5.8)  Medicaid 2.3 (1.0-5.4) 1.4 (0.7-2.6)  Other 0.9 (0.2-4.5) 0.8 (0.3-2.2) Median annual household income .038a .006a  0-25th percentile Reference Reference  26th-50th percentile 0.6 (0.2-1.6) 0.7 (0.3-1.6)  51st-75th percentile 0.4 (0.3-0.9) 0.4 (0.2-0.8)  76th-100th percentile 1.3 (0.5-3.8) 1.4 (0.7-2.7) Bed size of hospital .001a –  Small Reference –  Medium 0.2 (0.08-0.6) –  Large 0.3 (0.2-0.6) – Pre-existing comorbidityb  Depression 4.2 (1.9-9.4) <.001a 3.7 (1.7-7.7) .001a Type of surgery .014a .003a  Other surgery Reference Reference  Hemispherectomy 4.5 (1.4-14.7) 5.0 (1.7-14.4) Medical complication scorec 5.2 (2.08-13.2) .001a 4.3 (1.8-10.4) .012a Variable 30-d readmission 90-d readmission Total Patients (n) OR (95% CI) P value OR (95% CI) P value Payer .020a .043a  Private Reference Reference  Medicare 4.0 (1.6-10.1) 2.7 (1.3-5.8)  Medicaid 2.3 (1.0-5.4) 1.4 (0.7-2.6)  Other 0.9 (0.2-4.5) 0.8 (0.3-2.2) Median annual household income .038a .006a  0-25th percentile Reference Reference  26th-50th percentile 0.6 (0.2-1.6) 0.7 (0.3-1.6)  51st-75th percentile 0.4 (0.3-0.9) 0.4 (0.2-0.8)  76th-100th percentile 1.3 (0.5-3.8) 1.4 (0.7-2.7) Bed size of hospital .001a –  Small Reference –  Medium 0.2 (0.08-0.6) –  Large 0.3 (0.2-0.6) – Pre-existing comorbidityb  Depression 4.2 (1.9-9.4) <.001a 3.7 (1.7-7.7) .001a Type of surgery .014a .003a  Other surgery Reference Reference  Hemispherectomy 4.5 (1.4-14.7) 5.0 (1.7-14.4) Medical complication scorec 5.2 (2.08-13.2) .001a 4.3 (1.8-10.4) .012a aStatistical significance (P < .05) via Pearson chi-square test. bOdds ratio represents increased likelihood of readmission per each additional comorbidity. cOdds ratio represents increased likelihood of readmission per each additional complication. dIndicates variables not included in the final multivariable model. View Large DISCUSSION Cause of Readmissions The major cause of readmissions is persistent epilepsy and VEEG monitoring. It is well known that epilepsy surgery success rates range from 50% to 80%.14,15 Hence, it is not surprising that patients might require readmission for additional monitoring or management of persistent seizures. Similar findings in a smaller study were reported by Wilson et al16 who examined the records of 100 consecutive epilepsy patients undergoing anterior temporal lobectomy. Twenty-one (21%) of patients required readmission.16 The most common cause of readmission were psychiatric reasons (anxiety, depression; 53%), and the second most common cause was persistent epilepsy (28%).16 In our multivariable analysis of risk factors, depression was the only pre-existing comorbidity that was significantly associated with increased likelihood of readmission. Similarly, Kanner et al,17 in a retrospective review of 95 TLE patients, demonstrated that a life-time history of depression preoperatively was a predictor of worse postsurgical seizure outcome, specifically, less likely to be seizure free. Patient demographics and hospital type were found to influence the rate of readmissions. We found that Medicare but not Medicaid or privately insured were significantly more likely to be readmitted within both 30 and 90 d. Medicare patients are typically much older than Medicaid patients and may be more likely to have medical comorbidities. Upon further subgroup analysis, we found, however, that Medicaid patients had a higher rate of 3 or more comorbidities (19.3% vs 13.2%, P = .03) than Medicare patients. Medicare patients are less likely to have socioeconomic barriers to medical care compared to Medicaid patients. However, this could not be further evaluated in our dataset. Aggressive maintenance of antiepileptic medications after surgery might reduce readmissions for persistent epilepsy or monitoring. Likewise, scheduled visits with psychiatrists, therapists, and social workers might prevent psychiatric issues from arising as frequently after surgery. The categorical variable for median annual household income was statistically significant, but we were unable to identify clinically significant trends. For example, the analysis found that patients in the 51th to 75th percentile were less likely to be readmitted compared to patients in the 0 to 25th percentile. However, this trend did not apply to the 76th to 100th percentile in terms of income. Smaller hospitals (fewer beds) were significantly more likely to have 30-d readmissions compared to larger hospitals. Larger hospitals are known to have more resources, including advanced technology, and are more likely to be teaching facilities. Furthermore, smaller hospitals generally represent lower volume centers with less experience and are more likely to reach capacity, potentially prompting premature discharges. However, the precise characteristics of smaller hospitals that lead to more readmissions were not able to be queried from our database. A similar argument can be made for hemispherectomy surgery. These surgeries are often performed on young children and are long, tedious surgeries, which often require blood transfusions. Vadera et al11 utilized the NIS to identify 304 pediatric patients who underwent hemispherectomy between 1988 and 2010. They found that 56% of patients encountered a complication (42% related to surgery).11 In the pediatric population, undergoing such a long, complicated surgery prone to blood loss likely necessitates a transfusion as part of the procedure, and it is unclear if this is truly a preventable complication. The length and complexity of the surgery likely also increased the risk of postoperative infection, as borne out by our data. Frequency and Timing of Readmissions Prior studies of readmission for epilepsy have been examined as a subgroup of all craniotomies and not thoroughly analyzed. Moghavem et al18 investigated unplanned 30-d readmissions utilizing the SID for California, Florida, and New York. They identified 43 356 patients, of which 2.80% underwent cranial neurosurgery for seizure.18 Thirty-day readmissions were 13.89% for the seizure cohort.18 Significant predictors for readmission included male gender (OR 1.74; 95% CI 1.17-2.60) and initial admission via emergency department (OR 2.22; 95% CI 1.45-3.43).18 Taylor et al19 investigated unplanned readmissions following neurosurgery utilizing a statewide database for New York. One hundred fifty-four patients had epilepsy surgery, and the 30-d readmission rate was 7.14%.19 The overall most common reasons for unplanned readmission included infection (29.52%), medical complications (19.22%), stroke (6.09%), and venous thromboembolism (5.71%). Postoperative seizures accounted for less than 5%.19 Our study presents quite different conclusions for top reasons for readmission, but postoperative seizures accounted for approximately 5.88% of patients, similar to the aforementioned results. It is possible that results from our study vary due to the nationwide nature of the population as well as aspects of coding database queries. Our readmission rates for 30 and 90 d (11.5% and 16.8%, respectively) are of similar magnitude to previously reported studies: 7.14% to 21%. Almost two-thirds of the readmissions occurred within 30 d. Rates vary, however, due to study design including single-institution,16 single state,19 or select states18 from the SID. Our study adds data for readmissions outside the initial 30-d window and tracks patients longitudinally, perhaps more accurately reflecting readmissions. Reasons for admission are not qualitatively different at 30 vs 90 d, and this distinction will not alter prevention strategies. Predictors of 90-d readmission, in addition to Medicare payer status and depression, included increasing number of comorbidities and medical complications. Although pre-existing comorbidities were only significant in univariate analysis, each additional comorbidity and/or complication increased the likelihood of readmission. The medical complication score represents the OR for increased likelihood of readmission for each additional comorbidity or complication within 30 and 90 d. The most common medical complications likely resulted from mechanical ventilation and immobility or deep venous thrombosis (DVT). Therefore, additional attention is deserved for preventing these medical complications, ie, DVT prophylaxis, early ambulation, and incentive spirometry.20 Our results indicate that medical, rather than surgical, complications are a stronger contributor to the likelihood of readmission. However, no patient had more than one reported medical complication during index hospitalization. Furthermore, upon descriptive analysis, patients with surgical complications appeared to have higher rates of 30-d (16.2% vs 10.7%) and 90-d (37.1% vs 16%) readmission. Thus, these comparisons likely failed to reach statistical significance due to lack of statistical power resulting from the relatively low number of surgical complications. Limitations The NRD maintains patient-linkage numbers to track patients longitudinally, which is a limitation of other databases. The NRD does have limitations including its dependence on ICD-9-CM coding, with varying sensitivities and specificities, and small percentage of missing data for some variables.21-27 It is possible that additional causes of readmission listed in secondary diagnosis fields were missed. Since the NRD was only available for 1 calendar year when this study was conducted, our follow-up time was limited. In order to have a substantial population size for analysis as well as a reasonable follow-up period, we included surgeries performed during the first 9 mo of the year and left the remaining 3 mo (90 d) as a window to capture readmissions. Since the diagnoses in the dataset are compiled by medical coders, not clinicians, the diagnosis of stroke and hematoma is likely overestimated since most routine postoperative imaging will show some degree of blood and/or clinically insignificant ischemia. However, if this complication was not specifically entered into the patient chart by the clinical team, it is unlikely to have been coded or billed. Additionally, the NRD, like other databases, lacks functional status, clinical exams, radiology impressions, surgeon details, and surgical technique. Quality of life is not recorded or tracked in the NRD. Certain univariate variables could not be considered in the multivariate analysis due to small sample size. CONCLUSION Minimizing factors that contribute to readmission in various patient populations and procedures becomes important for patient care, resource utilization, and physician reimbursement. In epilepsy surgery, the majority of readmissions occurred within 30 d, and the causes of readmission were similar within 90 d: persistent epilepsy, postoperative infection, and postoperative CNS complications. Awareness of the reasons for readmission is important for patient counseling and surgical decision-making. In epilepsy surgery, readmissions might be reduced with optimization of pre-existing comorbidities particularly in the elderly, minimization of complications, close follow-up with hemispherectomy patients, and adequate control of epileptic seizures during the index visit. However, persistent seizures after a procedure with a known efficacy of only 50% to 80% might not be avoidable. Disclosures Dr Arnold has commercial relationships with Z-Plasty (stock, stock options, or other ownership interest); Medtronic Sofamor Danek (salary and any payment for services not otherwise identified as salary such as consulting, fees, honoraria, paid authorship, or other payments for services); Stryker Spine (sponsored or reimbursed travel, salary and any payment for services no otherwise identified as salary such as consulting, fees, honoraria, paid authorship, or other payments for services); AO Spine North America (sponsored or reimbursed travel); Invivo (salary and any payment for series not otherwise identified as salary such as consulting, fees, honoraria, paid authorship, or other payments for services). Dr Schwartz has stock options in Visionsense. The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article. Notes The content of this article was presented at the Neurosurgical Society of America meeting on April 3, 2017 in Jacksonville, Florida, as an Oral Presentation. REFERENCES 1. Kim Y , Gani F , Lucas DJ et al. Early versus late readmission after surgery among patients with employer-provided health insurance . Ann Surg . 2015 ; 262 ( 3 ): 502 – 511 . Google Scholar CrossRef Search ADS PubMed 2. Koubeissi MZ , Puwanant A , Jehi L , Alshekhlee A . In-hospital complications of epilepsy surgery: a six-year nationwide experience . Br J Neurosurg . 2009 ; 23 ( 5 ): 524 – 529 . Google Scholar CrossRef Search ADS PubMed 3. McClelland S 3rd , Guo H , Okuyemi KS . Racial disparities in the surgical management of intractable temporal lobe epilepsy in the United States . Arch Neurol . 2010 ; 67 ( 5 ): 577 – 583 . Google Scholar CrossRef Search ADS PubMed 4. Vale FL , Reintjes S , Garcia HG . Complications after mesial temporal lobe surgery via inferiortemporal gyrus approach . Neurosurg Focus . 2013 ; 34 ( 6 ): E2 . Google Scholar CrossRef Search ADS PubMed 5. NRD Overview . Healthcare Cost and Utilization Project (HCUP) . 2015 . Available at: www.hcup-us.ahrq.gov/nrdoverview.jsp. Accessed June 1, 2016 . 6. HCUP . Clinical Classifications Software (CCS) for ICD-9-CM . 2011 . Available at: www.hcup-us.ahrq.gov/toolssoftware/ccs/ccs.jsp. Accessed June 1, 2016 . 7. Schiltz NK , Koroukian SM , Lhatoo SD , Kaiboriboon K . Temporal trends in pre-surgical evaluations and epilepsy surgery in the U.S. from 1998–2009 . Epilepsy Res . 2013 ; 103 ( 2-3 ): 270 – 278 . Google Scholar CrossRef Search ADS PubMed 8. McClelland S 3rd , Curran CC , Davey CS , Okuyemi KS . Intractable pediatric temporal lobe epilepsy in the United States: examination of race, age, sex, and insurance status as factors predicting receipt of resective treatment . J Neurosurg . 2007 ; 107 ( 6 Suppl ): 469 – 473 . Google Scholar PubMed 9. HCUP . Comorbidity Software . 2012 . Available at: www.hcup-us.ahrq.gov/toolssoftware/comorbidity/comorbidity.jsp. Accessed June 1, 2016 . 10. Elixhauser A , Steiner C , Harris DR , Coffey RM . Comorbidity measures for use with administrative data . Med Care . 1998 ; 36 ( 1 ): 8 – 27 . Google Scholar CrossRef Search ADS PubMed 11. Vadera S , Griffith SD , Rosenbaum BP et al. National trends and In-hospital complication rates in more than 1600 hemispherectomies from 1988 to 2010 . Neurosurgery . 2015 ; 77 ( 2 ): 185 – 191 . Google Scholar CrossRef Search ADS PubMed 12. Sharma M , Sonig A , Ambekar S , Nanda A . Discharge dispositions, complications, and costs of hospitalization in spinal cord tumor surgery: analysis of data from the United States Nationwide Inpatient Sample, 2003–2010 . J Neurosurg Spine . 2014 ; 20 ( 2 ): 125 – 141 . Google Scholar CrossRef Search ADS PubMed 13. HCUP . Cost-to-Charge Ratio Files . 2012 . Available at: www.hcup-us.ahrq.gov/db/state/costtocharge.jsp. Accessed June 1, 2016 . 14. Spencer S , Huh L . Outcomes of epilepsy surgery in adults and children . Lancet Neurol . 2008 ; 7 ( 6 ): 525 – 537 . Google Scholar CrossRef Search ADS PubMed 15. Wiebe S , Blume WT , Girvin JP , Eliasziw M , Effectiveness, Efficiency of Surgery for Temporal Lobe Epilepsy Study Group. A randomized, controlled trial of surgery for temporal-lobe epilepsy . N Engl J Med . 2001 ; 345 ( 5 ): 311 – 318 . Google Scholar CrossRef Search ADS PubMed 16. Wilson SJ , Kincade P , Saling MM , Bladin PF . Patient readmission and support utilization following anterior temporal lobectomy . Seizure . 1999 ; 8 ( 1 ): 20 – 25 . Google Scholar CrossRef Search ADS PubMed 17. Kanner AM BR , Smith MC , Balabanov AJ , Frey M . Does a life-time history of depression predict a worse post-surgical seizure outcome following temporal lobectomy? Ann Neurol . 2006 ; 60 ( Suppl 3 ): S19 . 18. Moghavem N , Morrison D , Ratliff JK , Hernandez-Boussard T . Cranial neurosurgical 30-day readmissions by clinical indication . J Neurosurg . 2015 ; 123 ( 1 ): 189 – 197 . Google Scholar CrossRef Search ADS PubMed 19. Taylor BE , Youngerman BE , Goldstein H et al. Causes and timing of unplanned early readmission after neurosurgery . Neurosurgery . 2016 ; 79 ( 3 ): 356 – 369 . Google Scholar CrossRef Search ADS PubMed 20. Pavon JM , Adam SS , Razouki ZA et al. Effectiveness of intermittent pneumatic compression devices for venous thromboembolism prophylaxis in High-Risk surgical patients: A systematic review . J Arthroplasty . 2016 ; 31 ( 2 ): 524 – 532 . Google Scholar CrossRef Search ADS PubMed 21. Gologorsky Y , Knightly JJ , Chi JH , Groff MW . The Nationwide Inpatient Sample database does not accurately reflect surgical indications for fusion . J Neurosurg Spine . 2014 ; 21 ( 6 ): 984 – 993 . Google Scholar CrossRef Search ADS PubMed 22. Jalai CM , Worley N , Marascalchi BJ et al. The impact of advanced age on peri-operative outcomes in the surgical treatment of cervical spondylotic myelopathy . Spine . 2016 ; 41 ( 3 ): E139 – E147 . Google Scholar CrossRef Search ADS PubMed 23. King JT Jr , Abbed KM , Gould GC , Benzel EC , Ghogawala Z . Cervical spine reoperation rates and hospital resource utilization after initial surgery for degenerative cervical spine disease in 12 338 patients in Washington State . Neurosurgery . 2009 ; 65 ( 6 ): 1011 – 1023 . Google Scholar CrossRef Search ADS PubMed 24. Marquez-Lara A , Nandyala SV , Fineberg SJ , Singh K . Current trends in demographics, practice, and In-Hospital outcomes in cervical spine surgery . Spine . 2014 ; 39 ( 6 ): 476 – 481 . Google Scholar CrossRef Search ADS PubMed 25. Patil CG , Santarelli J , Lad SP , Ho C , Tian W , Boakye M . Inpatient complications, mortality, and discharge disposition after surgical correction of idiopathic scoliosis: a national perspective . Spine J . 2008 ; 8 ( 6 ): 904 – 910 . Google Scholar CrossRef Search ADS PubMed 26. Shamji MF , Cook C , Pietrobon R , Tackett S , Brown C , Isaacs RE . Impact of surgical approach on complications and resource utilization of cervical spine fusion: a nationwide perspective to the surgical treatment of diffuse cervical spondylosis . Spine J . 2009 ; 9 ( 1 ): 31 – 38 . Google Scholar CrossRef Search ADS PubMed 27. Wang MC , Chan L , Maiman DJ , Kreuter W , Deyo RA . Complications and mortality associated with cervical spine surgery for degenerative disease in the United States . Spine . 2007 ; 32 ( 3 ): 342 – 347 . Google Scholar CrossRef Search ADS PubMed Operative Neurosurgery Speaks! Audio abstracts available for this article at www.operativeneurosurgery-online.com. COMMENTS The authors have provided a valuable review of the NRD for data on patients undergoing open epilepsy surgery. Their findings are very much in line with experience, literature, and common sense. Failure to control epilepsy and surgical complications, especially in more complex operations were the most likely reasons for readmission. Another interesting finding is that patients covered by Medicare or coming from a lower socioeconomic background were more likely to be readmitted. This may be a proxy for more severe epilepsy, or for a lack of home resources for postoperative care. The unique predictor of 30-day readmission was a small treating hospital. It is not clear why this would be the case, but may be an indicator of a smaller epilepsy program. This data may be of use to epilepsy surgery centers as it may mark those patients who should be further supported with home healthcare services, or should be seen in the outpatient clinic soon after surgery for evaluation. Richard W. Byrne Chicago, Illinois The present study investigates rates, causes, and predictors of 30- and 90-day readmissions after epilepsy surgery, by using the Nationwide Readmission Database on a significantly large sample of readmitted cases. The authors should be congratulated for their effort, since the presented results may represent a useful basis to measure the impact of readmissions on additional costs for the healthcare system and to direct future actions aimed at limiting this phenomenon. Massimo Cossu Milano, Italy In their publication the authors report on the United States readmission rate after epilepsy surgery between January and September 2013 (roughly, an 8-month period). They queried a nationwide US database for readmission and used the ICD-9CM code to identify patients who were readmitted after being treated for intractable epilepsy (with surgery). Patients had undergone a variety of procedures (hemispherectomy, lobectomy, partial lobectomy, or amygdalohippocampectomy). A plethora of predictors were statistically analyzed to characterize readmission reasons. A total of 1587 patients were identified that fulfilled the mentioned criteria. The 30-day readmission rate was 11.5% and the 90-day readmission rate was 16.8%. Reasons for this were typically persistent epilepsy, postoperative complications, or severity of epilepsy (indicated by Video EEG). This could be - medically and politically - a very interesting paper. It probably sheds some light on the medical care given to patients with intractable epilepsy (and more so in other chronic conditions). In our experience a typical readmission happens when a patient was prematurely discharged; either a medication was not fully titrated (up to a therapeutic level) or a complication was not excluded/fully treated, or the patient was simply not fit enough to go home and take care of himself. Of course, another reason could be a high complication rate at a hospital that is not usually dealing with complicated procedures. Nevertheless, the pressure on neurosurgical units these days is high and often now financially driven. Decision making is not solely based on medical facts but often driven by financial considerations. We as physicians might draw conclusions out of information like this and try to raise pressure on politicians and hospital authorities against pure financial towards more medical reasoning when discharging a patient who otherwise might benefit from some more days under our care. Volker A. Coenen Freiburg im Breisgau, Germany Operative Neurosurgery Speaks (Audio Abstracts) Listen to audio translations of this paper's abstract into select languages by choosing from one of the selections below. Chinese: Zuowei Wang, MD. Department of Neurosurgery Beijing Hospital Beijing, China Chinese: Zuowei Wang, MD. Department of Neurosurgery Beijing Hospital Beijing, China Close French: Johan Pallud, MD, PhD. Department of Neurosurgery Medical School of Paris Descartes University Paris, France French: Johan Pallud, MD, PhD. Department of Neurosurgery Medical School of Paris Descartes University Paris, France Close English: William W. Ashley, MD, PhD, MBA. Department of Neurological Surgery Sinai Hospital and LifeBridge Health System Baltimore, Maryland English: William W. Ashley, MD, PhD, MBA. Department of Neurological Surgery Sinai Hospital and LifeBridge Health System Baltimore, Maryland Close Italian: Marco Cenzato, MD. Neurosurgical Department Ospedale Maggiore Niguarda Ca 'Grande Milan, Italy Italian: Marco Cenzato, MD. Neurosurgical Department Ospedale Maggiore Niguarda Ca 'Grande Milan, Italy Close Spanish: Alejandro Enriquez-Marulanda, MD. Department of Neurosurgery Beth Israel Deaconess Medical Center Boston, Massachusetts Spanish: Alejandro Enriquez-Marulanda, MD. Department of Neurosurgery Beth Israel Deaconess Medical Center Boston, Massachusetts Close Portuguese: José Luís Alves, MD. Department of Neurosurgery Centro Hospitalar e Universitário de Coimbra Coimbra, Portugal Portuguese: José Luís Alves, MD. Department of Neurosurgery Centro Hospitalar e Universitário de Coimbra Coimbra, Portugal Close Japanese: Masaru Aoyagi, MD. Department of Neurosurgery Tokyo Medical and Dental University Tokyo, Japan Japanese: Masaru Aoyagi, MD. Department of Neurosurgery Tokyo Medical and Dental University Tokyo, Japan Close Korean: Hye Ran Park, MD. Department of Neurosurgery Soonchunhyang University Seoul Hospital Seoul, Republic of Korea Korean: Hye Ran Park, MD. Department of Neurosurgery Soonchunhyang University Seoul Hospital Seoul, Republic of Korea Close Russian: Vsevolod Shurhkhay, MD. Burdenko Institute of Neurosurgery Moscow, Russian Federation Russian: Vsevolod Shurhkhay, MD. Burdenko Institute of Neurosurgery Moscow, Russian Federation Close Greek: Marios Themistocleous, MD. Department of Neurosurgery Aghia Sophia Children's Hospital Athens, Greece Greek: Marios Themistocleous, MD. Department of Neurosurgery Aghia Sophia Children's Hospital Athens, Greece Close Copyright © 2018 by the Congress of Neurological Surgeons This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Operative NeurosurgeryOxford University Press

Published: Jun 4, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off