Uptake and accessibility of surgical robotics in England

The distribution, utilisation and accessibility of surgical robotics in England is unknown.


| FOI Act requests
One hundred and forty-nine acute hospital trusts were identified from Estates Returns Information Collection data, publicly available from NHS Digital. 15 FOI Act requests were sent to all 149 acute NHS trusts on the 14th November 2019 in accordance with the FOI Act 2000. 16 The FOI Act gives the public the right of access to information held by public authorities. Members of the public are entitled to request information from public authorities and should be responded to within 20 working days. Trusts were asked to provide information concerning three broad areas: first, the number of robots in use; second, the number of procedures performed annually between 2013 and the date of the request and finally, the cost of the robot, disposables and maintenance. Whilst public authorities are obliged to respond to FOI requests, they are able to refuse requests should the request be deemed to take too much staff time to complete. 16 We therefore chose to limit our request to the specialties in which the most robotic procedures are performed. 17 We acknowledge the use of robotic approaches in other specialties including ENT, 18 cardiothoracic 19 and orthopaedic surgery. 20 The full FOI request is available in the Data S1.

| Spatial accessibility to surgical robotics
The locations of hospitals with surgical robots and travel time to nearest robot for England are mapped. Accessibility to surgical robots is calculated using the Enhanced Two-Step Floating Catchment Area (E2SFCA) Method. All results are described and represented spatially using choropleth maps.

| Travel time data
The traffic-free road travel times between all trusts with surgical robotics and the population-weighted centroids of all LSOAs in England were obtained from https://pythonhealthcare.org/ and supplemented with additional Open Street Map queries (openstreetmap. org) using the MapBox REST API (mapbox.com). Where trusts were composed of more than one site, the site of the surgical robot was further clarified by amendment to the original FOI request.

| Enhanced Two-Step Floating Catchment Area
Accessibility to surgical robotics was calculated for each LSOA in England using the E2SFCA method, 22 as follows: Step 1: In the first step, the supply: demand ratio (R j ) is calculated as the ratio of the number of surgical robots to the expected demand from the nearby population: � S j is the number of robots in the site.
� D k is the demand of the location -in this case is the crude LSOA population in 2018.
� Gðt kj ; t 0 Þ is the Gaussian distance decay function for the travel time between LSOA k and Hospital j (see below).
A Gaussian decay function was defined to preferentially weight populations closer to a hospital than those further away, as described in, 22 as follows: � t kj is the travel time between hospital j and LSOA k.
� t m is the outer limit of the region of LSOAs contributing cases to the hospital j.
In this study, an upper limit (t m ) of 60 min travel by road was taken as an appropriate maximal threshold. At 40 min a weight of 0.5 is applied. Outside of the 60 min catchment radius, all LSOAs receive a weighting of 0.
Step 2: In the second step, the accessibility of LSOA k (A k ) is calculated as the sum of the capacity:demand ratio (R j ) of all hospitals within 60 min road travel after weighting by the same Gaussian decay function as Equation (2).
� A k is the accessibility of surgical robots for LSOA k.
� R j is the capacity:demand ratio of hospital j, from the set of J hospital trusts with surgical robots.
� G(t kj,tm ) is the Gaussian distance decay function for the travel time between LSOA k and hospital j.

| Mapping of robotic centres
The geographic location of robotic centres in England are shown in

| Procedure data
Procedure data supplied by the 41 responding robotic centres are detailed in Table 1

| Spatial accessibility to robotic surgery
Travel time to robotic centres were determined and mapped in  3 and 4).

F I G U R E 1
Robotic centres in England-size of circle corresponds to number of robots LAM ET AL.
-3 of 7 Accessibility to robotic centres calculated using the E2SFCA method varied across England and was highest in the North East of England, the South West of England and in Greater London ( Figure   5). With the exception of Greater London, areas with high spatial accessibility to robotic surgical services were located in large towns or small cities with surrounding rural communities. In these cases, high accessibility is conferred by the presence of one or two robots in relatively sparsely populated areas, rather than many robots in a conurbation. The West Midlands, despite being near to many robotic surgical centres has lower spatial accessibility due to its larger constituent populations. 4.3% of the population of England (2 412 008 people) have zero accessibility to robotic surgical services, owing to the travel time to their nearest robotic centre exceeding 60 min.

| Economic data
Economic data received were limited by the fact that many trusts  Due to this competition, it is likely we will see innovation in business models and a move towards cheaper licencing and subscription models permitting access to the machine and disposable instruments. To date, five of the 22 trusts who reported robot cost data have opted to lease their robot rather than own them outright.
In keeping with the increasing number of surgical robots, there has been a rise in robotic surgical procedures performed in England.
However, there is significant variation in the volume of robotic procedures being performed between centres. In 2018, the total procedure volume per hospital provider ranged from 1 to 683 cases.
In centres performing robotic urological surgery the median number -5 of 7 unlikely to be safe or cost effective. A centralised approach can also deliver economies of scale. The range in the prices paid for robots, disposables and maintenance, and the number of trusts exempting to disclose financial data for commercial reasons, suggest that handling these purchases at a national level may be able to secure a better deal for the NHS as a whole. NHS strategy states that there is a desire to move to a more centralised procurement model, and there is a precedent for doing so. 24 Centralisation of procurement strategy for robotics is likely to represent better value for the NHS. Moreover, these centres will be able to properly assess and compare the merits of novel and competing robotic systems as they enter the market in the NHS. Implementation of a centralised robotic surgery, however, will not be without challenge. Trusts currently have licence to operate independently and funding is variable between trusts with many receiving charitable support. Centralisation will therefore require a national strategy. Encouragingly, there is precedent in the centralisation of specialist cancer surgery and specialist vascular surgery services; work would need to be undertaken to understand lessons learned from these projects and to apply them to centralisation of robotic surgery.
This strategy would also address the issue of learning curves in the adoption of robotics. Present learning curve estimates cannot be relied upon due to poor reporting standards and the significant degree of heterogeneity in methodology. 25 Minimum oncological resection standards exist in laparoscopic surgery for resections, but this has been much harder to properly assess in robotic surgery.
National robotic learning curve standards must now be urgently defined and this cannot be achieved without national coordination.
Covid-19 has fragmented surgical services and caused significant disruption to surgical volumes and education. The current robotic framework has little resilience and it is unlikely to maintain robotic surgical learning curves for the foreseeable future.
The National Bowel Cancer Audit in 2019 26  currently entering the market is likely not to be in the robots themselves but the data that comes from them. This data and its sharing will serve as the platform which may deliver the leap in clinical outcomes which robotic surgery has promised for so long but failed to deliver.
Whilst we have received a response from all acute NHS trusts surveyed, our study is limited by a fragmented data set with incomplete procedural and financial data. We are unable to make any economic conclusions with our data beyond descriptive statistics given fewer than 25% of respondents replied with full economic data. This study has mapped the real-world uptake of robotic systems in a public healthcare system. The lack of a national robotic strategy has led to variable accessibility to surgical robotics with decisions made only at a trust level. The issues caused by the absence of a national robotic registry have been demonstrated by the necessity to send individual FOI requests to every NHS trust. However, this study has also shown the use of FOI requests as a powerful means of collecting data that can be used to inform health policy and national surgical strategy.

| CONCLUSIONS
This is the first nationwide study to map out the extent of adoption of robotic surgery in a large public healthcare system and has demonstrated that the number of robots and robotic procedures continue to rise. Accessibility to robotic services is generally good, but certain areas of England such as the North West and the East Midlands are relatively underserved. The future of robotic surgery across England may lie not in further one-off acquisitions of robots by individual trusts, where the number of procedures may be relatively low, but in high volume robotic centres. However, further safety and economic analysis is required to determine whether this strategy is indeed the future for robotic surgery in the NHS. To allow this analysis to be carried out in the future, we should ensure that data from all robots within the NHS are tracked at trust level and shared at a national level.
This will allow evaluation of the value of robotic surgery to the NHS and to English surgeons beyond observational data. National data will be crucial in order to define our future national robotics strategy.