Correlation of Call Burden and Sleep Deprivation with Physician Burnout, Driving Crashes, and Medical Errors among US Neurointerventionalists

DISCUSSION

NIS may be at high risk of burnout, in large part due to the high proportion of emergency procedures and the frequency of on-call coverage.^18,19 Among those who participated in our study, we found that increased years of practice and a call frequency greater than or equal to once every 3 days (increased call frequency) were independently associated with burnout. In addition, an increased call frequency was associated with an increased risk of falling asleep at the wheel, motor vehicle crashes/near-crashes, and medical errors. In addition, sleeping <4 hours, which was more commonly encountered during call days, was an independent predictor of motor vehicle crashes and medical errors. To date, there are no standards for call frequency for NIS; when asked about optimal call frequency, most responding physicians selected a maximum of once every 3 days.

In our survey, 54% of respondents reported burnout in at least 1 domain of the CBI survey, with 13.4% reporting burnout in all domains. We used the CBI survey rather than the more commonly used Maslach Burnout Inventory for several reasons. The validity of the CBI has been supported by prior studies and has been used in multiple health care studies.^17,20,21 When they are compared directly, the CBI and Maslach Burnout Inventory showed very similar results, but unlike the Maslach Burnout Inventory, the CBI can measure personal burnout as a total score with better psychomotor properties, allowing it to measure the degree of burnout change across time (rather than assuming a stable individual trait).^22,23 The CBI is also nonproprietary and open access.

In a previous study of burnout, physician well-being, and the correlation to medical errors in 6695 physicians across different medical specialties, 54.3% of physicians reported burnout; 32.8%, excessive fatigue; and 10.5%, a major medical error within the prior 3 months. The study found that burnout and fatigue were independently associated with medical errors.²⁴ Risks for burnout included long working hours, frequent overnight call duties, sleep deprivation and sleep disturbances, high work intensity, and suboptimal compensation among other factors.¹⁹ In our study, call burden (being on call once or more every 3 days) was an independent predictor of burnout. This finding aligns with prior studies showing a 3% increased risk for burnout with every extra hour of work per week and a 3%–9% increased risk of burnout for additional nights and weekend call.²⁵

As per the American Heart Association/American Stroke Association guidelines, mechanical thrombectomy is standard-of-care treatment for large-vessel-occlusion acute ischemic strokes with salvageable brain tissue up to 24 hours from symptom onset or time last known well.^26,27 Stroke call is inherently stressful due to the challenging, emergent nature of the interventions and the often high morbidity of the disease, with increased burnout also demonstrated among noninterventional stroke neurologists.²⁸ Wilson et al²⁹ demonstrated that almost 60% of stroke consults in a 24-hour shift occurred during nonworking hours. The same study showed increasing stroke thrombectomy rates from once every 5 days to once every 3 days,²⁹ which are expected to further increase with expanding indications and place NIS at an even higher risk of burnout. When asked about the optimal maximal neurointerventional call burden in our survey, approximately 80% of respondents selected a call frequency not to exceed once every 3 days, with 39% selecting once every 4 days as the ideal call frequency, yet nearly half (48%) reported a current call frequency of greater than once every 3 days.

Our finding that >10 years of neurointerventional practice experience independently predicts burnout is contrary to prior studies that showed early-to-middle career physicians work more hours and have lower career satisfaction, less work-life balance, and more burnout and are more likely to quit the practice of medicine.³⁰ The discrepancy may be related to the steep learning curve and time required for proficiency as well as the novel, challenging nature of neurointervention being strong motivators in the early years, whereas senior practitioners are liable to experience greater monotony, frustration, and less opportunity for advancement.

As a direct consequence of call burden, sleep deprivation, and burnout, it is important to consider the physical well-being of physicians and the risk of medical errors, patient harm, and litigation. Most studies find that a majority of emergent stroke thrombectomy calls occur outside standard daytime work hours. Combined with stricter guidelines for earlier access and reperfusion times, this timing means that the NIS will often transition from sleeping to driving within minutes, potentially putting them at risk for motor vehicle crashes. Although the frequency of call coverage did not directly correlate with sleeping at the wheel or motor vehicle crashes, increased call-in frequency or being called-in >3 times per week was an independent predictor of sleeping at the wheel or being involved in motor vehicle crashes/near-crashes. This finding is alarming and raises an important question of who should be responsible for ensuring good backup coverage from partners when the interventionalist has been up all night treating patients. Should there be a societal guideline by the SNIS/SVIN to restrict the number of call days per week, number of call days taken back-to-back, and the number of hours worked continuously? Does the frequency of call coverage need to be tailored to practice volumes, the number of hospitals covered on call, and the likelihood of being called-in multiple times during call periods?

The relationship between sleep and cognitive and psychomotor function has been well-established, as well as the impact on medical errors, motor vehicle crashes, and health, such as obesity and heart disease. Sleep is a very complex process and varies significantly from one night to another. Frequent call coverage is likely to cause disturbed sleep. In our survey, 65% of respondents reported <6 hours of total sleep when on call, and 43% reported <4 hours of uninterrupted sleep. While the American College of Graduate Medical Education has regulated minimum hours of protected sleep among trainees, no such rules exist for attending physicians.^31,32 The American Academy of Sleep Medicine recently published a position statement on sleep, fatigue, and burnout, estimating that maladaptive sleep techniques led to a burnout prevalence of 50%, with sleep deprivation being one of the main culprits.³² A previous study demonstrated that the risk of procedural complication rates significantly increased if physicians experienced <6 hours of sleep (3.4% versus 6.2% with an OR of 1.72).³³ In our study, having <4 hours of uninterrupted sleep while on call was an independent predictor of motor vehicle crashes/near-crashes and medical errors, whereas we found a nonsignificant trend toward increased burnout.

Again, this sounds an alarm on what needs to be done to provide call coverage for emergency stroke cases, without affecting the patients’ outcomes or putting the providers at considerable risk. Should we train more neurointerventionalists to have adequate coverage, and if so, how would that affect the volume for other neurovascular pathologies treated by those physicians and would that eventually reduce individual expertise for treating those pathologies? On the other hand, should there be an alternative training pathway in which physicians are given the opportunity to do either 1 year of training to only perform stroke thrombectomies versus 2 years of training for all neurovascular pathologies, to ensure adequate stroke coverage without affecting the expertise in treatment of other neurovascular pathologies? Should there be a minimum number of emergent large-vessel occlusions in a hospital to allow the hospital to offer stroke thrombectomy service while having enough physicians to cover/backup the call? Should there be a minimum number of neurointerventionalists/stroke interventionalists employed by the hospital to provide thrombectomy service?

With regard to compensation models that use productivity/RVU as a basis, other specialties such as Emergency Medicine have developed formulas that assign a positive multiplication factor to RVUs generated during evening/night shifts. Should this model be applied to NIS as well? Given the tremendous societal burden of stroke, should productivity models be used at all for NIS compensation? These are all important questions that need to be addressed on a societal level to ensure continued provision of thrombectomy coverage without putting physicians at significant life-threatening risk or increasing medical errors that would affect patients’ outcomes.

Our study has several limitations. First, the survey results may be distorted by recall and selection bias in that a physician experiencing burnout may be more inclined to respond and lead to an overestimation of the burnout problem. However, the rates of burnout in our survey aligned with previously published data in the neurointerventional surgery field¹⁸ and other studies analyzing burnout in radiology, neurology, and neurosurgery.¹³ Second, the small proportion of responses limited the power of our study to find statistically significant associations. Due to the relatively small number of NIS compared with other specialties, we suggest conducting a repeat survey designed by a collaboration of the major neurointerventional societies, distributed more efficiently through conferences and e-mail lists, with emphasis on specific issues identified in these previous studies. Finally, we assessed burnout, accidents, and medical errors as a function of call burden, compensation, and sleep deprivation but did not include other, more difficult-to-categorize personal factors. Controlling for this myriad of influences will always pose a challenge to surveys addressing the complex causes of burnout, fatigue, and sleep disturbances.