Survey responses were analyzed using descriptive statistics. Differences in responses by disease category were analyzed using the student t test.

For hospital bed counts, there were five hospitals with more than 1500 beds, 12 with 1000–1499, 52 with 500–999, and eight with fewer than 500. There were 52 private university hospitals, 13 national university hospitals, nine secondary hospitals and three other hospitals, one for veterans, one military, and one a national medical center. Forty-three hospitals were tertiary level. For comparison, in the 2010 survey, a total of 50 hospitals (41 university hospitals and nine tertiary hospitals) participated (Table 3).

Nineteen hospitals (24.7%) employed a neurointensivist or faculty member; two hospitals employed two neurointensivists each, and the other 17 each had one neurointensivist on staff. All hospitals had more than 500 beds, and 16 hospitals (84.2%) were private. Among the neurointensivists, 14 had majored in neurosurgery, four in neurology, two in anesthesiology, and one in cardiothoracic surgery. Fourty seven percent (n=10) had more than 5 years of professional experience, and 47.6% (n=10) had a medical intensivist certificate from the Korean Society of Critical Care Medicine. In terms of ICU operating model, 13 hospitals were open, five were semi-closed, and one hospital were co-management. Tables 3 and 4 list the study ICUs’ operating models and the demographic data from this survey.

We compared NCU dependency and quarantine zones for infection control between the 2010 survey and our 2019 survey. In 2010, 21 hospitals (42%) operated their NCUs independently, and that increased to 35 hospitals (45.5%) in the current survey; of these, 12 hospitals did not operate NCUs independently from their neurology departments or intensive stroke units. The average number of NCU beds was 13.1 beds (range, 8–19) in this survey. Of the 35 hospitals that operated independent NCUs, 29 (82.9%) had an independent air ventilation system and 34 had a separate quarantine area for infection control; the average number of separated isolation beds was 2.35 (range, 1–6), in contrast with the average of 2.8 beds in a quarantine area in the 2010 survey. Hospitals equipped with an independent air ventilation system for infection control accounted for 70% of all respondent hospitals in the 2010 survey, and 82.9% of hospitals in that survey offered a quarantine area.

In the 2010 survey, 10 hospitals (34.5%) held neurosurgeryexclusive beds in the ICUs of other departments among 29 jointly operated hospitals. According to the findings of the 2019 survey, of the 42 hospitals that were not operating independent NCUs, 10 (23.8%) held an average of 9.4 beds in their surgical or general ICUs exclusively for their neurosurgery departments, and 32 were using nonexclusive beds in their surgical or general medical ICUs. In both surveys, hospital administrators used beds in the ICUs of other departments when a quarantine zone was not available. Table 5 gives comparative statistics for the two surveys.

This question was about the diseases that had presented for hospitalization in survey hospitals’ NCUs or NICUs over the previous year, and the disease frequencies were in order as follows : internal cerebral hemorrhage (ICH), TBI, subarachnoid hemorrhage, and others; the most common other cause of hospitalization was admission for routine post-operative care. There was no significant difference between the tertiary and non-tertiary hospital in the distribution of diseases categoriese in NCU or NICU under neurosurgery department (Table 6). However, the tertiary hospitals had high admission rates were high for postoperative care after elective surgery for neurosurgical disorders such as brain tumor and unruptured aneurysm, whereas TBI and ICH were the most frequent reasons for admission in the non-tertiary hospitals.

However, there was no statistical difference between tertiary and non-tertiary hospitals. There was also a marked difference in the distribution of disease categories between ICUs under neurosurgery departments and those under neurology. Most of the patients hospitalized in the neurologic ICUs were hospitalized with acute ischemic stroke (AIS) and other brain diseases such as nervous system infection, hypoxic encephalopathy after cardiac arrest, myasthenia gravis and demyelinating disease (Table 6).

Respondents from 37 hospitals (48.1%) reported high demand for neurointensivists, and respondents from 62 hospitals (80.5%) stated that the mandatory deployment of a neurointensivist improved the quality of patient care (Fig. 1). However, respondents from 38 hospitals (49.4%) gave low ratings for whether their hospitals had adequate human, physical, and financial resources support for installing an intensivist system, and 13 respondents (16.9%) reported difficulties related to deploying a neurointensivist (conflicts with the management team, treatment, or work).

Respondents from 44 hospitals (57.1%) believed that hiring neurointensivist would increase hospital costs, and in response to a question on potential earnings declines, 25 respondents (32.5%) anticipated conflicts regarding patients’ care between neurosurgeons and neurointensivists. Twentythree hospital respondents (29.9%) worried about loss of income to certain physician groups, but 69 hospitals of survey respondents (89.6%) observed that the most challenging aspect of recruiting neurointensivists was that they had no applicants for the position (Fig. 2).

Question 17 asked whether disclosing medical records such as hospital standardized mortality ratios and risk adjusted morality rates of ICUs was a barrier hiring neurointensivists, only 22 hospital respondents (28.6%) agreed or strongly agreed. Thirty-six respondents (46.8%) supported legislation for deployment of neurointensivists. In terms of potential solutions to these problems, 70 respondents (90.9%) maintained that additional fees were necessary for neurointensivists’ work, and 64 (83.1%) answered that direct support was needed of the personnel expenses for neurointensivists (Fig. 3).

Fifty-three hospital respondents (70.1%) agreed with the need for a new neurosurgeon-oriented NCU model; there was little negative response (strongly disagree/disagree) to this question (n=4, 5.2%) (Fig. 4).

For the survey questions related to the societies, most neurosurgeons (n=73, 94.8%) were well aware of the KNIC; 20.8% (n=16) felt that the KNIC was less competitive than the neurologist-driven KNCS.

In this survey, 65 hospital survey respondents (84.4%) believed a neurointensivist handbook was necessary. However, only 40 (61.5%) expressed willingness to participate in developing such a handbook (Fig. 5).

One survey question related to the survey respondents’ possessed NCU equipment at their hospitals, including intracranial pressure (ICP) monitors, tools for measuring cerebral blood flow velocity, transcranial Doppler, equipment for hypothermia treatment, etc., and compared their responses with data obtained from the 2010 survey. Sixty-one respondents (61/70, 87.1%) said their hospitals were equipped with ICP monitors, either Camino, Spiegelburg, or LiquidGard. Of the 62 survey respondents who answered the question, 15 (15/67, 22.4%) worked at hospitals that had devices for measuring central blood flow velocity, and of 63 respondents to a different question, 63 (63/71, 88.7%) had transcranial doppler equipment; at 34 of those hospitals, this was operated exclusively as neurologic equipment. Sixty-seven doctors responded to a question about hypothermia treatment, and 49 survey respondents (73.1%) did treat hypothermia. Fig. 6 presents these data as percentages. We found insignificant differences in the possessed equipment at survey respondents’ hospitals over the last decade.

We received responses from 77 hospitals (75 of 86 neurosurgical residency training hospitals and two neurosurgery specialist hospitals). Table 4 summarizes the demographic data from the survey. In 2010, there were 50 respondent hospitals, 41 university hospitals and nine tertiary facilities; in addition to there being more respondents to our survey, the mix of hospital types was broader. Nineteen hospitals employed a total of 21 neurointensivists (two hospitals had two each) in our survey; all of these were large hospitals with more than 500 beds, and 16 of the 19 were private. In 2010, 21 hospitals operated independent NCUs, and that increased to 35 hospitals (45.5%) in the current survey.

As shown in Table 5, there were no significant changes in independent operation of NCU and facilities such as air ventilation system for infection control, quarantined hospital beds and others over the last decade. There was no inconvenience about no facilities progression and indepent operation of NCU. Many patients with AIS and TBI admitted to stroke unit, trauma center and ICUs in emergency centers.

Although most hospital respondents to our survey reported that neurointensivists help improve the quality of patient care, only 37 reported high demand for neurointensivists at their hospitals. Our survey findings align with those from many previous studies that neurocritical care by neurointensivists contributed to quality improvement by reducing mortality and length of hospital stay [10].In some studies, a mandatory operation of a closed neurological ICU shortened the length of hospital stay and was beneficial in patient care [6,10], but few of the respondent hospitals in this survey had NCUs or NICUs staffed with neurointensivists. However, most respondents did report that they had difficulties hiring neurointensivists primarily because they had no applicants for the role.

We consider a number of reasons for the absence of applicants. First, neurointensivists perform only limited surgery and outpatient care, whereas most neurosurgeons want to use the full range of their skills; as such, many neurosurgeons do not find the neurointensivist role appealing. Second, neurointensive care is not recognized as an area of neurosurgery because critically ill patients are already treated in different neurosurgery sub-departments. Third, neurosurgeons tend to be unfamiliar with internal medicine-related knowledge and have difficulties interpreting neurophysiological tests such as electroencephalograms and electromyograms.

Markandaya et al. [2] proposed that neurologists are suitable for providing comprehensive neurological treatment as neurointensivists, but we have a different opinion. Many of the respondents to our survey thought that a new, neurosurgeryoriented model of intensive care, primarily because of the characteristics of patients hospitalized in NCUs. As shown in Table 6, the characteristics of patients hospitalized in NCUs differ entirely from the features of patients admitted to neurology ICUs [9]. Most patients will require surgery or have been admitted to an NCU after surgery; for this reason, neurointensivists in NCUs and NICUs under neurosurgery must have extensive knowledge of neurosurgical treatments, including being able to perform immediate surgical treatment if required including invasive procedures. Therefore, without education on pre- and postoperative management of neurosurgery-related treatments, neurointensivists may not be ideal for running independent NCUs.

Instead, we consider the ideal to be neurosurgery-led multidisciplinary care with teams from departments such as internal medicine, neurology, pharmacy, and rehabilitation [7,11]. We do acknowledge that neurosurgery-led multidisciplinary care may not be possible for many reasons such as economic problems and difficulties in staffing. When such a team approach is not an option for a hospital, we consider appropriate a semiclosed ICU or intensivist co-management model with neurosurgeons. In findings from a previous study, good teamwork and adequate communication between intensivists and surgeons contributed to improving the quality of perioperative care [3]. To increase support for neurosurgeons as neurointensivists, neurointensivists could be authorized to provide certain outpatient clinic and neurosurgical treatments. However, neurointensivists cannot perform all operations, and it is best if they are limited to providing surgery only in emergencies.

In terms of barriers to staffing hospital ICUs with neurointensivists, over half of respondents reported that their hospitals did not have the human, physical, and financial resource support for a neurointensivist system; a small number reported conflicts with hospital executives on the subject of neurointensivists, and more than half were concerned about the increased financial burden of hiring full-time neurointensivists (Fig. 2). The 19 hospitals in this survey that operated under a neurointensivist system were all large hospitals with more than 500 beds and most could financially support the cost of the position (Table 3). For other cases, most respondents to our survey proposed charging patients fees for critical care services by neurointensivists or providing financial support for ICU intensivist employment costs of ICU intensivists. Nearly half of respondents supported legislation on neurointensivists, but less than a third supported attracting social attention to neurology-related care by publicly reporting hospital standardized mortality ratios and risk-adjusted NCU mortality rates (Fig. 3).

Sixteen respondents felt that the KNIC was less competitive than the KNCS, which is mainly run by neurology departments, for a number of reasons such as lack of active participation by and interest from neurosurgeons, lack of good educational programs for fellowships, lack of publicity, and lack of a neurosurgery orientation. Duplicate academic interest and education also leads to a lack of professionalism. For example, basic knowledge such as ICP management considerably overlaps with clinical practice during residency and education programs of the Korean Neurotraumatology Society. To overcome this shortcoming, the KNIC needs to open up for collaborative practice by encouraging the involvement of neurologists, internal medicine specialists, and anesthesiologists.

Mutual exchange among relevant domestic and international medical societies is also crucial. Additionally, academic programs to share knowledge with other departments and learn from other specialists in treating critical neurosurgical patients are warranted, and there should be greater effort toward promoting these programs. Furthermore, the KNIC should establish training programs and qualification standards for neurointensivists because neurointensivists must have knowledge of all internal/surgical conditions and surgical treatment. For example, the Korean Society of Critical Care Medicine had the primary role in initiating subspecialities for managing critically ill patients.

We believe that an evidence-based manual or handbook related to NCUs’ clinical practice guidelines is essential for education and care quality control in institutions without sufficient certified specialists or neurosurgeons [8]. In the survey for this study, 65 hospital respondents thought such a handbook was necessary, but only 40 respondents expressed willingness to participate in producing such a handbook. The KNS and KNIC need to make efforts to update the published manual. Respondents to our survey had many opinions on making the handbook functional, and there were some opinions on developing mobile phone applications. However, there were some respondents who argued that a separate handbook is unnecessary because critical care guidelines by disease are already described in the textbooks of each specialty.

We compared four equipment retention rates with data from 10 years ago (Fig. 6) and found that despite a slight decrease in overall numbers, there were no significant changes in possessed equipment in departments of neurosurgery over the last decade. The reason may be that some hospitals were considering introducing more advanced monitoring equipment in the ICU such as near infrared spectroscopy, pupillometer, bispectral index, and minimal invasive hemodynamic monitoring system. Along with recent advances, a portable computed tomography (CT) or ultrasound scanner has been introduced because patients have difficulties moving frequently for CT or magnetic resonance imaging scans. Furthermore, monitoring equipment that can directly examine the brain is being used, and studies have verified its effectiveness [1,5].

This survey has several limitations that might affect interpretation of the data. First, the accuracy of equipment counts, facilities, and disease distributions is questionable, because some hospitals do not want to release these informations. To expand the availability of these data, we believe it is necessary for the KNS to conduct annual training hospital evaluations to obtain data. Second, selection bias may have occurred; most of the survey participants were neurosurgeons, and only two were neurologists. Thus, many responses heavily reflect the opinions of chief NCU managers, limiting the generalizability of these respondents’ opinions. Additionally, we did not solicit data for this survey from other NCU stakeholders such as hospital administrators, non-intensivist physicians, and policymakers. Third, implementation of full-time neurointensivist team systems will depend on many factors including individual NCU or NICU characteristics of each hospital. Considering strengths,weaknesses,threats and opportunities of neurosurgeons as neurointensivists (Table 7), we need to develop the appropriate education, training programs and optimum model for each hospital through further research.