Outcomes of Elective Cerebral Aneurysm Treatment Performed by Attending Neurosurgeons after Night Work

Outcomes of Elective Cerebral Aneurysm Treatment Performed by Attending Neurosurgeons after Night... Abstract BACKGROUND The association between long work hours and outcomes among attending surgeons remains an issue of debate. OBJECTIVE To investigate whether operating emergently the night before an elective case was associated with inferior outcomes among attending neurosurgeons. METHODS We executed a cohort study with unruptured cerebral aneurysm patients, who underwent endovascular coiling or surgical clipping from 2009 to 2013 and were registered in the Statewide Planning and Research Cooperative System database. We investigated the association of treatment by surgeons performing emergency procedures the night before with outcomes of elective cerebral aneurysm treatment using an instrumental variable analysis. RESULTS Overall, 4700 patients underwent treatment for unruptured cerebral aneurysms. There was no difference in inpatient mortality (adjusted difference, –0.7%; 95% confidence interval [CI], –1.4% to 0.02%), discharge to a facility (adjusted difference, –0.1%; 95% CI, –1.2% to 1.2%), or length of stay (adjusted difference, –0.58; 95% CI, –1.66 to 0.50) between patients undergoing elective cerebral aneurysm treatment by surgeons who performed emergency procedures the night before, and those who did not. CONCLUSION Using a comprehensive patient cohort in New York State for elective treatment of unruptured cerebral aneurysms, we did not identify an association of treatment by surgeons performing emergency procedures the night before, with mortality, discharge to a facility, or length of stay. Our study had 80% power to detect differences in mortality (our primary outcome), as small as 4.1%. The results of the present study do not support the argument for regulation of attending work hours. Unruptured cerebral aneurysm, Fatigue, Work hours, Instrumental variable, SPARCS ABBREVIATIONS ABBREVIATIONS 2SLS 2-stage least squares CI confidence interval ICD-9-CM International Classification of Disease-9-Clinical Modification LOS length of stay SD standard deviation SPARCS Statewide Planning and Research Cooperative System There is heightened public awareness and ongoing debate about the effects of sleep deprivation and fatigue on physician performance.1-6 Some studies have demonstrated that medical resident fatigue is associated with a higher rate of errors including percutaneous needle sticks and laceration injuries, as well as postcall motor vehicle crashes.2,7,8 Others have not replicated these findings among surgical residents.9,10 However, legislative reform and comprehensive regulatory efforts have transformed graduate medical education and have imposed strict work hour restrictions on trainees.11 Recent randomized trials demonstrated that loose adherence to these restrictions was not associated with increased complications, fueling further debate on this issue.12 Unlike residents, there is no restriction on the work hours of attending physicians in the United States.4,13 Prior studies on the effect of sleep deprivation of attending surgeons on outcomes have had conflicting results. Single-center studies14 have demonstrated an association with inferior outcomes, whereas multicenter Canadian investigations15,16 have identified no difference. The generalization of the latter studies in the United States is limited given the differences in the practice of medicine between the 2 countries. Additionally, these prior investigations have focused on general surgeons, who frequently subspecialize to only provide emergency care. However, neurosurgeons are very commonly called upon to treat trauma or other neurological pathology emergently, while maintaining a busy elective surgical schedule. No prior study has attempted to answer this question in a comprehensive cohort of neurosurgeons. We used the New York Statewide Planning and Research Cooperative System (SPARCS)17 to study the association of performing emergency procedures the night before an elective case with mortality, discharge to a facility, and length of stay for patients being treated for unruptured cerebral aneurysms. We used an instrumental variable analysis to simulate the effect of randomization. METHODS New York SPARCS The cohort included all patients with unruptured cerebral aneurysms who were registered in the SPARCS (New York State Department of Health, Albany, New York)17 database between 2009 and 2013. More information on SPARCS can be found at https://www.health.ny.gov/statistics/sparcs/. Cohort Definition This study was approved by the institutional review board and consent was waived due to deidentified data. In order to establish the cohort of patients, we used International Classification of Disease-9-Clinical Modification (ICD-9-CM) codes to select patients in the database who underwent surgical clipping (ICD-9-CM code 39.51) or endovascular coiling (ICD-9-CM code 39.52 [should also have a code 88.41 and no 39.51 during the same hospitalization], 39.72, 39.75, 39.76) for unruptured (ICD-9-CM code 437.3) cerebral aneurysms between 2009 and 2013. Outcome Variables Mortality during the hospitalization for unruptured cerebral aneurysm clipping or endovascular treatment was the primary outcome variable. Length of stay (LOS) during the initial hospitalization, and the rate of discharge to a facility (any facility other than the patient's home) were secondary outcomes. Exposure Variables Performing emergency procedures the night before (12 am to 6 am) elective cerebral aneurysm treatment was the primary exposure variable. Covariates (Table, Supplemental Digital Content 1) used in the regressions were age, race (African–American, Hispanic, Asian, Caucasian, other), gender, insurance (private, Medicare, Medicaid, uninsured, other), and procedure performed (clipping, endovascular intervention). Additional comorbidities used in the regressions were smoking, diabetes mellitus, chronic lung disease, hypercholesterolemia, hypertension, peripheral vascular disease, coronary artery disease, congestive heart failure, history of stroke, obesity, transient ischemic attack, chronic renal failure, alcohol abuse, and coagulopathy. Only variables identified by the database as “present on admission” were included in the patient's comorbidity profile. Statistical Analysis The association of the exposure variable of interest with our outcomes was examined in multivariable analyses. Physicians performing elective procedures after night work are nonrandomly selected. They are expected to be younger, since they are still actively involved in call, and less subspecialized (although most neurosurgeons take call regardless of subspecialty focus), which can be associated with inferior outcomes. In order to account for this unmeasured confounding, we used an instrumental variable analysis.18 The rate of performing emergency overnight procedures (at the provider level) was used as an instrument. Similar rates have been used before to create pseudorandomization on the exposure variable.19 This approach allows us to compare the outcomes of elective cases performed by surgeons operating the night before vs those performed by surgeons not operating the prior night, while controlling for unmeasured confounding. The methodology of instrumental variables has become increasingly popular over the past decade. This method provides a way to estimate a treatment effect without bias despite the presence of confounding (ie, unknown baseline functional clinical status of patients). Instrumental variables do indeed exist. We make them all the time when we run a randomized trial. Randomization is an example of an instrumental variable, because the flip of a coin or a computer-generated random number determines the treatment a subject is assigned and therefore the treatment a subject receives. However, when health services or clinical researchers think of instrumental variables, they are usually conceiving of some natural experiment. Favored choices for instrumental variables in the medical literature so far have been geographic distances, or geographic rates. We consider a good instrument a variable that is only associated with the outcome only through the exposure of interest.19 We used a 2-stage least squares (2SLS) approach. In the first stage of the 2SLS analysis, the F statistic was 30. This is compatible with a strong instrument (F statistic >10).18 In sensitivity analysis, the main regressions were repeated separately in subgroups of patients undergoing only endovascular coiling, or only surgical clipping. There was no change in the direction of the observed associations, and therefore we do not report these results further. A probit regression was used for the categorical outcomes (mortality, discharge to a facility),20 and a linear regression for the linear outcomes (LOS). In additional to sensitivity analysis, we repeated all analyses after logarithmic transformation of LOS to account for potentially positively skewed distribution of this variable. There was no change in the direction of the observed associations, and therefore we do not report these results further. The covariates used for risk adjustment in these models were age, race, gender, hospital ID, insurance, and all the previous comorbidities. We used the marginal effects of our independent variables to obtain interpretable coefficients. Regression diagnostics were performed on all models. All results are based on 2-sided tests, and the level of statistical significance was set at 0.05. This study, based on 4700 patients, has sufficient power (80%) at a 5% type I error rate to detect differences in mortality, as small as 4.1%. We used Stata version 13 (StataCorp, College Station, Texas) to perform the analyses. RESULTS Patient Characteristics Overall, there were 4700 patients being treated for unruptured cerebral aneurysms (mean age was 55.0 years, with 76.7% females) who were registered in SPARCS. A total of 2541 (54.1%) underwent elective aneurysm treatment by neurosurgeons who operated the night before; whereas 2159 (45.9%) underwent treatment by neurosurgeons who did not operate the night before. The baseline characteristic of the 2 cohorts can be seen in Table 1. TABLE 1. Patient Characteristics   Total n = 4700  Neurosurgeons operating after night work n = 2159  Neurosurgeons operating without prior night work n = 2541    Mean  SD  Mean  SD  Mean  SD  Age  55.00  14.12  54.09  14.82  55.82  13.56    n  %  n  %  n  %  Female gender  3605  76.7%  1662  77.0%  1906  75.0%  Comorbidities              Diabetes  484  10.3%  225  10.4%  267  10.5%  Smoking  912  19.4%  438  20.3%  458  18.7%  Obesity  291  6.2%  123  5.7%  162  6.6%  Ischemic stroke  24  0.5%  9  0.4%  18  0.7%  Coronary artery disease  348  7.4%  168  7.8%  185  7.3%  Chronic lung disease  804  17.1%  358  16.6%  462  18.2%  Congestive heart failure  94  2.0%  43  2.0%  51  2.0%  Coagulopathy  38  0.8%  15  0.7%  28  1.1%  Chronic renal failure  19  0.4%  6  0.3%  15  0.6%  Hypertension  2477  52.7%  1080  50.0%  1319  53.8%  Hypercholesterolemia  1344  28.6%  600  27.8%  701  28.6%  Alcohol  89  1.9%  37  1.7%  51  2.1%  Peripheral vascular disease  132  2.8%  54  2.5%  74  3.0%    Total n = 4700  Neurosurgeons operating after night work n = 2159  Neurosurgeons operating without prior night work n = 2541    Mean  SD  Mean  SD  Mean  SD  Age  55.00  14.12  54.09  14.82  55.82  13.56    n  %  n  %  n  %  Female gender  3605  76.7%  1662  77.0%  1906  75.0%  Comorbidities              Diabetes  484  10.3%  225  10.4%  267  10.5%  Smoking  912  19.4%  438  20.3%  458  18.7%  Obesity  291  6.2%  123  5.7%  162  6.6%  Ischemic stroke  24  0.5%  9  0.4%  18  0.7%  Coronary artery disease  348  7.4%  168  7.8%  185  7.3%  Chronic lung disease  804  17.1%  358  16.6%  462  18.2%  Congestive heart failure  94  2.0%  43  2.0%  51  2.0%  Coagulopathy  38  0.8%  15  0.7%  28  1.1%  Chronic renal failure  19  0.4%  6  0.3%  15  0.6%  Hypertension  2477  52.7%  1080  50.0%  1319  53.8%  Hypercholesterolemia  1344  28.6%  600  27.8%  701  28.6%  Alcohol  89  1.9%  37  1.7%  51  2.1%  Peripheral vascular disease  132  2.8%  54  2.5%  74  3.0%  View Large Outcomes Patients operated on by surgeons after night work demonstrated 1.3% mortality (33 patients), 17.8% discharge to rehabilitation (436 patients), and on average a 5.5-day LOS (standard deviation [SD] 15.0). Patients operated by surgeons who did not operate the night before demonstrated 1.5% mortality (32), 18.1% discharge to rehabilitation (391), and on average a 3.9-day LOS (SD 7.0). Instrumental Variable Analysis Using a probit regression with instrumental variable analysis, we did not identify an association of undergoing treatment by surgeons performing emergency procedures the night before with mortality (adjusted difference, –0.7%; 95% confidence interval [CI], –1.4% to 0.02%) after elective cerebral aneurysm treatment (Table 2). TABLE 2. Multivariable Models Examining the Association of Performing Emergency Procedures the Night Before an Elective Case with Outcomes of Unruptured Cerebral Aneurysm Treatment   Inpatient Mortalitya    Discharge to rehabilitationa    Length of stayb      Adjusted difference (95% CI)  P-value  Adjusted difference (95% CI)  P-value  Adjusted difference (95% CI)  P-value  Instrumental variable analysisc  –0.7% (–1.4% to 0.02%)  .684  –0.1% (–1.2% to 1.2%)  .970  –0.58 (–1.66 to 0.50)  .994    Inpatient Mortalitya    Discharge to rehabilitationa    Length of stayb      Adjusted difference (95% CI)  P-value  Adjusted difference (95% CI)  P-value  Adjusted difference (95% CI)  P-value  Instrumental variable analysisc  –0.7% (–1.4% to 0.02%)  .684  –0.1% (–1.2% to 1.2%)  .970  –0.58 (–1.66 to 0.50)  .994  CI: confidence intervals; OR: odds ratio. aBased on a probit model. bAll regressions were based on linear models. cThe rate of performing emergency overnight procedures (at the provider level) was used as an instrument. View Large Likewise, there was no association of undergoing treatment by surgeons performing overnight emergency procedures with discharge to a facility (adjusted difference, –0.1%; 95% CI, –1.2% to 1.2%) after elective cerebral aneurysm treatment (Table 2). Lastly, there was no association of undergoing treatment by surgeons performing overnight emergency procedures with longer LOS (adjusted difference, –0.58; 95% CI, –1.66 to 0.50) during hospitalization after elective cerebral aneurysm treatment (Table 2). DISCUSSION Using a comprehensive all-payer cohort of patients in New York State, there was no association of undergoing treatment by a surgeon operating on emergencies the night before with mortality, discharge to a facility, or LOS for elective cerebral aneurysm treatment. Contrary to our findings, it is generally accepted by the public that sleep deprivation and fatigue adversely affect physician performance.4,13 This is contributing to an ongoing debate about whether attending surgeon work hours need to be regulated more tightly, similar to what applies for trainees.4,13 Prior observational studies attempting to answer this question have shown mixed results. Rothschild et al14 in a single-center study demonstrated that among patients undergoing procedures in a tertiary care center, those operated by obstetricians or general surgeons working the night before did not experience inferior outcomes. However, less than 6 h of sleep was associated with a higher rate of complications. The generalizability of these findings is limited given the single-center nature of the study and its focus on obstetric procedures. In addition, this investigation was limited to a teaching facility, where procedures are performed by larger teams that can compensate for individual fatigue, a setting not replicated in the community. Studies demonstrating no difference in surgical outcomes for sleep-deprived cardiac surgeons suffer from similar biases.21,22 In 2 multicenter studies in Ontario, Canada, general surgical procedures performed by surgeons operating the night before were not associated with inferior outcomes.15,16 However, the realities of their practice might have been different from the United States. In addition, the authors did not adjust for unmeasured confounders in either study. Surgeons with higher burden of emergency cases are not identical to those having mainly an elective practice. They can differ in skill, motivation, patient selection and comorbidity profile, as well as case complexity. Other potential confounders are the duration and complexity of the overnight case(s), call schedules that do not overlap with elective case work, possibility of nonoperative overnight work, the identity of the physician actually, performing the case, and potential alterations in next day case schedule. Not accounting for this questions the reliability of their findings. Our study, purposefully addresses this potential bias, by utilizing an instrumental variable analysis, to account for unmeasured confounding. Lastly, prior studies were based on fairly common operations, which likely do not require significant concentration by experienced surgeons. On the contrary, we utilized cerebral aneurysm treatment as a paradigm, one of the most technically and mentally challenging procedures in neurosurgery. The implementation of policies limiting attending surgeon work hours is controversial. The 2008 Institute of Medicine report on resident duty hours did not address limitations for attending physician work hours.11 Some argue that attending physicians have more experience, which may compensate for worsening performance. Critics of tighter regulation suggest it could result in fragmented error-prone care.23 Preventing surgeons from operating during the day after they operated the night before would require significant restructuring of healthcare delivery with major cost, staffing, and resource implications, without tangible benefit as demonstrated in this study.24 Given the shortage of neurosurgeons nationwide, restricting their work hours in some rural underserved areas would create significant barriers to care. The results of the present study do not support the argument for further regulation. Limitations Our study has several limitations. Residual confounding could bias our results. This is minimized by using a good instrument, as suggested by our F statistic. In addition, coding inaccuracies can impact our estimates. However, multiple studies have shown that cerebrovascular disease coding has excellent correlation with medical records.25,26 SPARCS includes all New York hospitals, but these results are not necessarily generalizable to the entire United States. SPARCS does not include any information on the size, location, or structure of the aneurysms. However, we are attempting to control for this, using an instrumental variable analysis. Additionally, we did not have postdischarge and long-term results on our patients. Quality measures are also not available through this source, and therefore we cannot assess these outcomes. This will be feasible through national registries such as the first cerebrovascular module created by the NeuroPoint Alliance.27 Although disposition does not always represent the patient's functional outcome, discharge location has been shown28 to correlate with modified Rankin Scale score. Finally, we cannot definitely establish causality, despite the use of advanced techniques. CONCLUSION The association between long work hours and outcomes among attending surgeons remains an issue of debate. We investigated whether operating emergently the night before an elective case was associated with inferior outcomes among attending neurosurgeons. Using a comprehensive all-payer cohort of patients in New York State, we did not identify an association of undergoing treatment by a surgeon operating on emergencies the night before with mortality, discharge to a facility, or LOS for elective cerebral aneurysm treatment. Our study had 80% power to detect differences in mortality (our primary outcome), as small as 4.1%. The results of the present study do not support the argument for regulation of attending work hours. Disclosures This work was supported by grants from the National Center for Advancing Translational Sciences (NCATS) of the NIH (Dartmouth Clinical and Translational Science Institute-UL1TR001086). The funders had no role in the design or execution of the study. The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article. REFERENCES 1. Nurok M, Czeisler CA, Lehmann LS. Sleep deprivation, elective surgical procedures, and informed consent. N Engl J Med . 2010; 363( 27): 2577- 2579. Google Scholar CrossRef Search ADS PubMed  2. Landrigan CP, Rothschild JM, Cronin JW et al.   Effect of reducing interns' work hours on serious medical errors in intensive care units. N Engl J Med . 2004; 351( 18): 1838- 1848. Google Scholar CrossRef Search ADS PubMed  3. Fletcher KE, Davis SQ, Underwood W, Mangrulkar RS, McMahon LFJ, Saint S. Systematic review: effects of resident work hours on patient safety. Ann Intern Med . 2004; 141( 11): 851- 857. Google Scholar CrossRef Search ADS PubMed  4. Hyman NH. Attending work hour restrictions: is it time? Arch Surg . 2009; 144( 1): 7- 8. Google Scholar CrossRef Search ADS PubMed  5. Gaba DM, Howard SK. Patient safety: fatigue among clinicians and the safety of patients. N Engl J Med . 2002; 347( 16): 1249- 1255. Google Scholar CrossRef Search ADS PubMed  6. Pellegrini CA, Britt LD, Hoyt DB. Sleep deprivation and elective surgery. N Engl J Med . 2010; 363( 27): 2672- 2673. Google Scholar CrossRef Search ADS PubMed  7. Ayas NT, Barger LK, Cade BE et al.   Extended work duration and the risk of self-reported percutaneous injuries in interns. JAMA . 2006; 296( 9): 1055- 1062. Google Scholar CrossRef Search ADS PubMed  8. Barger LK, Cade BE, Ayas NT et al.   Extended work shifts and the risk of motor vehicle crashes among interns. N Engl J Med . 2005; 352( 2): 125- 134. Google Scholar CrossRef Search ADS PubMed  9. Hoh BL, Neal DW, Kleinhenz DT, Hoh DJ, Mocco J, Barker FGn. Higher complications and no improvement in mortality in the ACGME resident duty-hour restriction era: an analysis of more than 107,000 neurosurgical trauma patients in the Nationwide Inpatient Sample database. Neurosurgery . 2012; 70( 6): 1369- 1381. Google Scholar CrossRef Search ADS PubMed  10. Babu R, Thomas S, Hazzard MA et al.   Morbidity, mortality, and health care costs for patients undergoing spine surgery following the ACGME resident duty-hour reform: Clinical article. J Neurosurg Spine . 2014; 21( 4): 502- 515. Google Scholar CrossRef Search ADS PubMed  11. National Research Council. Institute of Medicine; Committee on Optimizing Graduate Medical Trainee (Resident) Hours and Work Schedules to Improve Patient Safety . Washington, DC 2008. 12. Bilimoria KY, Chung JW, Hedges LV et al.   National cluster-randomized trial of duty-hour flexibility in surgical training. N Engl J Med . 2016; 374( 8): 713- 727. Google Scholar CrossRef Search ADS PubMed  13. MacDonald NE, Hébert PC, Flegel K, Stanbrook MB. Working while sleep-deprived: not just a problem for residents. CMAJ . 2011; 183( 15): 1689. Google Scholar CrossRef Search ADS PubMed  14. Rothschild JM, Keohane CA, Rogers S et al.   Risks of complications by attending physicians after performing nighttime procedures. JAMA . 2009; 302( 14): 1565- 1572. Google Scholar CrossRef Search ADS PubMed  15. Govindarajan A, Urbach DR, Kumar M et al.   Outcomes of daytime procedures performed by attending surgeons after night work. N Engl J Med . 2015; 373( 9): 845- 853. Google Scholar CrossRef Search ADS PubMed  16. Vinden C, Nash DM, Rangrej J et al.   Complications of daytime elective laparoscopic cholecystectomies performed by surgeons who operated the night before. JAMA . 2013; 310( 17): 1837- 1841. Google Scholar CrossRef Search ADS PubMed  17. Health NYSDo. Statewide Planning and Research Cooperative System (SPARCS) ; 2015. Available at: https://www.health.ny.gov/statistics/sparcs/. Accessed February 13, 2015. 18. Staiger D, Stock JH. Instrumental variables regression with weak instruments. Econometrica . 1997; 65( 3): 557- 586. Google Scholar CrossRef Search ADS   19. Garabedian LF, Chu P, Toh S, Zaslavsky AM, Soumerai SB. Potential bias of instrumental variable analyses for observational comparative effectiveness research. Ann Intern Med . 2014; 161( 2): 131- 138. Google Scholar CrossRef Search ADS PubMed  20. Foster EM. Instrumental variables for logistic regression: an illustration. Soc Sci Res.  1997; 26: 287- 504. Google Scholar CrossRef Search ADS   21. Chu MW, Stitt LW, Fox SA et al.   Prospective evaluation of consultant surgeon sleep deprivation and outcomes in more than 4000 consecutive cardiac surgical procedures. Arch Surg . 2011; 146( 9): 1080- 1085. Google Scholar CrossRef Search ADS PubMed  22. Ellman PI, Law MG, Tache-Leon C et al.   Sleep deprivation does not affect operative results in cardiac surgery. Ann Thorac Surg . 2004; 78( 3): 906- 911. Google Scholar CrossRef Search ADS PubMed  23. Stavert RR, Lott JP. The bystander effect in medical care. N Engl J Med . 2013; 368( 1): 8- 9. Google Scholar CrossRef Search ADS PubMed  24. Payette M, Chatterjee A, Weeks WB. Cost and workforce implications of subjecting all physicians to aviation industry work-hour restrictions. Am J Surg . 2009; 197( 6): 820- 825. Google Scholar CrossRef Search ADS PubMed  25. Kokotailo RA, Hill MD. Coding of stroke and stroke risk factors using international classification of diseases, revisions 9 and 10. Stroke . 2005; 36( 8): 1776- 17781. Google Scholar CrossRef Search ADS PubMed  26. Tirschwell DL, Longstreth WTJ. Validating administrative data in stroke research. Stroke . 2002; 33( 10): 2465- 2470. Google Scholar CrossRef Search ADS PubMed  27. NeuroPoint Alliance. The National Neurosurgery Quality and Outcomes Database (N2QOD) ; 2015. Available at: http://www.neuropoint.org/NPAN2QOD.html. Accessed January 15, 2015. 28. Bonita R, Beaglehole R. Recovery of motor function after stroke. Stroke . 1988; 19( 12): 1497- 1500. Google Scholar CrossRef Search ADS PubMed  Supplemental digital content is available for this article at www.neurosurgery-online.com. COMMENTS The authors are to be congratulated for adding an important study to the literature on a very important topic. As they note, the effect of physician fatigue on patient safety has been hotly debated. It is because of the concerns about fatigued residents being a threat to patients that resident duty hour restrictions were instituted in the United States. However, after 13 years of duty hour restrictions we still have no good evidence that such restrictions improve patient care.1-4 Studies and surveys since 2003 have documented adverse effects from duty hour restrictions. Residents miss vital educational opportunities by reducing their hours of surgical experience, using midlevel practitioners for educationally valuable activities, reducing time in elective operations where surgical techniques are refined and reducing research and conference time. In addition, these rules compromise the continuity of care of neurologically unstable patients. Perhaps most important, duty hour rules foster a shift-work mentality with its attendant loss of the individual's commitment to the patient. Duty hour restrictions force our residents to choose between adherence to regulations requiring them to end their shift or their commitment to patients who could still benefit from their care. If they choose the latter, they must lie or put their program at risk. We should not make our residents feel they must lie about doing the right thing. An ever-increasing volume of data, including data from prospective randomized trials,1 demonstrates that the premise that we accepted - that restricting duty hours would improve patient safety - is false. We have not improved the safety of today's patients, but we are sacrificing the safety of future patients by diminishing resident training and making shift workers of those who should be learning to be consummate professionals. Despite this, there are those who call for more stringent duty hour restrictions for both residents and staff physicians. Imposing work hour restrictions on staff neurosurgeons would have profound negative consequences for our patients and our specialty. Neurosurgery is a demanding technical specialty, but we do much more than perform procedures. We care for our patients in the clinic, the emergency room, the operating room, the recovery room, the intensive care unit, and on the hospital wards. We are specialists in the care of patients with neurological disease, not technicians who have mastered a motor skill. We have always taken care of our patients whenever they need us, for as long as they need us. This is a founding principle of our specialty that we must not abandon.5 Work hour restrictions for residents or staff physicians is not the right answer to insure patient safety. Fatigue is a fact of life in a surgical career. It cannot be eliminated but it can be recognized and managed6 in a way that best serves our patients and our specialty. Papers like this one are the data we need to make our point. Robert E. Harbaugh Hershey, Pennsylvania 1. Bilimoria KY, Chung JW, Hedges LV, Dahlke AR, Love R, Cohen ME et al.   National Cluster-Randomized Trial of Duty-Hour Flexibility in Surgical Training. N Engl J Med . 2016; 374: 713- 727. Google Scholar CrossRef Search ADS PubMed  2. Dacey RG Jr.: Editorial: Resident duty hour regulations: time for reassessment and revision. J Neurosurg . 2015; 1- 2. 3. Ganju A, Kahol K, Lee P, Simonian N, Quinn SJ, Ferrara JJ et al.   The effect of call on neurosurgery residents' skills: implications for policy regarding resident call periods. J Neurosurg . 2012; 116: 478- 482. Google Scholar CrossRef Search ADS PubMed  4. Hoh BL, Neal DW, Kleinhenz DT, Hoh DJ, Mocco J, Barker FGn. Higher complications and no improvement in mortality in the ACGME resident duty hour restriction era: an analysis of more than 107,000 neurosurgical trauma patients in the Nationwide Inpatient Sample database. Neurosurgery . 2012; 70: 1369- 1381. Google Scholar CrossRef Search ADS PubMed  5. Harbaugh RE. Neurosurgery's founding principles. J Neurosurg . 2015; 123: 1351- 1357. Google Scholar CrossRef Search ADS PubMed  6. Govindarajan A, Urbach DR, Kumar M et al.   Outcomes of Daytime Procedures Performed by Attending Surgeons after Night Work. N Engl J Med . 2015; 373: 845- 853. Google Scholar CrossRef Search ADS PubMed  The authors have analyzed data from the SPARCS (NY) database from 2009–2013. In this, they reviewed the effect of operating the night prior to treating an elective aneurysm. They have used an instrumental variable analysis, which attempted to eliminate the uncontrolled confounding present in many medical treatment studies. The primary outcome of inpatient mortality was not different between the groups, along with the secondary outcomes of length of stay and discharge to a facility other than the patient's home. The authors confirm what many neurosurgeons have suggested for years. Namely our ability to monitor our personal functional status for safe and effective care of patients is intact. While there remain numerous variables to this question (surgeon's experience, technical difficulty of the coiling/clipping, time and length of the previous night's surgery, resident assistance at one or the other procedures, etc), in general, one can conclude that neurosurgeons can effectively police their capabilities. The manuscript does not analyze quality outcomes but hopefully these conclusions will be addressed with ongoing collaborative databases. Nathan E. Simmons Lebanon, New Hampshire Brutal on-call schedules represent an archaic tradition among medical trainees, one that has been rectified in recent years. In the case of surgical residents, some would say too much so. Although studies linking residents' work hours to performance have been inconclusive, there are calls to limit attending physician on-call hours. The authors address this issue by looking at indices of success in surgery for unruptured cerebral aneurysms, comparing a cohort of neurosurgeons who performed emergency surgery the night before the procedure with those who did not. They demonstrated no differences between the 2 groups with regards to perioperative mortality, length of hospital stay, or discharge to a facility, thus refuting the argument to regulate attending work hours. The erudite statistical analysis is adjusted for a number of covariate risks and for a number of possible confounders. Instrumental variable analysis was used to adjust for the fact that membership in each cohort is not random. This choice seems eminently reasonable, although I lack the statistical sophistication to critique its application. The same might be said for the sensitivity analyses and other complex regression techniques. Sherman C. Stein Philadelphia, Pennsylvania Copyright © 2017 by the Congress of Neurological Surgeons http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Neurosurgery Oxford University Press

Outcomes of Elective Cerebral Aneurysm Treatment Performed by Attending Neurosurgeons after Night Work

Loading next page...
 
/lp/ou_press/outcomes-of-elective-cerebral-aneurysm-treatment-performed-by-00Qf9xAmaD
Publisher
Oxford University Press
Copyright
Copyright © 2017 by the Congress of Neurological Surgeons
ISSN
0148-396X
eISSN
1524-4040
D.O.I.
10.1093/neuros/nyx174
Publisher site
See Article on Publisher Site

Abstract

Abstract BACKGROUND The association between long work hours and outcomes among attending surgeons remains an issue of debate. OBJECTIVE To investigate whether operating emergently the night before an elective case was associated with inferior outcomes among attending neurosurgeons. METHODS We executed a cohort study with unruptured cerebral aneurysm patients, who underwent endovascular coiling or surgical clipping from 2009 to 2013 and were registered in the Statewide Planning and Research Cooperative System database. We investigated the association of treatment by surgeons performing emergency procedures the night before with outcomes of elective cerebral aneurysm treatment using an instrumental variable analysis. RESULTS Overall, 4700 patients underwent treatment for unruptured cerebral aneurysms. There was no difference in inpatient mortality (adjusted difference, –0.7%; 95% confidence interval [CI], –1.4% to 0.02%), discharge to a facility (adjusted difference, –0.1%; 95% CI, –1.2% to 1.2%), or length of stay (adjusted difference, –0.58; 95% CI, –1.66 to 0.50) between patients undergoing elective cerebral aneurysm treatment by surgeons who performed emergency procedures the night before, and those who did not. CONCLUSION Using a comprehensive patient cohort in New York State for elective treatment of unruptured cerebral aneurysms, we did not identify an association of treatment by surgeons performing emergency procedures the night before, with mortality, discharge to a facility, or length of stay. Our study had 80% power to detect differences in mortality (our primary outcome), as small as 4.1%. The results of the present study do not support the argument for regulation of attending work hours. Unruptured cerebral aneurysm, Fatigue, Work hours, Instrumental variable, SPARCS ABBREVIATIONS ABBREVIATIONS 2SLS 2-stage least squares CI confidence interval ICD-9-CM International Classification of Disease-9-Clinical Modification LOS length of stay SD standard deviation SPARCS Statewide Planning and Research Cooperative System There is heightened public awareness and ongoing debate about the effects of sleep deprivation and fatigue on physician performance.1-6 Some studies have demonstrated that medical resident fatigue is associated with a higher rate of errors including percutaneous needle sticks and laceration injuries, as well as postcall motor vehicle crashes.2,7,8 Others have not replicated these findings among surgical residents.9,10 However, legislative reform and comprehensive regulatory efforts have transformed graduate medical education and have imposed strict work hour restrictions on trainees.11 Recent randomized trials demonstrated that loose adherence to these restrictions was not associated with increased complications, fueling further debate on this issue.12 Unlike residents, there is no restriction on the work hours of attending physicians in the United States.4,13 Prior studies on the effect of sleep deprivation of attending surgeons on outcomes have had conflicting results. Single-center studies14 have demonstrated an association with inferior outcomes, whereas multicenter Canadian investigations15,16 have identified no difference. The generalization of the latter studies in the United States is limited given the differences in the practice of medicine between the 2 countries. Additionally, these prior investigations have focused on general surgeons, who frequently subspecialize to only provide emergency care. However, neurosurgeons are very commonly called upon to treat trauma or other neurological pathology emergently, while maintaining a busy elective surgical schedule. No prior study has attempted to answer this question in a comprehensive cohort of neurosurgeons. We used the New York Statewide Planning and Research Cooperative System (SPARCS)17 to study the association of performing emergency procedures the night before an elective case with mortality, discharge to a facility, and length of stay for patients being treated for unruptured cerebral aneurysms. We used an instrumental variable analysis to simulate the effect of randomization. METHODS New York SPARCS The cohort included all patients with unruptured cerebral aneurysms who were registered in the SPARCS (New York State Department of Health, Albany, New York)17 database between 2009 and 2013. More information on SPARCS can be found at https://www.health.ny.gov/statistics/sparcs/. Cohort Definition This study was approved by the institutional review board and consent was waived due to deidentified data. In order to establish the cohort of patients, we used International Classification of Disease-9-Clinical Modification (ICD-9-CM) codes to select patients in the database who underwent surgical clipping (ICD-9-CM code 39.51) or endovascular coiling (ICD-9-CM code 39.52 [should also have a code 88.41 and no 39.51 during the same hospitalization], 39.72, 39.75, 39.76) for unruptured (ICD-9-CM code 437.3) cerebral aneurysms between 2009 and 2013. Outcome Variables Mortality during the hospitalization for unruptured cerebral aneurysm clipping or endovascular treatment was the primary outcome variable. Length of stay (LOS) during the initial hospitalization, and the rate of discharge to a facility (any facility other than the patient's home) were secondary outcomes. Exposure Variables Performing emergency procedures the night before (12 am to 6 am) elective cerebral aneurysm treatment was the primary exposure variable. Covariates (Table, Supplemental Digital Content 1) used in the regressions were age, race (African–American, Hispanic, Asian, Caucasian, other), gender, insurance (private, Medicare, Medicaid, uninsured, other), and procedure performed (clipping, endovascular intervention). Additional comorbidities used in the regressions were smoking, diabetes mellitus, chronic lung disease, hypercholesterolemia, hypertension, peripheral vascular disease, coronary artery disease, congestive heart failure, history of stroke, obesity, transient ischemic attack, chronic renal failure, alcohol abuse, and coagulopathy. Only variables identified by the database as “present on admission” were included in the patient's comorbidity profile. Statistical Analysis The association of the exposure variable of interest with our outcomes was examined in multivariable analyses. Physicians performing elective procedures after night work are nonrandomly selected. They are expected to be younger, since they are still actively involved in call, and less subspecialized (although most neurosurgeons take call regardless of subspecialty focus), which can be associated with inferior outcomes. In order to account for this unmeasured confounding, we used an instrumental variable analysis.18 The rate of performing emergency overnight procedures (at the provider level) was used as an instrument. Similar rates have been used before to create pseudorandomization on the exposure variable.19 This approach allows us to compare the outcomes of elective cases performed by surgeons operating the night before vs those performed by surgeons not operating the prior night, while controlling for unmeasured confounding. The methodology of instrumental variables has become increasingly popular over the past decade. This method provides a way to estimate a treatment effect without bias despite the presence of confounding (ie, unknown baseline functional clinical status of patients). Instrumental variables do indeed exist. We make them all the time when we run a randomized trial. Randomization is an example of an instrumental variable, because the flip of a coin or a computer-generated random number determines the treatment a subject is assigned and therefore the treatment a subject receives. However, when health services or clinical researchers think of instrumental variables, they are usually conceiving of some natural experiment. Favored choices for instrumental variables in the medical literature so far have been geographic distances, or geographic rates. We consider a good instrument a variable that is only associated with the outcome only through the exposure of interest.19 We used a 2-stage least squares (2SLS) approach. In the first stage of the 2SLS analysis, the F statistic was 30. This is compatible with a strong instrument (F statistic >10).18 In sensitivity analysis, the main regressions were repeated separately in subgroups of patients undergoing only endovascular coiling, or only surgical clipping. There was no change in the direction of the observed associations, and therefore we do not report these results further. A probit regression was used for the categorical outcomes (mortality, discharge to a facility),20 and a linear regression for the linear outcomes (LOS). In additional to sensitivity analysis, we repeated all analyses after logarithmic transformation of LOS to account for potentially positively skewed distribution of this variable. There was no change in the direction of the observed associations, and therefore we do not report these results further. The covariates used for risk adjustment in these models were age, race, gender, hospital ID, insurance, and all the previous comorbidities. We used the marginal effects of our independent variables to obtain interpretable coefficients. Regression diagnostics were performed on all models. All results are based on 2-sided tests, and the level of statistical significance was set at 0.05. This study, based on 4700 patients, has sufficient power (80%) at a 5% type I error rate to detect differences in mortality, as small as 4.1%. We used Stata version 13 (StataCorp, College Station, Texas) to perform the analyses. RESULTS Patient Characteristics Overall, there were 4700 patients being treated for unruptured cerebral aneurysms (mean age was 55.0 years, with 76.7% females) who were registered in SPARCS. A total of 2541 (54.1%) underwent elective aneurysm treatment by neurosurgeons who operated the night before; whereas 2159 (45.9%) underwent treatment by neurosurgeons who did not operate the night before. The baseline characteristic of the 2 cohorts can be seen in Table 1. TABLE 1. Patient Characteristics   Total n = 4700  Neurosurgeons operating after night work n = 2159  Neurosurgeons operating without prior night work n = 2541    Mean  SD  Mean  SD  Mean  SD  Age  55.00  14.12  54.09  14.82  55.82  13.56    n  %  n  %  n  %  Female gender  3605  76.7%  1662  77.0%  1906  75.0%  Comorbidities              Diabetes  484  10.3%  225  10.4%  267  10.5%  Smoking  912  19.4%  438  20.3%  458  18.7%  Obesity  291  6.2%  123  5.7%  162  6.6%  Ischemic stroke  24  0.5%  9  0.4%  18  0.7%  Coronary artery disease  348  7.4%  168  7.8%  185  7.3%  Chronic lung disease  804  17.1%  358  16.6%  462  18.2%  Congestive heart failure  94  2.0%  43  2.0%  51  2.0%  Coagulopathy  38  0.8%  15  0.7%  28  1.1%  Chronic renal failure  19  0.4%  6  0.3%  15  0.6%  Hypertension  2477  52.7%  1080  50.0%  1319  53.8%  Hypercholesterolemia  1344  28.6%  600  27.8%  701  28.6%  Alcohol  89  1.9%  37  1.7%  51  2.1%  Peripheral vascular disease  132  2.8%  54  2.5%  74  3.0%    Total n = 4700  Neurosurgeons operating after night work n = 2159  Neurosurgeons operating without prior night work n = 2541    Mean  SD  Mean  SD  Mean  SD  Age  55.00  14.12  54.09  14.82  55.82  13.56    n  %  n  %  n  %  Female gender  3605  76.7%  1662  77.0%  1906  75.0%  Comorbidities              Diabetes  484  10.3%  225  10.4%  267  10.5%  Smoking  912  19.4%  438  20.3%  458  18.7%  Obesity  291  6.2%  123  5.7%  162  6.6%  Ischemic stroke  24  0.5%  9  0.4%  18  0.7%  Coronary artery disease  348  7.4%  168  7.8%  185  7.3%  Chronic lung disease  804  17.1%  358  16.6%  462  18.2%  Congestive heart failure  94  2.0%  43  2.0%  51  2.0%  Coagulopathy  38  0.8%  15  0.7%  28  1.1%  Chronic renal failure  19  0.4%  6  0.3%  15  0.6%  Hypertension  2477  52.7%  1080  50.0%  1319  53.8%  Hypercholesterolemia  1344  28.6%  600  27.8%  701  28.6%  Alcohol  89  1.9%  37  1.7%  51  2.1%  Peripheral vascular disease  132  2.8%  54  2.5%  74  3.0%  View Large Outcomes Patients operated on by surgeons after night work demonstrated 1.3% mortality (33 patients), 17.8% discharge to rehabilitation (436 patients), and on average a 5.5-day LOS (standard deviation [SD] 15.0). Patients operated by surgeons who did not operate the night before demonstrated 1.5% mortality (32), 18.1% discharge to rehabilitation (391), and on average a 3.9-day LOS (SD 7.0). Instrumental Variable Analysis Using a probit regression with instrumental variable analysis, we did not identify an association of undergoing treatment by surgeons performing emergency procedures the night before with mortality (adjusted difference, –0.7%; 95% confidence interval [CI], –1.4% to 0.02%) after elective cerebral aneurysm treatment (Table 2). TABLE 2. Multivariable Models Examining the Association of Performing Emergency Procedures the Night Before an Elective Case with Outcomes of Unruptured Cerebral Aneurysm Treatment   Inpatient Mortalitya    Discharge to rehabilitationa    Length of stayb      Adjusted difference (95% CI)  P-value  Adjusted difference (95% CI)  P-value  Adjusted difference (95% CI)  P-value  Instrumental variable analysisc  –0.7% (–1.4% to 0.02%)  .684  –0.1% (–1.2% to 1.2%)  .970  –0.58 (–1.66 to 0.50)  .994    Inpatient Mortalitya    Discharge to rehabilitationa    Length of stayb      Adjusted difference (95% CI)  P-value  Adjusted difference (95% CI)  P-value  Adjusted difference (95% CI)  P-value  Instrumental variable analysisc  –0.7% (–1.4% to 0.02%)  .684  –0.1% (–1.2% to 1.2%)  .970  –0.58 (–1.66 to 0.50)  .994  CI: confidence intervals; OR: odds ratio. aBased on a probit model. bAll regressions were based on linear models. cThe rate of performing emergency overnight procedures (at the provider level) was used as an instrument. View Large Likewise, there was no association of undergoing treatment by surgeons performing overnight emergency procedures with discharge to a facility (adjusted difference, –0.1%; 95% CI, –1.2% to 1.2%) after elective cerebral aneurysm treatment (Table 2). Lastly, there was no association of undergoing treatment by surgeons performing overnight emergency procedures with longer LOS (adjusted difference, –0.58; 95% CI, –1.66 to 0.50) during hospitalization after elective cerebral aneurysm treatment (Table 2). DISCUSSION Using a comprehensive all-payer cohort of patients in New York State, there was no association of undergoing treatment by a surgeon operating on emergencies the night before with mortality, discharge to a facility, or LOS for elective cerebral aneurysm treatment. Contrary to our findings, it is generally accepted by the public that sleep deprivation and fatigue adversely affect physician performance.4,13 This is contributing to an ongoing debate about whether attending surgeon work hours need to be regulated more tightly, similar to what applies for trainees.4,13 Prior observational studies attempting to answer this question have shown mixed results. Rothschild et al14 in a single-center study demonstrated that among patients undergoing procedures in a tertiary care center, those operated by obstetricians or general surgeons working the night before did not experience inferior outcomes. However, less than 6 h of sleep was associated with a higher rate of complications. The generalizability of these findings is limited given the single-center nature of the study and its focus on obstetric procedures. In addition, this investigation was limited to a teaching facility, where procedures are performed by larger teams that can compensate for individual fatigue, a setting not replicated in the community. Studies demonstrating no difference in surgical outcomes for sleep-deprived cardiac surgeons suffer from similar biases.21,22 In 2 multicenter studies in Ontario, Canada, general surgical procedures performed by surgeons operating the night before were not associated with inferior outcomes.15,16 However, the realities of their practice might have been different from the United States. In addition, the authors did not adjust for unmeasured confounders in either study. Surgeons with higher burden of emergency cases are not identical to those having mainly an elective practice. They can differ in skill, motivation, patient selection and comorbidity profile, as well as case complexity. Other potential confounders are the duration and complexity of the overnight case(s), call schedules that do not overlap with elective case work, possibility of nonoperative overnight work, the identity of the physician actually, performing the case, and potential alterations in next day case schedule. Not accounting for this questions the reliability of their findings. Our study, purposefully addresses this potential bias, by utilizing an instrumental variable analysis, to account for unmeasured confounding. Lastly, prior studies were based on fairly common operations, which likely do not require significant concentration by experienced surgeons. On the contrary, we utilized cerebral aneurysm treatment as a paradigm, one of the most technically and mentally challenging procedures in neurosurgery. The implementation of policies limiting attending surgeon work hours is controversial. The 2008 Institute of Medicine report on resident duty hours did not address limitations for attending physician work hours.11 Some argue that attending physicians have more experience, which may compensate for worsening performance. Critics of tighter regulation suggest it could result in fragmented error-prone care.23 Preventing surgeons from operating during the day after they operated the night before would require significant restructuring of healthcare delivery with major cost, staffing, and resource implications, without tangible benefit as demonstrated in this study.24 Given the shortage of neurosurgeons nationwide, restricting their work hours in some rural underserved areas would create significant barriers to care. The results of the present study do not support the argument for further regulation. Limitations Our study has several limitations. Residual confounding could bias our results. This is minimized by using a good instrument, as suggested by our F statistic. In addition, coding inaccuracies can impact our estimates. However, multiple studies have shown that cerebrovascular disease coding has excellent correlation with medical records.25,26 SPARCS includes all New York hospitals, but these results are not necessarily generalizable to the entire United States. SPARCS does not include any information on the size, location, or structure of the aneurysms. However, we are attempting to control for this, using an instrumental variable analysis. Additionally, we did not have postdischarge and long-term results on our patients. Quality measures are also not available through this source, and therefore we cannot assess these outcomes. This will be feasible through national registries such as the first cerebrovascular module created by the NeuroPoint Alliance.27 Although disposition does not always represent the patient's functional outcome, discharge location has been shown28 to correlate with modified Rankin Scale score. Finally, we cannot definitely establish causality, despite the use of advanced techniques. CONCLUSION The association between long work hours and outcomes among attending surgeons remains an issue of debate. We investigated whether operating emergently the night before an elective case was associated with inferior outcomes among attending neurosurgeons. Using a comprehensive all-payer cohort of patients in New York State, we did not identify an association of undergoing treatment by a surgeon operating on emergencies the night before with mortality, discharge to a facility, or LOS for elective cerebral aneurysm treatment. Our study had 80% power to detect differences in mortality (our primary outcome), as small as 4.1%. The results of the present study do not support the argument for regulation of attending work hours. Disclosures This work was supported by grants from the National Center for Advancing Translational Sciences (NCATS) of the NIH (Dartmouth Clinical and Translational Science Institute-UL1TR001086). The funders had no role in the design or execution of the study. The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article. REFERENCES 1. Nurok M, Czeisler CA, Lehmann LS. Sleep deprivation, elective surgical procedures, and informed consent. N Engl J Med . 2010; 363( 27): 2577- 2579. Google Scholar CrossRef Search ADS PubMed  2. Landrigan CP, Rothschild JM, Cronin JW et al.   Effect of reducing interns' work hours on serious medical errors in intensive care units. N Engl J Med . 2004; 351( 18): 1838- 1848. Google Scholar CrossRef Search ADS PubMed  3. Fletcher KE, Davis SQ, Underwood W, Mangrulkar RS, McMahon LFJ, Saint S. Systematic review: effects of resident work hours on patient safety. Ann Intern Med . 2004; 141( 11): 851- 857. Google Scholar CrossRef Search ADS PubMed  4. Hyman NH. Attending work hour restrictions: is it time? Arch Surg . 2009; 144( 1): 7- 8. Google Scholar CrossRef Search ADS PubMed  5. Gaba DM, Howard SK. Patient safety: fatigue among clinicians and the safety of patients. N Engl J Med . 2002; 347( 16): 1249- 1255. Google Scholar CrossRef Search ADS PubMed  6. Pellegrini CA, Britt LD, Hoyt DB. Sleep deprivation and elective surgery. N Engl J Med . 2010; 363( 27): 2672- 2673. Google Scholar CrossRef Search ADS PubMed  7. Ayas NT, Barger LK, Cade BE et al.   Extended work duration and the risk of self-reported percutaneous injuries in interns. JAMA . 2006; 296( 9): 1055- 1062. Google Scholar CrossRef Search ADS PubMed  8. Barger LK, Cade BE, Ayas NT et al.   Extended work shifts and the risk of motor vehicle crashes among interns. N Engl J Med . 2005; 352( 2): 125- 134. Google Scholar CrossRef Search ADS PubMed  9. Hoh BL, Neal DW, Kleinhenz DT, Hoh DJ, Mocco J, Barker FGn. Higher complications and no improvement in mortality in the ACGME resident duty-hour restriction era: an analysis of more than 107,000 neurosurgical trauma patients in the Nationwide Inpatient Sample database. Neurosurgery . 2012; 70( 6): 1369- 1381. Google Scholar CrossRef Search ADS PubMed  10. Babu R, Thomas S, Hazzard MA et al.   Morbidity, mortality, and health care costs for patients undergoing spine surgery following the ACGME resident duty-hour reform: Clinical article. J Neurosurg Spine . 2014; 21( 4): 502- 515. Google Scholar CrossRef Search ADS PubMed  11. National Research Council. Institute of Medicine; Committee on Optimizing Graduate Medical Trainee (Resident) Hours and Work Schedules to Improve Patient Safety . Washington, DC 2008. 12. Bilimoria KY, Chung JW, Hedges LV et al.   National cluster-randomized trial of duty-hour flexibility in surgical training. N Engl J Med . 2016; 374( 8): 713- 727. Google Scholar CrossRef Search ADS PubMed  13. MacDonald NE, Hébert PC, Flegel K, Stanbrook MB. Working while sleep-deprived: not just a problem for residents. CMAJ . 2011; 183( 15): 1689. Google Scholar CrossRef Search ADS PubMed  14. Rothschild JM, Keohane CA, Rogers S et al.   Risks of complications by attending physicians after performing nighttime procedures. JAMA . 2009; 302( 14): 1565- 1572. Google Scholar CrossRef Search ADS PubMed  15. Govindarajan A, Urbach DR, Kumar M et al.   Outcomes of daytime procedures performed by attending surgeons after night work. N Engl J Med . 2015; 373( 9): 845- 853. Google Scholar CrossRef Search ADS PubMed  16. Vinden C, Nash DM, Rangrej J et al.   Complications of daytime elective laparoscopic cholecystectomies performed by surgeons who operated the night before. JAMA . 2013; 310( 17): 1837- 1841. Google Scholar CrossRef Search ADS PubMed  17. Health NYSDo. Statewide Planning and Research Cooperative System (SPARCS) ; 2015. Available at: https://www.health.ny.gov/statistics/sparcs/. Accessed February 13, 2015. 18. Staiger D, Stock JH. Instrumental variables regression with weak instruments. Econometrica . 1997; 65( 3): 557- 586. Google Scholar CrossRef Search ADS   19. Garabedian LF, Chu P, Toh S, Zaslavsky AM, Soumerai SB. Potential bias of instrumental variable analyses for observational comparative effectiveness research. Ann Intern Med . 2014; 161( 2): 131- 138. Google Scholar CrossRef Search ADS PubMed  20. Foster EM. Instrumental variables for logistic regression: an illustration. Soc Sci Res.  1997; 26: 287- 504. Google Scholar CrossRef Search ADS   21. Chu MW, Stitt LW, Fox SA et al.   Prospective evaluation of consultant surgeon sleep deprivation and outcomes in more than 4000 consecutive cardiac surgical procedures. Arch Surg . 2011; 146( 9): 1080- 1085. Google Scholar CrossRef Search ADS PubMed  22. Ellman PI, Law MG, Tache-Leon C et al.   Sleep deprivation does not affect operative results in cardiac surgery. Ann Thorac Surg . 2004; 78( 3): 906- 911. Google Scholar CrossRef Search ADS PubMed  23. Stavert RR, Lott JP. The bystander effect in medical care. N Engl J Med . 2013; 368( 1): 8- 9. Google Scholar CrossRef Search ADS PubMed  24. Payette M, Chatterjee A, Weeks WB. Cost and workforce implications of subjecting all physicians to aviation industry work-hour restrictions. Am J Surg . 2009; 197( 6): 820- 825. Google Scholar CrossRef Search ADS PubMed  25. Kokotailo RA, Hill MD. Coding of stroke and stroke risk factors using international classification of diseases, revisions 9 and 10. Stroke . 2005; 36( 8): 1776- 17781. Google Scholar CrossRef Search ADS PubMed  26. Tirschwell DL, Longstreth WTJ. Validating administrative data in stroke research. Stroke . 2002; 33( 10): 2465- 2470. Google Scholar CrossRef Search ADS PubMed  27. NeuroPoint Alliance. The National Neurosurgery Quality and Outcomes Database (N2QOD) ; 2015. Available at: http://www.neuropoint.org/NPAN2QOD.html. Accessed January 15, 2015. 28. Bonita R, Beaglehole R. Recovery of motor function after stroke. Stroke . 1988; 19( 12): 1497- 1500. Google Scholar CrossRef Search ADS PubMed  Supplemental digital content is available for this article at www.neurosurgery-online.com. COMMENTS The authors are to be congratulated for adding an important study to the literature on a very important topic. As they note, the effect of physician fatigue on patient safety has been hotly debated. It is because of the concerns about fatigued residents being a threat to patients that resident duty hour restrictions were instituted in the United States. However, after 13 years of duty hour restrictions we still have no good evidence that such restrictions improve patient care.1-4 Studies and surveys since 2003 have documented adverse effects from duty hour restrictions. Residents miss vital educational opportunities by reducing their hours of surgical experience, using midlevel practitioners for educationally valuable activities, reducing time in elective operations where surgical techniques are refined and reducing research and conference time. In addition, these rules compromise the continuity of care of neurologically unstable patients. Perhaps most important, duty hour rules foster a shift-work mentality with its attendant loss of the individual's commitment to the patient. Duty hour restrictions force our residents to choose between adherence to regulations requiring them to end their shift or their commitment to patients who could still benefit from their care. If they choose the latter, they must lie or put their program at risk. We should not make our residents feel they must lie about doing the right thing. An ever-increasing volume of data, including data from prospective randomized trials,1 demonstrates that the premise that we accepted - that restricting duty hours would improve patient safety - is false. We have not improved the safety of today's patients, but we are sacrificing the safety of future patients by diminishing resident training and making shift workers of those who should be learning to be consummate professionals. Despite this, there are those who call for more stringent duty hour restrictions for both residents and staff physicians. Imposing work hour restrictions on staff neurosurgeons would have profound negative consequences for our patients and our specialty. Neurosurgery is a demanding technical specialty, but we do much more than perform procedures. We care for our patients in the clinic, the emergency room, the operating room, the recovery room, the intensive care unit, and on the hospital wards. We are specialists in the care of patients with neurological disease, not technicians who have mastered a motor skill. We have always taken care of our patients whenever they need us, for as long as they need us. This is a founding principle of our specialty that we must not abandon.5 Work hour restrictions for residents or staff physicians is not the right answer to insure patient safety. Fatigue is a fact of life in a surgical career. It cannot be eliminated but it can be recognized and managed6 in a way that best serves our patients and our specialty. Papers like this one are the data we need to make our point. Robert E. Harbaugh Hershey, Pennsylvania 1. Bilimoria KY, Chung JW, Hedges LV, Dahlke AR, Love R, Cohen ME et al.   National Cluster-Randomized Trial of Duty-Hour Flexibility in Surgical Training. N Engl J Med . 2016; 374: 713- 727. Google Scholar CrossRef Search ADS PubMed  2. Dacey RG Jr.: Editorial: Resident duty hour regulations: time for reassessment and revision. J Neurosurg . 2015; 1- 2. 3. Ganju A, Kahol K, Lee P, Simonian N, Quinn SJ, Ferrara JJ et al.   The effect of call on neurosurgery residents' skills: implications for policy regarding resident call periods. J Neurosurg . 2012; 116: 478- 482. Google Scholar CrossRef Search ADS PubMed  4. Hoh BL, Neal DW, Kleinhenz DT, Hoh DJ, Mocco J, Barker FGn. Higher complications and no improvement in mortality in the ACGME resident duty hour restriction era: an analysis of more than 107,000 neurosurgical trauma patients in the Nationwide Inpatient Sample database. Neurosurgery . 2012; 70: 1369- 1381. Google Scholar CrossRef Search ADS PubMed  5. Harbaugh RE. Neurosurgery's founding principles. J Neurosurg . 2015; 123: 1351- 1357. Google Scholar CrossRef Search ADS PubMed  6. Govindarajan A, Urbach DR, Kumar M et al.   Outcomes of Daytime Procedures Performed by Attending Surgeons after Night Work. N Engl J Med . 2015; 373: 845- 853. Google Scholar CrossRef Search ADS PubMed  The authors have analyzed data from the SPARCS (NY) database from 2009–2013. In this, they reviewed the effect of operating the night prior to treating an elective aneurysm. They have used an instrumental variable analysis, which attempted to eliminate the uncontrolled confounding present in many medical treatment studies. The primary outcome of inpatient mortality was not different between the groups, along with the secondary outcomes of length of stay and discharge to a facility other than the patient's home. The authors confirm what many neurosurgeons have suggested for years. Namely our ability to monitor our personal functional status for safe and effective care of patients is intact. While there remain numerous variables to this question (surgeon's experience, technical difficulty of the coiling/clipping, time and length of the previous night's surgery, resident assistance at one or the other procedures, etc), in general, one can conclude that neurosurgeons can effectively police their capabilities. The manuscript does not analyze quality outcomes but hopefully these conclusions will be addressed with ongoing collaborative databases. Nathan E. Simmons Lebanon, New Hampshire Brutal on-call schedules represent an archaic tradition among medical trainees, one that has been rectified in recent years. In the case of surgical residents, some would say too much so. Although studies linking residents' work hours to performance have been inconclusive, there are calls to limit attending physician on-call hours. The authors address this issue by looking at indices of success in surgery for unruptured cerebral aneurysms, comparing a cohort of neurosurgeons who performed emergency surgery the night before the procedure with those who did not. They demonstrated no differences between the 2 groups with regards to perioperative mortality, length of hospital stay, or discharge to a facility, thus refuting the argument to regulate attending work hours. The erudite statistical analysis is adjusted for a number of covariate risks and for a number of possible confounders. Instrumental variable analysis was used to adjust for the fact that membership in each cohort is not random. This choice seems eminently reasonable, although I lack the statistical sophistication to critique its application. The same might be said for the sensitivity analyses and other complex regression techniques. Sherman C. Stein Philadelphia, Pennsylvania Copyright © 2017 by the Congress of Neurological Surgeons

Journal

NeurosurgeryOxford University Press

Published: Mar 1, 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