Performing concurrent operations in academic vascular neurosurgery does not affect patient outcomes

View More View Less
  • 1 Department of Neurological Surgery,
  • 2 Center for Healthcare Value,
  • 3 Department of Quality, UCSF Health, and
  • 4 Department of Anesthesiology, University of California, San Francisco, California
Full access

OBJECTIVE

Concurrent surgeries, also known as “running two rooms” or simultaneous/overlapping operations, have recently come under intense scrutiny. The goal of this study was to evaluate the operative time and outcomes of concurrent versus nonconcurrent vascular neurosurgical procedures.

METHODS

The authors retrospectively reviewed 1219 procedures performed by 1 vascular neurosurgeon from 2012 to 2015 at the University of California, San Francisco. Data were collected on patient age, sex, severity of illness, risk of mortality, American Society of Anesthesiologists (ASA) status, procedure type, admission type, insurance, transfer source, procedure time, presence of resident or fellow in operating room (OR), number of co-surgeons, estimated blood loss (EBL), concurrent vs nonconcurrent case, severe sepsis, acute respiratory failure, postoperative stroke causing neurological deficit, unplanned return to OR, 30-day mortality, and 30-day unplanned readmission. For aneurysm clipping cases, data were also obtained on intraoperative aneurysm rupture and postoperative residual aneurysm. Chi-square and t-tests were performed to compare concurrent versus nonconcurrent cases, and then mixed-effects models were created to adjust for different procedure types, patient demographics, and clinical indicators between the 2 groups.

RESULTS

There was a significant difference in procedure type for concurrent (n = 828) versus nonconcurrent (n = 391) cases. Concurrent cases were more likely to be routine/elective admissions (53% vs 35%, p < 0.001) and physician referrals (59% vs 38%, p < 0.001). This difference in patient/case type was also reflected in the lower severity of illness, risk of death, and ASA class in the concurrent versus nonconcurrent cases (p < 0.01). Concurrent cases had significantly longer procedural times (243 vs 213 minutes) and more unplanned 30-day readmissions (5.7% vs 3.1%), but shorter mean length of hospital stay (11.2 vs 13.7 days), higher rates of discharge to home (66% vs 51%), lower 30-day mortality rates (3.1% vs 6.1%), lower rates of acute respiratory failure (4.3% vs 8.2%), and decreased 30-day unplanned returns to the OR (3.3% vs 6.9%; all p < 0.05). Rates of severe sepsis, postoperative stroke, intraoperative aneurysm rupture, and postoperative aneurysm residual were equivalent between the concurrent and nonconcurrent groups (all p values nonsignificant). Mixed-effects models showed that after controlling for procedure type, patient demographics, and clinical indicators, there was no significant difference in acute respiratory failure, severe sepsis, 30-day readmission, postoperative stroke, EBL, length of stay, discharge status, or intraoperative aneurysm rupture between concurrent and nonconcurrent cases. Unplanned return to the OR and 30-day mortality were significantly lower in concurrent cases (odds ratio 0.55, 95% confidence interval 0.31–0.98, p = 0.0431, and odds ratio 0.81, p < 0.001, respectively), but concurrent cases had significantly longer procedure durations (odds ratio 21.73; p < 0.001).

CONCLUSIONS

Overall, there was a significant difference in the types of concurrent versus nonconcurrent cases, with more routine/elective cases for less sick patients scheduled in an overlapping fashion. After adjusting for patient demographics, procedure type, and clinical indicators, concurrent cases had longer procedure times, but equivalent patient outcomes, as compared with nonconcurrent vascular neurosurgical procedures.

ABBREVIATIONS AVM = arteriovenous malformation; ASA = American Society of Anesthesiologists; CM = cavernous malformation; EBL = estimated blood loss; ED = emergency department; LOS = length of stay; MGH = Massachusetts General Hospital; OR = operating room; UCSF = University of California, San Francisco.

OBJECTIVE

Concurrent surgeries, also known as “running two rooms” or simultaneous/overlapping operations, have recently come under intense scrutiny. The goal of this study was to evaluate the operative time and outcomes of concurrent versus nonconcurrent vascular neurosurgical procedures.

METHODS

The authors retrospectively reviewed 1219 procedures performed by 1 vascular neurosurgeon from 2012 to 2015 at the University of California, San Francisco. Data were collected on patient age, sex, severity of illness, risk of mortality, American Society of Anesthesiologists (ASA) status, procedure type, admission type, insurance, transfer source, procedure time, presence of resident or fellow in operating room (OR), number of co-surgeons, estimated blood loss (EBL), concurrent vs nonconcurrent case, severe sepsis, acute respiratory failure, postoperative stroke causing neurological deficit, unplanned return to OR, 30-day mortality, and 30-day unplanned readmission. For aneurysm clipping cases, data were also obtained on intraoperative aneurysm rupture and postoperative residual aneurysm. Chi-square and t-tests were performed to compare concurrent versus nonconcurrent cases, and then mixed-effects models were created to adjust for different procedure types, patient demographics, and clinical indicators between the 2 groups.

RESULTS

There was a significant difference in procedure type for concurrent (n = 828) versus nonconcurrent (n = 391) cases. Concurrent cases were more likely to be routine/elective admissions (53% vs 35%, p < 0.001) and physician referrals (59% vs 38%, p < 0.001). This difference in patient/case type was also reflected in the lower severity of illness, risk of death, and ASA class in the concurrent versus nonconcurrent cases (p < 0.01). Concurrent cases had significantly longer procedural times (243 vs 213 minutes) and more unplanned 30-day readmissions (5.7% vs 3.1%), but shorter mean length of hospital stay (11.2 vs 13.7 days), higher rates of discharge to home (66% vs 51%), lower 30-day mortality rates (3.1% vs 6.1%), lower rates of acute respiratory failure (4.3% vs 8.2%), and decreased 30-day unplanned returns to the OR (3.3% vs 6.9%; all p < 0.05). Rates of severe sepsis, postoperative stroke, intraoperative aneurysm rupture, and postoperative aneurysm residual were equivalent between the concurrent and nonconcurrent groups (all p values nonsignificant). Mixed-effects models showed that after controlling for procedure type, patient demographics, and clinical indicators, there was no significant difference in acute respiratory failure, severe sepsis, 30-day readmission, postoperative stroke, EBL, length of stay, discharge status, or intraoperative aneurysm rupture between concurrent and nonconcurrent cases. Unplanned return to the OR and 30-day mortality were significantly lower in concurrent cases (odds ratio 0.55, 95% confidence interval 0.31–0.98, p = 0.0431, and odds ratio 0.81, p < 0.001, respectively), but concurrent cases had significantly longer procedure durations (odds ratio 21.73; p < 0.001).

CONCLUSIONS

Overall, there was a significant difference in the types of concurrent versus nonconcurrent cases, with more routine/elective cases for less sick patients scheduled in an overlapping fashion. After adjusting for patient demographics, procedure type, and clinical indicators, concurrent cases had longer procedure times, but equivalent patient outcomes, as compared with nonconcurrent vascular neurosurgical procedures.

ABBREVIATIONS AVM = arteriovenous malformation; ASA = American Society of Anesthesiologists; CM = cavernous malformation; EBL = estimated blood loss; ED = emergency department; LOS = length of stay; MGH = Massachusetts General Hospital; OR = operating room; UCSF = University of California, San Francisco.

Concurrent surgeries, also known as “running two rooms” or simultaneous/overlapping operations, have come under intense scrutiny after a recent Boston Globe article drawing attention to this practice at Massachusetts General Hospital (MGH).1 In response to media and public concern, MGH has released public statements about the definition and use of concurrent, or overlapping, surgery in academic medical centers.6 More specifically, the hospital defines concurrent surgeries as those with either overlapping procedure or in-room times, and highlights the importance of overlapping surgeries in handling trauma, improving operating room (OR) and surgical team efficiency, promoting surgical education for residents and fellows, and allowing more timely access to high-demand surgical subspecialties.6 Leading the way for many institutions, MGH has updated and publicized its perioperative policy for concurrent staffing of 2 ORs.8

In light of this recent controversy, our hospital is currently reviewing its concurrent surgery policy. Like MGH and other academic medical centers, the goal of our institution is to provide the best care possible to all patients in a timely fashion and to educate the next generation of physicians and surgeons. Although high-profile opinion pieces in the surgical literature argue that concurrent surgery is safe and efficient,2,5 there is very little academic research on this topic. An abstract from a presentation at the American Association of Thoracic Surgery meeting reports no significant difference in operative times or patient outcomes in 1378 thoracic surgery procedures performed at the University of Virginia.9 However, there has been no analysis of concurrent surgery in neurosurgical procedures. Therefore, the goal of our study was to evaluate the operative times and patient outcomes of concurrent versus nonconcurrent vascular neurosurgical procedures performed by a single surgeon at our institution.

Methods

Data Collection

We collected data from the electronic medical records (APeX, EPIC Inc.) of all 1219 neurosurgical cases performed by 1 vascular neurosurgery attending physician from January 1, 2012, through December 31, 2015, at the University of California, San Francisco (UCSF). This study was performed for internal quality improvement purposes, and its publication was approved by our IRB.

For each case, patient demographics (age, sex) and clinical indicators (American Society of Anesthesiologists [ASA] physical status classification, severity of illness, risk of mortality) were collected. Severity of illness and risk of mortality were derived from the All Patient Refined-Diagnosis Related Groups licensed software (3M), and incorporated the following discharge data elements: principal diagnosis, principal procedure, secondary diagnoses, secondary procedures, age, sex, birth weight, discharge date, status of discharge, and days on mechanical ventilator.

We also obtained admission type (routine/elective, emergency, urgent), insurance status (Medicare, not Medicare), and source of transfer (emergency department [ED], physician referral, transfer from acute hospital, transfer from outside ED, other) for each patient. We recorded the performed procedure type (as defined by our UCSF-specific booking codes: for example, “carotid endarterectomy,” “crani for aneurysm clipping < 5 hours”), as well as the procedure date (year, day of the week), surgical procedure time (from surgical incision to wound closure), and total patient in-room time for each procedure. We also determined whether there was a neurosurgical fellow or resident present in the room and the total number of co-surgeons for each case. Concurrent cases were defined by whether the attending surgeon had another case whose procedure time overlapped by ≥ 1 second with a case in another room. In general, the surgeon had 2 scheduled ORs on Tuesdays and Fridays and performed simultaneous cases on those days, but had only 1 scheduled OR on Monday, during which he performed nonoverlapping cases.

Finally, we collected the following patient outcomes: acute respiratory failure (identified by discharge codes of ICD-9-CM 518.81 or ICD-10-CM J96.00 not present on admission), severe sepsis (identified by discharge codes ICD-9-CM 998.92 or ICD-10-CM R65.2 not present on admission), procedure duration (in minutes), 30-day unplanned readmission, 30-day unplanned return to OR, 30-day mortality, postoperative stroke, estimated blood loss (EBL, in milliliters), length of stay (LOS; in days), and discharge location (home vs not home). For aneurysm cases, we also obtained data on the intraoperative aneurysm rupture and postoperative residual aneurysm from our neurosurgical research database.

All 30-day readmissions were manually reviewed to determine if the readmission was related to the initial surgery. Each case of 30-day return to the OR was manually reviewed to determine if this was an unplanned return to the OR. Planned staged procedures (e.g., first-sided bypass followed by second-sided bypass for bilateral moyamoya disease) and shunt placement for patients with subarachnoid hemorrhage after ruptured aneurysm clip placement were excluded from unplanned 30-day returns to the OR. For postoperative stroke, a manual chart review was performed to ensure that this was a procedural-related stroke causing a new, unexpected neurological deficit after surgery. Subarachnoid hemorrhage cases with vasospasm occurring many days after the initial surgery, for instance, were not included.

Statistical Analysis

Data aggregation and statistical analysis was performed in SAS version 9.4 (SAS Institute, Inc.). Chi-square and student t-tests were used to compare individual variables between the overlapping and nonoverlapping cases. We then created multiple mixed-effects models to examine the effect of multiple variables (patient age, sex, insurance status, admission source, severity of illness, presence of fellow, presence of resident, and concurrent versus nonconcurrent cases) on the following outcome variables: acute respiratory failure, severe sepsis, procedure duration, 30-day return to OR, 30-day readmission, 30-day mortality, postoperative stroke, EBL, LOS, discharge status, and intraoperative aneurysm rupture. Note that we could not perform a multivariate model for postoperative aneurysm residual, due to the very low frequency of this outcome (only 2 cases of residual aneurysms across all routine/elective craniotomies for aneurysm clippings). In our mixed-effects models, the procedure type was chosen as the random effect, to account for differences between procedure types. All other variables were treated as fixed effects.

Results

Overall, 828 (68%) of 1219 vascular neurosurgical procedures were performed concurrently. There were no significant differences in age, sex, or insurance status between patients whose surgeries were performed concurrently versus those performed nonconcurrently (chi-square test; Table 1). Patients who underwent overlapping procedures had significantly lower ASA class, severity of illness, and risk of death than those who underwent nonoverlapping procedures (Table 1; all p < 0.0001). Consistent with this finding, concurrent cases were more likely to be routine/elective admissions, as compared with emergency/urgent admissions (53% routine/elective cases in concurrent vs 35% routine/elective cases in nonconcurrent surgeries, p < 0.0001), and physician referrals, as compared with ED admissions or transfers (59% physician referrals in concurrent vs 38% physician referrals in nonconcurrent surgeries; p < 0.0001; Table 1).

TABLE 1.

Patient demographics and clinical indicators of concurrent versus nonconcurrent vascular neurosurgical procedures at UCSF from 2012 to 2015

VariableConcurrent Cases (%)Nonconcurrent Cases (%)p Value*
No. of cases828391
Demographics
  Mean age ± SD (yrs)52.8 ± 17.053.4 ± 17.00.6063
  Sex0.6183
    Female531 (64)245 (63)
    Male297 (36)146 (37)
  Insurance
    Medicare251 (30)119 (30)0.9659
    Not Medicare577 (70)272 (70)
Clinical indicators
  ASA Class<0.0001
    I18 (2.4)3 (0.8)
    II304 (41)103 (29)
    III342 (46)180 (50)
    IV79 (11)69 (19)
    V3 (0.4)3 (0.8)
Severity of illness<0.0001
  Minor124 (15)58 (15)
  Moderate337 (41)98 (25)
  Major242 (29)126 (32)
  Extreme124 (15)108 (28)
Risk of death<0.0001
  Minor483 (58)188 (48)
  Moderate141 (17)74 (19)
  Major98 (12)40 (10)
  Extreme105 (13)88 (23)
Admission type<0.0001
  Routine/elective439 (53)135 (35)
  Emergency141 (17)98 (25)
  Urgent248 (30)158 (40)
Source of transfer<0.0001
  Physician referral490 (59)148 (38)
  ED46 (5.6)37 (9.5)
  Transfer (acute hospital)99 (12)72 (18)
  Transfer (other ED)187 (23)129 (33)
  Other6 (0.7)5 (1.3)

p values calculated using the chi-square test or t-test where appropriate for all categories under the particular variable. Boldface type indicates statistical significance.

Table 2 shows the different case mix in concurrent versus nonconcurrent cases, with a larger percentage of craniotomies for arteriovenous malformation (AVMs) and cavernous malformation (CM) resections noted in the overlapping case group (p < 0.001; Table 2). There was no difference in concurrent versus nonconcurrent case distribution from 2012 to 2015 (p = 0.5452; Table 2). The higher percentage of concurrent cases on Tuesdays and Fridays (44% and 42%, respectively; p < 0.001; Table 2) reflected the surgeon's OR block time: 2 ORs on Tuesdays/Fridays and only 1 OR allocated on Mondays. The number of co-surgeons was slightly higher for nonconcurrent as compared with concurrent surgeries (2.56 vs 2.35), and nonconcurrent cases were more likely to have a neurosurgical fellow present in the OR (47% vs 34%, p < 0.0001; Table 2). Surgical procedure time and in-room time were both slightly higher for concurrent than nonconcurrent cases (243 vs 213 minutes and 326 vs 293 minutes, respectively; p < 0.0001; Table 2).

TABLE 2.

Procedure characteristics and metrics of concurrent versus nonconcurrent vascular neurosurgical procedures at UCSF from 2012 to 2015

Procedure VariableConcurrent Cases (%)Nonconcurrent Cases (%)p Value
Procedure Type<0.001
  Craniotomy for aneurysm clipping < 5 hrs340 (41)167 (43)
  Craniotomy for AVM resection < 5 hrs140 (17)44 (11)
  Craniotomy for CM resection94 (11)24 (6.1)
  Craniotomy for EC-IC low-flow bypass43 (5.2)16 (4.1)
  Laminotomy for AVM resection27 (3.3)5 (1.3)
  VP shunt insertion (no GS assist)27 (3.3)22 (5.6)
  Carotid endarterectomy24 (2.9)11 (2.8)
  Craniotomy for aneurysm clipping > 7 hrs20 (2.4)11 (2.8)
  Craniotomy for EC-IC high-flow bypass13 (1.6)7 (1.8)
  Craniotomy for cranioplasty11 (1.3)1 (0.3)
Year of procedure0.5452
  2012125 (15)69 (18)
  2013223 (27)110 (28)
  2014250 (30)106 (27)
  2015230 (28)106 (27)
Day of week of procedure<0.001
  Monday103 (12)169 (43)
  Tuesday364 (44)78 (20)
  Wednesday10 (1.2)53 (14)
  Thursday0 (0)20 (5.1)
  Friday351 (42)54 (14)
  Saturday0 (0)11 (2.8)
  Sunday0 (0)6 (1.5)
Mean procedure time ± SD (mins)243 ± 109213 ± 114<0.0001
Mean in-room time ± SD (mins)326 ± 119293 ± 129<0.0001
Mean no. of co-surgeons ± SD2.35 ± 0.62.56 ± 0.7<0.0001
Presence of fellow<0.0001
  Yes282 (34)184 (47)
  No546 (66)207 (53)
Presence of resident0.2903
  Yes692 (84)336 (86)
  No136 (16)55 (14)

EC-IC = extracranial–intracranial; GS = general surgery; VP = ventriculoperitoneal.

Only the top 10 most common procedure types (as determined by the booking code) for this surgeon are shown. p values were calculated using the chi-square test or t-test where appropriate for all categories under the particular variable. Boldface type indicates statistical significance.

As shown in Table 3, concurrent cases had significantly shorter LOS (11.2 vs 13.7 days), higher rates of discharge to home (66% vs 51%), lower 30-day mortality rates (3.1% vs 6.1%), lower rates of acute respiratory failure (4.3% vs 8.2%), and decreased 30-day unplanned returns to the OR (3.3% vs 6.9%; all p < 0.05; Table 3). There was no statistically significant difference in rates of severe sepsis (1.4% vs 2.0%), postoperative stroke (4.8% vs 3.3%), or EBL (354 vs 364 ml) between the 2 groups (Table 3). However, there was a higher rate of unplanned 30-day readmissions in the concurrent versus nonconcurrent cases (5.7% vs 3.1%, p = 0.0477; Table 3).

TABLE 3.

Patient outcomes of concurrent versus nonconcurrent vascular neurosurgical procedures at UCSF from 2012 to 2015

VariableConcurrent Cases (%)Nonconcurrent Cases (%)p Value*
Mean LOS ± SD (days)11.2 ± 15.213.7 ± 13.20.0042
Mean EBL ± SD (ml)354 ± 508364 ± 5350.7496
Discharge disposition<0.0001
  Home, self or home care542 (66)201 (51)
  Skilled nursing facility76 (9.2)50 (13)
  Acute rehabilitation facility134 (16)69 (18)
  Deceased24 (2.9)24 (6.1)
  Other acute care hospital36 (4.4)38 (9.7)
  Other (e.g., hospice, LTAC facility, jail)16 (1.9)9 (2.3)
30-day mortality
  Deceased26 (3.1)24 (6.1)0.0138
  Survived802 (97)367 (94)
30-day unplanned readmission
  Yes47 (5.7)12 (3.1)0.0477
  No781 (94)379 (97)
30-day unplanned return to OR0.0039
  Yes27 (3.3)27 (6.9)
  No801 (97)364 (93)
Complication: acute respiratory failure0.0064
  Yes36 (4.3)32 (8.2)
  No792 (96)359 (92)
Complication: severe sepsis0.4439
  Yes12 (1.4)8 (2.0)
  No816 (99)383 98
Complication: postop stroke0.2288
  Yes40 (4.8)13 (3.3)
  No788 (95)378 (97)

LTAC = long-term acute care.

p values were calculated using the chi-square test or t-test where appropriate for all categories under the particular variable. Boldface type indicates statistical significance.

Given the significant difference in case mix/type between the 2 groups, we next focused specifically on routine/elective craniotomy for aneurysm clippings (n = 252). There were no statistically significant differences in surgical procedure time, LOS, EBL, discharge status, 30-day mortality, 30-day unplanned readmission, 30-day unplanned return to the OR, acute respiratory failure, severe sepsis, postoperative stroke, intraoperative aneurysm rupture, or postoperative aneurysm residual in concurrent versus nonconcurrent routine/elective aneurysm cases (Table 4).

TABLE 4.

Surgical time and patient outcomes of concurrent versus nonconcurrent routine/elective craniotomy for aneurysm clip placements (i.e., nonruptured aneurysm clippings; n = 252) at UCSF from 2012 to 2015

VariableConcurrent Cases (%)Nonconcurrent Cases (%)p Value*
Mean LOS ± SD (days)6.4 ± 5.86.4 ± 6.40.9474
Mean surgical procedure time ± SD (mins)264 ± 89263 ± 830.9226
Mean EBL ± SD (ml)372 ± 365402 ± 3050.5837
Discharge disposition0.3157
  Home, self or home care164 (83)41 (76)
  Skilled nursing facility9 (4.6)3 (5.6)
  Acute rehabilitation facility17 (8.6)9 (16.7)
  Deceased6 (3.0)0 (0)
  Other acute care facility1 (0.5)0 (0)
  Other (e.g., hospice, LTAC, jail)1 (0.5)1 (1.9)
30-day mortality0.1954
  Deceased6 (3.3)0 (0)
  Survived192 (97)54 (100)
30-day unplanned readmission0.5679
  Yes17 (8.6)6 (11)
  No81 (91)48 (89)
30-day unplanned return to OR0.2334
  Yes5 (2.5)0 (0)
  No193 (98)54 (100)
Complication: acute respiratory failure0.3078
  Yes10 (5.1)1 (2.8)
  No188 (85)53 (98)
Complication: severe sepsis0.4584
  Yes2 (1)0 (0)
  No196 (99)54 (100)
Complication: postop stroke0.2683
  Yes16 (8.1)2 (3.7)
  No182 (92)52 (96)
Intraop aneurysm rupture0.6102
  Yes11 (5.6)4 (7.4)
  No187 (94)50 (93)
Postop aneurysm residual0.4584
  Yes2 (1)0 (0)
  No196 (99)54 (100)

p values calculated using the chi-square test or t-test where appropriate.

Finally, mixed-effects models showed that, after controlling for procedure type, there was no difference in acute respiratory failure, severe sepsis, unplanned 30-day readmission, postoperative stroke, EBL, LOS, discharge status, or intraoperative aneurysm rupture between concurrent and nonconcurrent cases (Table 5). Concurrent cases had longer procedure durations (odds ratio = 21.73; p < 0.001), but rates of 30-day unplanned return to OR and 30-day mortality were lower in concurrent versus nonconcurrent cases (odds ratio 0.55, 95% confidence interval [CI] 0.31–0.98, p = 0.0431, and odds ratio = 0.81, p < 0.0001, respectively; Table 5).

TABLE 5.

Mixed-effects model (with procedure type as random effect) for the effect of patient/clinical factors and concurrent vs nonconcurrent cases on 11 different outcomes

EffectAgeFemale (vs male)Medicare (vs other)Routine/Elective (vs emergency/urgent)Illness Severity: Extreme/Major (vs minor/moderate)Presence of Fellow (yes/no)Presence of Resident (yes/no)Concurrent Cases (vs nonconcurrent)
Acute respiratory failure1.03 (1.01–1.06);

p = 0.0073
0.66 (0.38–1.13);

p = 0.1281
1.18 (0.57–2.43);

p = 0.6631
1.61 (0.88–2.96);

p = 0.126
84.26 (11.3–628.8);

p < 0.001
1.37 (0.76–2.44);

p = 0.2921
1.80 (0.77–4.23);

p = 0.177
0.71 (0.41–1.22);

p = 0.2133
Severe sepsis (2)1.05*;

p < 0.001
1.24 (0.46–3.36);

p = 0.6678
1.81 (0.52–6.34);

p = 0.3551
1.11 (0.38–3.25);

p = 0.8483
>999*;

p = 0.9941
1.31 (0.47–3.62);

p = 0.6034
1.31 (0.33–5.26);

p = 0.7054
1.04 (0.40–2.7);

p = 0.9306
Procedure duration (min)−0.60;

p <0.001
−2.28;

p = 0.6664
−12.57;

p = 0.0754
11.34;

p = 0.0616
20.13;

p = 0.0008
27.65;

p < 0.0001
34.54;

p < 0.0001
21.73;

p <0.001
Unplanned return to OR1.00 (0.98–1.03);

p = 0.6958
0.82 (0.46–1.45);

p = 0.695
1.18 (0.54–2.58);

p = 0.671
0.99 (0.50–1.97);

p = 0.9765
3.10(1.48–6.50);

p = 0.0028
1.22 (0.66–2.26);

p = 0.5187
1.16 (0.50–2.69);

p = 0.7338
0.55 (0.31–0.98);

p = 0.0431
Readmission1.00 (0.98–1.02);

p = 0.8278
1.12 (0.63–1.99);

p = 0.7116
0.57 (0.27–1.21);

p = 0.1416
1.61 (0.83–3.1);

p = 0.1586
0.50 (0.25–0.98);

p = 0.0437
1.41 (0.78–2.54);

p = 0.2505
2.13 (0.84–5.39);

p = 0.1107
1.72 (0.88–3.37);

p = 0.113
30-day mortality1.01*;

p < 0.001
0.90*;

p < 0.001
0.81 (0.36–1.84);

p = 0.6189
0.68 (0.30–1.55);

p = 0.3532
>999*;

p = 0.9951
1.48 (0.77–2.84);

p = 0.2426
2.34 (0.84–6.53);

p = 0.1032
0.81*;

p < 0.0001
Postop stroke1.00 (0.98–1.03);

p = 0.9019
1.06 (0.55–2.03);

p = 0.8641
0.53 (0.23–1.22);

p = 0.1355
3.91 (2.07–7.38);

p <0.001
57.41 (13.35–247);

p < 0.001
0.94 (0.47–1.90);

p = 0.8725
1.13 (0.44–2.91);

p = 0.7985
1.68 (0.84–3.38);

p = 0.1418
EBL (ml)−0.01;

p = 0.9919
−0.39;

p = 0.9905
−28.17;

p = 0.5164
−4.83;

p = 0.8979
154.9;

p < 0.001
115.5;

p = 0.0011
101.4;

p = 0.0299
29.07;

p = 0.3937
LOS (days)0.08;

p = 0.006
0.23;

p = 0.7621
3.32;

p = 0.0014
−6.02;

p <0.001
9.45;

p < 0.001
−1.85;

p = 0.0276
−0.03;

p = 0.9766
0.36;

p = 0.6549
Discharge (home vs not home)0.97 (0.96–0.98);

p = 0.001
0.98 (0.73–1.32);

p = 0.9015
1.18 (0.80–1.75);

p = 0.4027
2.46 (1.77–3.42);

p < 0.0001
0.16 (0.11–0.21);

p < 0.0001
0.90 (0.66–1.25);

p = 0.5379
1.01 (0.66–1.56);

p = 0.952
1.30 (0.96–1.77);

p = 0.0899
Intraop aneurysm rupture1.04 (0.97–1.11);

p = 0.3048
0.68 (0.20–2.39);

p = 0.5497
4.53 (0.98–20.88);

p = 0.0527
NA5.28 (1.65–16.9);

p = 0.0052
1.31 (0.40–4.34);

p = 0.6522
1.03 (0.18–5.83);

p = 0.972
0.77 (0.22–2.76);

p = 0.6865

NA = not applicable.

For most variables, the odds ratio is shown, followed by the 95% confidence interval (in parentheses) and the p value. For numerical outcomes (procedure duration, EBL, and LOS), the coefficient is shown, followed by the p value. Boldface type indicates statistical significance.

No 95% confidence interval can be calculated because standard error is 0.

Intraoperative aneurysm rupture analysis was only performed for routine/elective craniotomies for aneurysm clippings.

Our mixed-effects models also showed that increasing patient age was associated with higher rates of acute respiratory failure, severe sepsis, 30-day mortality, discharge to location other than home, and longer lengths of stay (Table 5). Higher severity of illness was significantly correlated with higher rates of acute respiratory failure, return to OR, postoperative stroke, discharge to location other than home, intraoperative aneurysm rupture, longer procedure duration, longer LOS, and higher EBL (Table 5). Of note, the presence of a fellow or resident in the OR was associated with longer procedure duration and higher EBL, but did not negatively impact any other patient outcomes (Table 5).

Discussion

In summary, our work shows a significant difference in the types of concurrent versus nonconcurrent cases performed by a single vascular neurosurgeon at our institution, with more routine/elective cases for less sick patients scheduled in an overlapping fashion. After controlling for procedure type and patient, clinical, and operative factors in mixed-effects models, we found slightly longer operative times, but equivalent patient outcomes, in concurrent versus nonconcurrent cases (Table 5). Despite longer procedure times in concurrent cases, we did not find higher rates of acute respiratory failure (as might be expected from longer anesthesia times) or severe sepsis (which could be correlated with increased OR traffic and procedure duration). It is important to note that the link between OR traffic/time and infection is controversial,3 and a recent prospective trial failed to demonstrate that reduction in OR traffic leads to lower rates of surgical site infections in neurosurgical procedures.4

Our overall conclusions are consistent with the only prior surgical analysis of concurrent thoracic surgeries,9 as well as the unpublished review of concurrent cases available on the MGH website.7 In their review of 418 cases, MGH did not find significant differences in mortality, morbidity, unplanned readmission, or unplanned return to the OR or intensive care unit.7 However, they did not identify the types of cases or characteristics of patients, and did not perform rigorous multivariate analyses controlling for patient, clinical, and operative factors, as we have done here. Nevertheless, the combination of our work and these prior studies9 suggests that concurrent surgery may be performed safely and efficiently in an academic setting.

Simultaneous case scheduling allows for more timely access to high-demand surgeries and promotes education and independence of surgical residents and fellows. However, it is important that concurrent surgeries adhere to the following guidelines. First, patients must be appropriately and explicitly informed of concurrent surgeries and the extent of participation of trainees in their operation.5 This is an integral part of the UCSF consent form, and has recently been updated to reflect that the surgeon may perform operations simultaneously. Second, concurrent surgeries should be carefully scheduled and well planned. Surgeons should consider scheduling routine and straightforward, rather than complex, cases in overlapping ORs. They should also ensure that they have trainees with the appropriate skills for the particular case and that attending surgeon backup is available immediately when needed. Our institution requires that the backup surgeon is available on campus throughout the duration of the procedure, but he or she may perform his or her own case during that time period. The backup surgeon must be listed on the surgical booking form, and the identity and location of the backup surgeon is discussed by the OR team during the timeout at the beginning of each case. The identity of the backup surgeon must also be disclosed to the patient. And finally, surgeons should prospectively collect their data on concurrent cases in a standardized fashion and review outcomes to demonstrate that they can perform concurrent surgery safely.

An important limitation of our work is that this is a retrospective, single-site, single-provider study. Given the low frequency of several outcome variables (especially severe sepsis, 30-day mortality, postoperative residual aneurysm, and intraoperative aneurysm rupture), our study may be underpowered to detect a significant difference between the concurrent and nonconcurrent surgery groups. We acknowledge that a larger multisite or multiprovider study could provide more generalizable and robust data, although we intentionally examined a single surgeon's cases for purposes of patient, clinical, and surgical management homogeneity. In addition, we understand that institutions that operate without residents or fellows, for example, may not be able to safely and efficiently perform simultaneous vascular neurosurgeries, as described here. However, given the lack of data on concurrent neurosurgery, we believe that this study is an important step toward understanding the effect of concurrent surgery on neurosurgical practice, demonstrating its safety, and increasing patient access to highly specialized surgical experts.

Conclusions

To the best of our knowledge, our study represents the first report of the operative time and patient outcomes in concurrent versus nonconcurrent neurosurgical procedures. We found a significant difference in the types of concurrent versus nonconcurrent cases, with more routine/elective cases for less sick patients scheduled in an overlapping fashion. Given the single-site and single-provider nature of this study, we caution that our results may not be generalizable to other providers or institutions. However, our retrospective analysis provides Level III evidence that there are slightly longer operative times, but equivalent patient outcomes, in concurrent versus nonconcurrent procedures performed by a single vascular neurosurgeon. We therefore believe that concurrent surgery may be performed safely and efficiently in these specific circumstances at our institution, and we are updating our institutional concurrent surgery policy to reflect these findings.

Acknowledgments

Dr. Zygourakis is supported by a research fellowship from the UCSF Center for Healthcare Value.

Disclosures

Dr. Zygourakis has received educational expenses from Stryker and travel stipends to attend resident education spine courses from Globus and Nuvasive. Dr. Lawton has received consulting fees from Zeiss and Stryker.

Author Contributions

Conception and design: Lawton, Zygourakis, Lobo. Acquisition of data: Zygourakis, Lee, Barba. Analysis and interpretation of data: Zygourakis, Lee, Barba. Drafting the article: Zygourakis. Critically revising the article: Lawton, Zygourakis. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Lawton. Statistical analysis: Lee. Administrative/technical/material support: Lawton, Lobo. Study supervision: Lawton.

References

  • 1

    Abelson J, Saltzman J, Kowalczyk L, Allen S: Clash in the name of care. Boston Globe Oct 25 2015. (https://apps.bostonglobe.com/spotlight/clash-in-the-name-of-care/story/) [Accessed November 15, 2016]

    • Search Google Scholar
    • Export Citation
  • 2

    Beasley GM, Pappas TN, Kirk AD: Procedure delegation by attending surgeons performing concurrent operations in academic medical centers: balancing safety and efficiency. Ann Surg 261:10441045, 2015

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 3

    Birgand G, Saliou P, Lucet JC: Influence of staff behavior on infectious risk in operating rooms: what is the evidence?. Infect Control Hosp Epidemiol 36:93106, 2015

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 4

    Bohl M, Clark JC, Oppenlander ME, Meeusen AJ, Budde A, Porter RW, : The Barrow Randomized OR Traffic (BRITE) Trial: The effect of OR traffic on infection rates. Neurosurgery 62:Suppl 1 196197, 2015. (Abstract)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 5

    Langerman A: Concurrent surgery and informed consent. JAMA Surg 151:601602, 2016

  • 6

    Massachusetts General Hospital: About Concurrent/Overlapping Surgery Fact Sheet. (http://www.massgeneral.org/overlapping-surgery/about.aspx) [Accessed November 15, 2016]

    • Export Citation
  • 7

    Massachusetts General Hospital: Monitoring Outcomes for Procedural Overlap Surgeries at MGH. (http://www.massgeneral.org/news/assets/pdf/MonitoringOutcomes.pdf) [Accessed November 15, 2016]

    • Export Citation
  • 8

    Massachusetts General Hospital: Perioperative Policy for Concurrent Surgical Staffing of Two Rooms. (http://www.massgeneral.org/news/assets/pdf/MGHConcurrentSurgeryPolicy.pdf) [Accessed November 15, 2016]

    • Export Citation
  • 9

    Yount KW, Gillen JR, Kron IL, Kern JA, Kozower BD, Ailawadi G, : Attendings' performing simultaneous operations in academic cardiothoracic surgery does not increase operative duration or negatively affect patient outcomes. 94th AATS Annual Meeting Toronto April 26–30, 2014 (Abstract) (http://aats.org/annualmeeting/Program-Books/2014/2.cgi) [Accessed November 15, 2016]

    • Export Citation

If the inline PDF is not rendering correctly, you can download the PDF file here.

Contributor Notes

Correspondence Michael T. Lawton, Department of Neurological Surgery, University of California, San Francisco, 505 Parnassus Ave., Rm. M779, San Francisco, CA 94143. email: michael.lawton@ucsf.edu.

ACCOMPANYING EDITORIAL DOI: 10.3171/2016.9.JNS161500.

INCLUDE WHEN CITING Published online January 20, 2017; DOI: 10.3171/2016.6.JNS16822.

Disclosures Dr. Zygourakis has received educational expenses from Stryker and travel stipends to attend resident education spine courses from Globus and Nuvasive. Dr. Lawton has received consulting fees from Zeiss and Stryker.

  • 1

    Abelson J, Saltzman J, Kowalczyk L, Allen S: Clash in the name of care. Boston Globe Oct 25 2015. (https://apps.bostonglobe.com/spotlight/clash-in-the-name-of-care/story/) [Accessed November 15, 2016]

    • Search Google Scholar
    • Export Citation
  • 2

    Beasley GM, Pappas TN, Kirk AD: Procedure delegation by attending surgeons performing concurrent operations in academic medical centers: balancing safety and efficiency. Ann Surg 261:10441045, 2015

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 3

    Birgand G, Saliou P, Lucet JC: Influence of staff behavior on infectious risk in operating rooms: what is the evidence?. Infect Control Hosp Epidemiol 36:93106, 2015

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 4

    Bohl M, Clark JC, Oppenlander ME, Meeusen AJ, Budde A, Porter RW, : The Barrow Randomized OR Traffic (BRITE) Trial: The effect of OR traffic on infection rates. Neurosurgery 62:Suppl 1 196197, 2015. (Abstract)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 5

    Langerman A: Concurrent surgery and informed consent. JAMA Surg 151:601602, 2016

  • 6

    Massachusetts General Hospital: About Concurrent/Overlapping Surgery Fact Sheet. (http://www.massgeneral.org/overlapping-surgery/about.aspx) [Accessed November 15, 2016]

    • Export Citation
  • 7

    Massachusetts General Hospital: Monitoring Outcomes for Procedural Overlap Surgeries at MGH. (http://www.massgeneral.org/news/assets/pdf/MonitoringOutcomes.pdf) [Accessed November 15, 2016]

    • Export Citation
  • 8

    Massachusetts General Hospital: Perioperative Policy for Concurrent Surgical Staffing of Two Rooms. (http://www.massgeneral.org/news/assets/pdf/MGHConcurrentSurgeryPolicy.pdf) [Accessed November 15, 2016]

    • Export Citation
  • 9

    Yount KW, Gillen JR, Kron IL, Kern JA, Kozower BD, Ailawadi G, : Attendings' performing simultaneous operations in academic cardiothoracic surgery does not increase operative duration or negatively affect patient outcomes. 94th AATS Annual Meeting Toronto April 26–30, 2014 (Abstract) (http://aats.org/annualmeeting/Program-Books/2014/2.cgi) [Accessed November 15, 2016]

    • Export Citation

Metrics