Healthcare organization policy recommendations for the governance of surgical innovation: review of NHS policies

Abstract Background The governance for introducing innovative surgical procedures/devices differs from the research requirements needed for new drugs. New invasive procedures/devices may be offered to patients outside of research protocols with local organization oversight alone. Such institutional arrangements exist in many countries and written policies provide guidance for their use, but little is known about their scope or standards. Methods One hundred and fifty acute NHS trusts in England and seven health boards in Wales were systematically approached for information about their policies. A modified framework approach was used to analyse when policies considered new procedures/devices to be within local organization remit and/or requiring research ethics committee (REC) approval. Results Of 113 policies obtained, 109 and 34 described when local organization and REC approval was required, respectively. Procedures/devices being used for the first time in the organization (n = 69) or by a clinician (n = 67) were commonly within local remit, and only 36 stated that evidence was required. Others stated limited evidence as a rationale for needing REC approval (n = 13). External guidance categorizing procedures as ‘research only’ was the most common reason for gaining REC approval (n = 15). Procedures/devices with uncertain outcomes (n = 28), requiring additional training (n = 26), and not previously used (n = 6) were within the remit of policies, while others recommended REC application in these situations (n = 5, 2 and 7, respectively). Conclusion This study on NHS policies for surgical innovation shows variability in the introduction of procedures/devices in terms of local oversight and/or need for REC approval. Current NHS standards allow untested innovations to occur without the safety of research oversight and thus a standard approach is urgently needed.


Introduction
Innovation in invasive procedures, including surgery, is important. It is often driven by the desire to improve care and is encouraged 1 . Innovation ranges from completely new (first-in-human) procedures to modifications and adoption of existing techniques. Worldwide, the regulation governing innovation in surgery and invasive procedures/devices differs from the rigorous incremental standards required to introduce and modify pharmaceutical products. In the UK, this is via the Medicines and Healthcare products Regulatory Agency (MHRA), in the USA it is the Food and Drug Administration's (FDA) Centre for Drug Evaluation and Research, and in the European Union it is the European Medicines Agency (EMA). No such dedicated body or regulation for the development and evaluation of surgical innovations exists. Innovative invasive procedures and devices may be delivered to patients with local oversight via policies implemented by clinical effectiveness committees in the organization 2 . They may also be introduced within formal research studies with research ethics committee (REC) approval, although historically this has been uncommon and innovations in surgery have occurred without formal registration anywhere [3][4][5] . The distinction between local organization oversight and governance via REC is important as it has implications for patient information provision, and reporting outcomes, harms, and adverse events. Research structures provide a systematic framework for governing the evaluation and introduction of innovative procedures/devices. Under research settings it is required that patients are given written information and provided with a choice about receiving the intervention, adverse events are reported, and an appropriate sponsor who is responsible for the study is identified. Under local hospital governance requirements, it is less clear what standards are necessary and there is little known about existing policies. Limited publications examining these issues are available 2,6 . There are several examples of patient harm caused by procedures and devices being used in clinical practice before evaluation of safety and effectiveness are well known [7][8][9] . Understanding the remit of local organization policies for the introduction of innovation invasive procedures and devices is therefore important.
Guidance from national and international professional bodies [10][11][12][13] suggests that surgeons refer to local organization policies when introducing new procedures/devices. Although these exist across the NHS and worldwide 2, 14 , there has been no previous research found that has systematically or widely examined them. Previous work is limited to case studies of individual hospitals and qualitative investigations of surgeon decision-making in this area 2,6 . It is not known when new invasive surgical procedures/ devices are appropriate for introduction via local governance policies, and thus can be delivered to patients with local oversight, or when new procedures require evaluation within formal research studies with REC approval.
The INTRODUCE study 15 examined NHS organization governance policies for the introduction of new invasive procedures and devices. The aim of this paper is to describe when policies consider the introduction of new procedures/devices to be under local NHS organization clinical effectiveness committee governance remit and when they recommend that REC approval is sought.

Methods
A content analysis of policy documents using a modified framework approach was undertaken 16 provides a systematic and rigorous approach involving coding of text into descriptive themes to form a framework in an iterative process of constant comparison.

Sampling and data collection
All acute trusts in England (n = 150) and seven health boards in Wales were approached and asked for governance policies for introducing new invasive and surgical procedures between November 2017 and November 2018. Document inclusion criteria are described in the protocol 15 .

Data extraction and analysis
The content analysis started with careful reading and re-reading of all policies by one reviewer (H.S.R.) to ensure familiarization and an in-depth understanding of the data. Verbatim text describing when 'new' invasive procedures/devices could be delivered with local NHS organization governance and when instead REC application was recommended was extracted and imported into a data-management program (Microsoft ® Excel, version 2015). This formed two data sets. The number of policies contributing to each were counted. A second reviewer independently extracted data from 20 per cent of policies (S.C., J.B., J.A.M., H.R., and J.Z.), rather than the 10 per cent specified in the protocol 15 because initial reading revealed very heterogeneous text that required careful categorization. The use of multiple reviewers to extract data ensured that mistakes (e.g. missed data) were minimized and reduced the risk that data selection was influenced by a single-person biases 18 . Disagreements were discussed until consensus was reached.
For each data set, individual themes and subthemes were developed from iterative inductive 'open' coding process performed iteratively by three reviewers (H.S.R., S.C., and J.B.). Where coded text could not be allocated to an existing theme, a new theme was created. Overarching themes were created by grouping together similar individual themes. 'Combined overarching' themes, linking two overarching themes, were created if descriptions of procedures and devices within local organization remit or requiring REC application contained the conjunction 'and', linking two thematically discrete elements. The approach was 'modified' in that the content of the policies was examined only with reference to the specific research question (i.e. how policies described 'new' procedures and devices within local NHS organization governance remit and those recommended for REC application).
The final frameworks including all coded text for all overarching and combined overarching themes, individual themes, and subthemes were reviewed for a final time (H.S.R., S.C., and J.B.), and then all policies re-read (H.S.R.) to ensure relevant text had been comprehensively extracted and coded.
Descriptive statistics summarized the characteristics of organizations approached and those with policies, including commissioning region, foundation status, and acute trust type, and information about the policy issue date and planned review date/expiry dates were extracted.

Patient and public involvement
The concept and design stages of the study were informed by two public engagement events, hosted by the surgical innovation theme and the patient and public involvement group of the Bristol Biomedical Research Centre (BRC) on 20 March and 9 November 2019. The first took place at a shopping centre near Bristol University and the second at an annual supporters' celebration for university alumni. Both events attracted members of the public and provided an opportunity for people to consider innovation in surgery with interactive life-size 'operation' games ('Operation Game' (Altitude Events, Felbridge, UK) and 'Beat the Surgeon' (University of Bristol, Bristol, UK)) and posters describing the issues, alongside members of the BRC dressed in 'scrubs'. Opportunities to respond to questions around surgical innovation were given, including preferences for undergoing surgery with robotic technology or standard methods, whether information about the surgeon's experience of a new procedure was required before receiving a new operation, and what was understood by the term 'innovation'. People more often voted for new operations in place of standard procedures, but raised issues related to safety, levels of previous testing/evaluation, and surgeon experience. Individual feedback on study findings and recommendations described in this paper was provided by two members of the BRC patient and public involvement group. Both patient representatives were surprised at the variance between policies and highlighted how this contrasted with the introduction of new drugs, questioning why there were such differences. Regarding the recommendations made in the paper, they stated that standardization was important and felt it would be beneficial. Concern was expressed about creating barriers to innovation, especially if different processes required more REC applications, which may discourage surgeons from engaging in the process.

Response rates, organization characteristics, and policy review dates
Of the 157 organizations approached, 20 said they did not have a policy and nine did not respond (Fig. 1). There were 113 policies eligible for inclusion. Policies were received from across all parts of England and three obtained from Wales ( Table 1). The majority were from foundation status trusts, with policies from all sizes and types (e.g. teaching hospitals) of trust. Fifty policies were out of date (i.e. the planned review date had passed) on receipt and five did not report a planned review date. The average period for planned reviews and updates of policies was 33 months (minimum 12 months, maximum 134 months).

Overarching and combination overarching themes
Six overarching and seven combined overarching themes were identified (detailed below) from descriptions in policies about when procedures/devices may be introduced via local NHS organization governance (included in 109 policies), and when they instead require REC approval (n = 34). All verbatim text coded to all themes is given in the supplementary material (Tables S1 and S2).

Personnel
The expertise and skill of the proceduralist and team was often used by policies to place procedures/devices within the local NHS organization governance remit (n = 81; Table 2). Most said that procedures being undertaken by a clinician for the first time (n = 67) should be brought to the local committee (Table S3). Some policies stated that the procedure/device being undertaken by the clinician for the first time would only be within local governance remit if it was also 'established' (n = 5; i.e. within the combined overarching theme-Personnel and Evidence) or being delivered for the first time in the organization (n = 3; combined overarching theme-Personnel and Place ( Table 2)). The need for additional training (n = 26), delivering the procedure/ device in an extended role (n = 18), or those personally developed by a clinician for an individual case (n = 6) were also specified as being suitable for approval by local NHS committees (Table S3). Policies less commonly used the expertise of the proceduralist/team as a rationale for recommending REC application ( Table 3). Two policies said if the procedure/device required additional training then it should go to a REC (Table S4).

Place
First-time delivery of the procedure/device in a location was used frequently by policies to signal when procedures/devices could be introduced with local governance (n = 78; Table 2). This was most commonly when it was first-time delivery in the organization (n = 69; Table S3). Some policies said that procedures/devices being delivered for the first time in the organization were only within the local remit if they: had an evidence base (n = 13) or were established in clinical practice (n = 10; combined overarching theme-Place and Evidence); were significantly different from current practice (n = 3; combined overarching theme-Place and Procedure); had financial implications (n = 2; combined overarching theme-Place and Economic) ( Table S3). Importantly, none of the policies defined these terms. Clinicians wishing to deliver procedures/devices for the first time in the NHS or for the first time anywhere were advised to apply to the local committee for approval, in seven and six policies, respectively. In contrast, seven policies said procedures/devices being used for the first time anywhere should be assessed and approval given by a REC ( Table 3 and Table S4). One policy stated if procedures/devices had not been delivered/used in the organization before they should seek REC approval, but only if they were not established techniques/devices (combined overarching theme-Place and Evidence). Again, in these policies no definition of what was understood by 'an established technique' was provided.

Procedure
Proposed changes to components of the procedure/device itself could be introduced via the local approval in 70 policies ( Table 2). Forty-two said that procedures/devices where major modifications were planned were within their remit, and nine gave examples of what a major modification would constitute (Table S3). Procedures/devices being used for a different indication (n = 20), in a different part of the body (n = 5), or combined with other treatments (n = 3) were also described as appropriate for introduction via local governance processes. The need for enhanced or modified consent processes (n = 4) or whether different aftercare was required (n = 1) was also used as an indication as to whether the procedure or device was within local NHS organization governance remit. Descriptions of how procedures/devices may have changed were not used by any policies to indicate that REC approval should be sought, except in one policy, which only said this was needed if the procedure involves gene therapy (n = 1; Table 3).

Evidence
Just under half of the policies (n = 42; Table 2) used the degree of underpinning evidence as a reason for the procedure/device to be introduced with local governance. Policies described procedures/devices with 'uncertain' or changed outcomes (n = 28), those being used in research studies (n = 17) or prior to the start of a study (n = 1) as being appropriate for introduction with local governance approvals (Table S3). No policies expanded on what was meant by uncertain or changed outcomes. The degree of underpinning evidence was highlighted by 13 of the 34 policies that gave recommendations for when REC approval should be sought ( Table 3). Policies similarly talked about procedures/devices with 'insufficient' evidence (n = 6) and 'uncertain' or changed outcomes (n = 5) but recommended that these required REC approvals, rather than local governance, before delivery in the organization (Table S4). Only two policies suggested what level of evidence would be considered 'insufficient' (i.e. if the procedure was not in use elsewhere as described in a peer-reviewed publication or if there was no evidence supporting use in nationally or internationally recognized centres; Table S2). Four policies used terms like 'original', 'completely new', and 'experimental' to describe procedures/devices requiring REC application, but none defined what was meant by these. Two policies said that if the procedure/device was not established and its delivery/use had not been notified to the National Institute for Health and Care Excellence's (NICE) Interventional Procedures Advisory Committee (IPAC) then REC application was necessary (combined overarching theme-Evidence and External guidance).

Economic
Ten policies said that where the introduction of procedures/ devices had financial implications they should be brought to the committee ( Table 2). No policies used financial/resource implications as a reason to apply for REC approval.

External guidance
Three policies said that procedures/devices recommended for use by NICE were appropriate for introduction by the local policy ( Table 2). Policies placed more emphasis on external guidance

Inconsistencies across policies
For some procedures/devices conflicting recommendations regarding necessary governance approvals were given across policies ( Table 4). Twenty-eight stated that procedure or devices with uncertain or changed outcomes were within local NHS policy remit (Table S3). However, two of these, along with three different polices, also recommended REC application in these circumstances (Table S4). Twenty-six policies said procedures/ devices requiring additional training were within local NHS remit, and two others recommended REC application. Finally, six policies stated procedures/devices not previously used anywhere could be introduced via the local NHS organization policy, but seven others recommended REC application.

Discussion
This study has systematically examined policies for the introduction of new invasive procedures and devices into NHS organizations and clinical practice across England and Wales. Specifically, it examined guidance for when local NHS organization oversight and/or REC application is recommended. Policies were obtained from 113 of 157 hospitals approached, with 20 reporting the absence of a written policy. Whereas almost all described when new procedures and devices may be introduced with local organization oversight, only a third gave guidance on when REC application was advised. Detailed analyses identified considerable variation and inconsistencies about what governance (local organization approval versus REC) was required, and some common themes were identified. The placement of novel procedures/devices under local governance was largely determined by the experience of the personnel (e.g. if a clinician was delivering it for the first time in the NHS) or whether they were being delivered for the first time in a given location (e.g. first time in the organization). In contrast, recommendations to seek REC approval were mainly based around external guidance (e.g. NICE) and evidence underpinning the procedure. Policies lacked details about required degrees of published evidence to support delivery of a new procedure.
Definitions of important terms such as 'uncertain', 'sufficient', and 'experimental' were limited. The ambiguity and inconsistencies observed in the documents highlight an important gap in the governance of innovation in surgery. This has direct implications for patient safety. It means that untested procedures can be introduced and delivered in clinical practice with local NHS oversight alone. Without the framework of research regulation patients may not be informed of the innovative nature of procedures and that may limit their choice. It also means that outcome and safety data may not be collected and reported. Although examples of patient harm caused by innovative procedures and devices are well known and published, the full extent of the problem is unknown [7][8][9]19 . This analysis of NHS policies identifies the need to improve the governance and regulation of innovation in invasive procedures and surgery. Worldwide 10,12,13,20 , there are similar approaches to those in the UK for the introduction of innovative surgical procedures. Publications from the USA 21,22 , Australia 23,24 , and Canada 6 include recommendations for local hospital review if a procedure has significantly changed 6,23 or if it has not been delivered in the organization 6,24 or on a large scale 21 previously. Two of the papers proposed a staged rollout of innovative procedures, overseen by the local committees 21,23 and one 22 recommended that interventions that were first in human and/ or high-risk procedures, and those with no outcome data to be overseen by research (i.e. an FDA-regulated Institutional Review Board). All the publications and UK guidance lack clarity about when innovations should be delivered within a research study with ethical approval. They are also unclear about what is understood by 'new' procedures that fall within local remit. Understanding what surgical innovative interventions require research oversight is therefore important. The Health Research Authority in the UK does define what constitutes research. It says that research involves changing treatment/care services from accepted standards 25 and stipulates REC approval is required if this involves NHS patients 26 . However, no further detail is given regarding the necessary magnitude or nature of change. It defines research based on its characteristics, such as the use of randomization, and the deliberate use of methods intended to collect quantitative or qualitative data 27 . This is similar to the US Department of Health and Human Services, which defines research as systematic investigations designed to develop or contribute to generalizable knowledge 28 . These definitions of research currently overlap with the recommendations from local hospitals about new procedures making it confusing to surgeon innovators about what oversight is appropriate. The inconsistent approach to governance for the oversight of innovative surgical interventions and devices in surgery is evident in systematic reviews of specific innovations. Reviews of novel techniques in minimally invasive transthoracic liver resection, magnetic augmentation of the lower oesophageal sphincter, and laparoendoscopic colonic polyp resection, for example, show variability in the number of observational and non-comparative studies reporting REC approval (between 25 per cent and 67 per cent) [3][4][5] . There may be several reasons why this happens. It may reflect historical  32,33 , and none of the guidance is mandated. The strengths of this report include the novelty in examination of these documents and the good response rate. Only nine of 157 organizations did not respond, reflecting the rigorous approach. This study is also detailed and the verbatim text policy content was analysed exhaustively using an in-depth qualitative approach. Text was coded by a second reviewer and there was extensive iteration between throughout analysis. All extracted and coded data are available in Table S1 and S2 for transparency. The study does have weaknesses: it is possible that organizations may have updated policies since this work was undertaken. Policy update review periods vary; however, over a third of policies received were technically 'out of date', and so it is unclear to what extent policies are updated regularly. It is unlikely that organizations updated these during the COVID-19 pandemic. This potential weakness must be balanced with the rigorous analysis conducted. The degree to which policies are practically utilized could not be ascertained by this study. If policies are implemented to the letter, there is huge variability across organizations. If practice differs from that outlined in policies, it would not be inappropriate to assume that the picture may, in fact, be worse. Finally, the current study only examined UK NHS policies. Given the lack of surgery-specific regulatory agencies worldwide, the inconsistencies found are likely to be mirrored in other healthcare systems.
The variation, lack of clarity, and contradictions identified across policies for the introduction of innovative procedures has, for the first time, identified a major risk within NHS governance systems. The development of standardized national guidance for surgeons and clinicians introducing new and modified invasive procedures is essential. Recommendations for improvement are echoed by the 2020 Cumberlege report 34 , which highlighted the disjointed nature of guidance across healthcare bodies. Guidance needs to have clarity about what procedures can be approved by local hospitals and what requires research ethics oversight. It should define clearly what magnitude of 'novelty' of a procedure/modification requires research approval, and what constitutes 'sufficient evidence' to allow a procedure or device to be used with local oversight, without ethical approval. Additional ways of managing patient information provision and outcome monitoring could be outlined when this is the case. Possible methods for monitoring outcomes could include the mandated use of registries 35 . It is the recommendation of the authors that local organizations only approve procedures with supporting published evidence of safety and efficacy. In these instances, it is possible that clinician and team training are sufficient for delivery. The authors also consider that local processes for registering the procedure and following up on patient outcomes need to be more rigorous. Work to make research processes for overseeing early-phase studies of surgical innovation more efficient is needed and clinicians require support to engage with these processes meaningfully. Informed debate and work with key stakeholders is underway. Ultimately, it is hoped that the provision of clearer guidance in all these matters will support innovation and safeguard patients, surgeons, and the healthcare providers.

Funding
This study was funded by the National Institute for Health and Care Research Bristol Biomedical Research Centre. The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care. The study funders had no role in the design of the study, data collection, data analysis, interpretation of data, writing of the report or the decision to submit the article for publication. editing. J.M.B.: conceptualization; funding acquisition; data curation; formal analysis; investigation; methodology; supervision; validation; visualization; writing-original draft.
Disclosure. The authors declare no conflict of interest.

Supplementary material
Supplementary material is available at BJS online.

Data availability
The study protocol has been published 15 and all analysed data are provided in the Supplementary Material.