Re-visiting the impact of the first wave of COVID-19 on neurosurgical practice and training in a large UK neurosurgery unit: a retrospective review

Changes to facilities and resources

At the start of the pandemic, the department of neurosurgery at CUH counts 20 consultants and provides neurosurgical care for the East of England region (population: 6.235 million). Pre-COVID-19, 4 fully dedicated neurosurgery operating theatres ran each day on weekdays and an additional operating theatre on two working days a week. One of the operating theatres is a 24/7 emergency list. There is only emergency care provision on weekends. There is anaesthetic cover for three general anaesthetic neuro-endovascular lists a week.

Following the UK Government advice and the NHS England COVID-19 mitigation guidance our trust suspended all elective operating with effect from 23^rd March 2020. This meant that the neurosurgical operating was only available in one 24/7 emergency operating theatre. There was no additional anaesthetic provision for the neuro-endovascular service. Rather, emergency interventional neuroradiology was provided using the same staff for the single emergency theatre.

Significant changes were made to the neurosurgical bed pool, reducing the number of beds to allow for emergency neurosurgery care only and repurposing other neurosurgical wards as medical. Further, all wards were labelled green, amber or red wards in a hospital-wide triage system based on the likelihood of COVID-19 infection. The 16-bed female neurosurgical ward was changed into an amber COVID-19 area and the 21-bed trauma rehab unit into a red COVID-19 area.

Changes to the neurosurgical workforce included the re-deployment of 1 registrar and 1 junior clinical fellow to the COVID-19 medical workforce. Due to the high numbers of aerosol generating procedures within neurosurgery nursing staff who were deemed to be at risk (based on health risk assessments) were allocated to other areas. The trust policies for self-isolation and sickness for COVID-19 meant at times there were large numbers of nursing staff unavailable to work. A further reduction to the availability of nursing staff was caused by government guidance on pregnancy which caused several nursing staff to have to work from home.

It was rapidly recognised that this skeleton neurosurgical service was inadequate and, in the weeks and months to come, the Trust and the Department regularly revised the strategy taking into account the national guidelines, the need for (neuro)surgical services, and the COVID-19 patient load. As a result, on the 30^th March, a second neurosurgical theatre was opened for emergency and urgent cases, and on the 8^th April, the third neurosurgical operating theatre was opened for urgent cases. Additionally, the Trust started running three NHS operating lists every week in the independent sector (two at Nuffield Hospital Cambridge and one at the Spire Centre Cambridge) (for timeline, see Figure 1⁷). There was still no separate provision for emergency and urgent neuro-endovascular cases. These procedures continued to be carried out using the staff allocated to the 3 neurosurgical operating theatres, consequently interrupting the neurosurgical lists.

Figure 1. Timeline of Cambridge University Hospital (CUH) theatre changes against daily positive COVID tests in the UK.

With increased operative capacity, the neurosurgical bed and ward staff pool gradually increased over this period of time, however not yet reaching pre-COVID-19 levels.

Referrals

The number of regional referrals made in the three-month period of the lockdown did not substantially reduce when compared with the number of referrals in the same time-period from previous years (Figure 2). Comparing the referrals in 2020 with this trend, the number of referrals may have reduced during the peak of the pandemic.

Figure 2. Number of referrals per month over Feb to May, 2018-2020.

Admissions

The reduced operative capacity, bed pool and ward staffing levels as outlined above, produced a substantial reduced capacity for admissions, affecting both elective and emergency admissions (Figure 3). The number of elective and emergency admissions had begun to recover by May 2020 yet remained below pre-COVID-19 levels.

Figure 3. Number of emergency and elective admissions per month from 2016-2020.

The average age of patients admitted did not change compared to previous years (Figure 4) (Kruskal-Wallis test: emergency p=0.1531, elective p=0.5447). However, the average length of stay did significantly reduce for emergency cases but did not change for elective cases (Kruskal-Wallis test: emergency p=0.0029, elective p=0.1473).

Figure 4. Average age and length of stay for emergency and elective admissions from 2016-2020.

Theatre efficiency

The average duration of operations (time from the patient entering the operating theatre to entering recovery) and the average time of knife-to-skin (time from the patient entering the theatre to the start of operating) did not change substantially during the three-month pandemic period. However, the average downtime in which no operating was occurring increased substantially compared to previous years (Figure 5).

Figure 5. Average theatre timings in the Mar-June periods from 2016 to 2020.

The total number of theatre sessions during the March to June period (where a full day of operating is defined as 2 sessions) was lower in 2020 by 28.5% compared to the average of the four years prior (Table 1). Concomitant with the increase in theatre downtime, the rate of cases per session was reduced by 18.2%, further indicating a reduction in theatre efficiency. Using the average of cases performed and session count over the four years prior to 2020, the cases per session ratio decreased significantly in 2020 (two tailed z-statistic = 52.5, p<0.0001).

Theatre activity by pathology

Across all subspecialties, a reduction in total number of operations performed during the three-month pandemic period was seen (Figure 6). Shown in figure is the percent change for each subspecialty in total, elective, and emergency change in the three-month period compared to the average of the same time period over the four years prior. Elective work is substantially reduced while the reduction in emergency work is more modest. The emergency work for paediatrics did not reduce substantially, while for oncology the emergency work increased.

Figure 6. Operations performed in the Mar-June period from 2016 to 2020 broken down by emergency/elective and by subspecialty interest of the overseeing consultant.

Operating numbers and effect on training

Comparing the three-month period 8^th March to 7^th June in 2020 with previous years we found a significant reduction in the total number of operations performed per month (Figure 7a). This reduction was primarily driven by a reduction in elective work (Figure 7b) (Ordinary one-way ANOVA: p<0.0001).

Figure 7. 7a (left) Operations per month and 7b (right) total emergency and elective operations from 2016 to 2020.

The number of total operations carried out by consultants and juniors (includes trainees and non-trainees) was globally reduced (Figure 8). This was primarily driven by a reduction in elective operations for all grades. However, a modest reduction in emergency cases was also seen for juniors.

Figure 8. Number of operations performed in the Mar-June period from 2016-2020 broken down by emergency/elective and grade of operating surgeon

The number of operations performed by trainees followed a similar pattern to that of all juniors (Figure 9). Currently, to obtain a Neurosurgical Certificate of Completion of training (CCT), trainees are required to obtain a minimum of 1200 operations and the formal duration of training is eight years. To assess the magnitude of this reduction, we estimated that if such an operative rate was sustained achieving CCT would be delayed by a further 1.5 to 4 years (Figure 9).

Figure 9. Total number of operations performed by trainees with projected time to completion of CCT.

Principal findings

The COVID-19 first wave and lockdown in the UK spanned the time period of 8^th March to 7^th June. The response to the first wave has highlighted a substantial lack of preparedness within the NHS. Further waves have occurred since this time and despite a substantial national vaccination campaign, a further winter wave in 2021 has occurred⁶. Our review of the impact of the first wave on our trust and the reaction to it demonstrates a significant hit to admissions, operative throughput, and training. Recalling these consequences is crucial at this juncture with future pandemics or epidemics likely. There is therefore a need to reiterate the importance of protecting acute services such as Neurosurgery.

During the first wave, we found a very minimal reduction in referrals from local hospitals (Figure 2), despite nation-wide reduction in all accident and emergency (A+E) activity^8,9. Any minimal reduction may have been contributed to by a change in referrer practice, a change in the pathologies occurring e.g. reduced trauma from reduced traffic during lockdown¹⁰, or public fear and avoidance of hospitals. Such a reduction in referrals may not occur in the face of future waves if the public become desensitised to life under repeated waves.

The response to the first wave caused a substantial reduction in admissions, including emergency cases (Figure 3). Concomitantly, there was a reduction in the average length of stay (Figure 4). While we cannot be certain what caused this pattern, potential causative factors include: stricter triage of referred patients and/or increased use of on-call advice provision to assist with remote care of patients at their local referring hospitals. This is an area for future investigation to assess whether this apparent increased efficiency was detrimental to patient care or a useful side effect of the stresses of the pandemic.

While the total number of operations did significantly decrease, the number of emergency operations did not (Figure 7). In the context of a modest reduction in referrals and a significant reduction in admissions, a stable rate of emergency operations likely reflects that urgent cases were still able to obtain necessary treatment.

Our data demonstrate a reduction in theatre efficiency as theatre downtime increased and operations per theatre session decreased (Figure 5 and Table 1). Since patient time in theatre and time of knife-to-skin did not change, the reduced efficiency is likely caused by perioperative processes. Namely, the increased precautions and use of personal protective equipment, anaesthetic time, occasional need for patient post-op recovery to occur in theatre and more rigorous cleaning of facilities between cases. An awareness of this effect is needed to permit future resources to allow more efficient use of theatre resources, such as dedicated post-operative recovery areas to not delay further operative throughput.

As expected, all pathology subtypes faced a reduction in elective work (Figure 6). However, urgent work did not decrease significantly across paediatric and oncology subspecialties. From observational assessment of our practice, we believe this may be due to a practice of admitting oncology or paediatric patients for treatment on the emergency list where their condition is urgent but not an emergency.

The number of emergency cases performed by consultants did not significantly reduce while those being performed by junior trainees did reduce. This reflects a trust-wide drive to increase consultant-led care and the redeployment of juniors for other clinical purposes. Furthermore, anecdotally, where juniors did operate this was primarily for emergency cases with very few complex or index cases performed by trainees. This low rate of operating would create a substantial deficit in the training and surgical acumen of trainees. To estimate the size of this effect, we calculated that if such an operative rate were maintained, achieving completion of training would add an additional 1.5 to 4 years to the 8-year neurosurgical training period (Figure 9). To ameliorate these effects, in our institution cadaver microscopic dissections were formally introduced. Trainees were allocated time slots as they would for theatre slots and encouraged to use this facility in a structured way to maintain and progress their microscopic skills and anatomic knowledge. Other possible remedies include the increased exposure of trainees to operating that occurs in the independent sector.

In the context of relevant literature

The effect of the COVID-19 pandemic on neurosurgery has been palpable globally with major changes to practice¹. Neurosurgical institutions have adapted with various changes including stricter triaging of patients^2,3, suspending all or most non-urgent or elective cases¹¹, increased testing^2,12 and even reductions in bone drilling speed to limit aerosol generation². Similar changes have been implemented specifically for spinal services^12,13.

Although we found a minimal reduction in the number of referrals, more substantial reductions in referrals have been reported elsewhere¹⁰. Furthermore, while our study demonstrates that emergency cases were not significantly affected, the impact on non-emergency urgent cases was detrimental (Figure 6). This is further corroborated by a reduction in the provision of necessary surgery for malignant brain tumours across the UK¹⁴.

It is critical to also note the severe impact of the pandemic on neurosurgical training. While the pandemic has led to some positive educational changes in the UK and globally including more online teaching, an increased use of open access resources¹⁵, reduced working hours with more time for activities such as research^16–18 there is nonetheless evidence of an expected detrimental effect on trainee morale and wellbeing¹⁹. Furthermore, the effect on operative experience cannot be underestimated and future protections for training using ancillary resources such as simulation should be encouraged and supported²⁰. Other avenues include the increased use of the operative resources in the independent sector. Additionally, changes to training completion criteria could be considered. This may include, for example, an active change to a focus on index cases and the degree of independence of the trainee as opposed to operative numbers alone when assessing the trainee’s progression.

Strengths and limitations

The principal strengths of our study include the contemporaneous capture of extensive data to understand the neurosurgical patient journey through the region. In addition, through the use of electronic databases, we were able to capture key parameters including the grade of operating surgeon and therefore present a quantified estimate of the effects on neurosurgical training.

Limitations of our study include the method of assessing the pathology subtype for the operations performed by classifying according to the subspecialty of the named overseeing consultant. Also, while our study demonstrates important lessons for the future response to further infection waves, it is important to consider that the rates of COVID-19 infections in the East of England region were amongst the lowest in the UK.

Underlying data

Harvard Dataverse: Re-visiting the impact of the first wave of COVID-19 on Neurosurgical Practice and Training in a Large UK Neurosurgery Unit. https://doi.org/10.7910/DVN/EXRTD7²¹

This project contains the following files:

Data are available under the terms of the Creative Commons Zero "No rights reserved" data waiver (CC0 1.0 Public domain dedication).