Autumn update: The impact of COVID-19 on UK hospital departments

Introduction

We usually present and discuss our benchmarking data for our client hospital organisations at our twice-yearly events in spring and autumn. These benchmarks allow departments to examine key metrics relating to their service delivery and to compare these with other departments in their cohort.

Due to 2020’s unpredictable and turbulent nature, our usual metrics are perhaps less relevant as NHS departments look to examine the impact of COVID-19 on service delivery. Here we instead provide an update to our May report on anaesthetics departments and COVID-19, extending this to look at all UK hospital departments using CLWRota or Medirota between October 2019 and October 2020, which includes about 19,000 staff. CLWRota is used to roster anaesthetists and Medirota is used to roster surgical and medical staff.

We focus on the impact of COVID-19 on our users’ service delivery and how departments rapidly adapted their rotas to meet changing demand, in particular looking at:

Frequency of Rota Changes

Analysing rota changes over the past 12 months indicates how rapidly departments responded to COVID-19. A rota change is defined as anytime a rota master assigns or removes a person from a session on the rota.

In April there was a dramatic increase in the frequency of rota changes, doubling pre-March levels as departments rapidly altered their rotas, in many cases using the tools described in our COVID-19 assistance guide here. This was in response to demand which required NHS departments to restructure their usual service delivery, redeploying clinicians outside of their usual patterns of activity.

Figure 1. Line graph showing the number of rota changes in Medirota and CLWRota where a rota change is defined as any time a rota master assigns or removes a person from a slot on the rota.

The frequency of rota changes has remained above pre-March levels across both services, with Medirota departments seeing a slower return to the baseline frequency. This could be a result of re-scheduling elective clinical and surgical work.

Rota sessions per person: CLWRota

A key change over the past 7 months was the shift in distribution of work from in-hours to out-of-hours. The stacked graph below shows the average weekly rota sessions worked per person. A CLWRota rota session is a slot in the day – either “morning”, “afternoon”, “evening” or “night”.

Figure 2. Stacked graph showing weekly rota sessions per person in and out of hours in CLWRota. The figures are averaged over a 7-day rolling window.

In April, a large proportion of clinicians’ in-hours (mornings and afternoons, Monday-Friday) work was replaced by out-of-hours work. The balance between in-hours and out-of-hours work now resembles pre-March figures again, after a decline in out-of-hours work in June and July followed by the restoration of in-hours work in mid-August.

Figure 3. Heatmap showing the average per-person activity in April in CLWRota broken down by time of the week, as a percentage change compared to January-March 2020.

The heatmap above, which compares April to the start of the year, further highlights the redistribution of work across the week. We see a 20-30% drop in in-hours activity and an increase in out-of-hours activity peaking at 92.9% on Sunday am sessions.

Rota sessions per person: Medirota

For Medirota departments, there is a less significant change in the distribution of work. This might be because Medirota data involves a mix of specialities with varying levels of involvement with COVID-19 related work.

Figure 4. Stacked graph showing weekly rota sessions per person in and out of hours in Medirota. The figures are averaged over a 7-day rolling window.

In-hours work only saw a small, brief drop in April, and there is once again an increase in out-of-hours work.

Figure 5. Heatmap showing the average per-person activity in April in Medirota broken down by time of the week, as a percentage change compared to January-March 2020. Note that April included a bank holiday Monday and Friday for Easter.

Cancellations

The number of cancellations recorded in CLWRota and Medirota gives an indication of the large scale of rescheduling of planned work as departments adjusted to deal with COVID-19.

The graph below looks at all cancellations across both systems and cancellations for which COVID-19 was noted as the reason are represented by the pink area. After a dramatic increase in April, the rate of cancellations steadily dropped until October but have since started rising again. Currently, the cancellation rate is almost twice as high as it was before March.

Figure 6. Count of total sessions cancelled per week in Medirota and CLWRota. The pink area shows cancellations where the recorded reason mentioned COVID-19.

The x‑axis represents the date on which the cancelled sessions were scheduled to take place (rather than the date they were cancelled).

Cancellations by speciality: CLWRota

By breaking down cancellation data from CLWRota by speciality, we get an idea of which areas of delivery have been prioritised.

In the graph below, each of the 15 most common specialities on CLWRota is represented by a grey line showing the percentage of sessions of that speciality that were cancelled per week. ENT saw the highest percentage of cancellations which may be due to the high risk nature of the work involving close contact with the airways and the fact that it involves fewer urgent cases than many other specialities.

Figure 7. Line graph showing the percentage of sessions cancelled for the 15 most common specialities in CLWRota, with 4 of these highlighted.

In comparison, specialities such as ICU, emergencies and obstetrics saw the fewest cancellations as these are essential services which couldn't be temporarily suspended.

This data highlights the scale of missed theatre work throughout the pandemic and the resultant pressure on the NHS to catch up on missed work in upcoming months.

Leave booked

Annual leave bookings were severely impacted by COVID-19. The graph below looks at consultants' leave data across both our systems who have an April-to-April leave year, and shows the percentage of their total yearly annual leave allowance had been booked by a given point in the year. Data is shown from the current and last leave years for comparison.

Figure 8. Line graph showing the cumulative percentage of yearly annual leave allowance booked by Consultants in CLWRota and Medirota with an April-April leave year. The figures are averaged over a 7-day rolling window.

The graph shows a dip in March 2020 as annual leave was cancelled in response to COVID-19. Bookings start increasing again around June but at the same rate as the previous leave year, meaning the gap between current and normal amounts of booked leave is unlikely to close.

This has implications for clinicians who have had less time off and for trusts who will need to pay back leave or carry over leave allowances while at the same time moving service delivery back to pre-COVID levels.

Sickness and COVID-19 related leave

Despite annual leave bookings dropping, departments had fewer staff available to work as a result of COVID-19. The graph below shows how the number of people per day that were unavailable to work due to sickness or COVID-related reasons spiked sharply at the beginning of April. These types of absence have recently begun to gradually rise again, and currently 44% more leave of these types is being recorded than at the start of the year.

Figure 9. Count of people on sick and COVID-related leave per day in CLWRota and Medirota averaged over a 7-day rolling window. The pink shaded area represents leave for which the user-inputted details mention "covid" or similar

Questions

If you have any questions about the above ideas or would like to know more about how to get reports from your system please contact the Rotamap support team at support@rotamap.net or 020 7631 1555.

Notes

The data covers all NHS departments who have been using Medirota and CLWRota since October 2019 and covers the date range of October 2019-2020. The identification of leave and cancellations related to COVID-19 is not built into CLWRota and Medirota so it relies on processing free text input by its users. This is inevitably subject to variation between departments' procedures.

articles data