Luers, ethics, research and service evaluation: even more shades of grey?

Journals of standing require all research to have been reviewed by an independent Research Ethics Committee (REC) and deemed ethical. Local Research and Development (R&D) approval is also required in the UK to confirm necessary resources and research governance arrangements. For interventional research involving patients, written patient consent to participate is expected. However, for other activities (service evaluation and audit) the rules are less clear: despite attempts to provide clear definitions and guidance on the difference between research and audit, the boundary remains blurred and its interpretation variable. The last decade has seen numerous changes, and a general increase, in the processes and governance around research and associated activities 1. The current scope, remit, composition and standard operating policies for RECs in the UK are defined in the document Governance Arrangements for Research Ethics Committees: a Harmonised Edition 2, known as GAfREC. The GAfREC document attempts to clarify which projects are research, requiring REC approval, and which are not, based both on UK legislation and Health Departments' policy. Thus, “RECs are not expected to consider applications in respect of activities that are not research, for example clinical audit, service evaluation and public health surveillance” but “Care providers owe a duty of care to users of their services. They are responsible for ensuring that ethical issues and risks in the course of the care they provide are considered.” In August 2011, the National Research Ethics Service (NRES) published a guide, Does my project require review by a Research Ethics Committee? 3. This document provides an algorithm to assist researchers, sponsors and R&D offices in determining the need for REC review and repeats that REC review is not required for clinical audit or service evaluation. It refers to a 2009 NRES document for definitions of categories 4 which can be summarised: research is an attempt to derive new generalisable knowledge (including both hypothesis generation and testing) by studies that ask specific questions, often involving interventions or collection of data additional to routine care; service evaluation observes practice solely to define or judge current care; clinical audit observes and evaluates practice compared with a predetermined standard. Research may include randomisation, while the other activities do not. The 2011 document includes additional categories of research exempt from REC review, including ‘research involving staff’: REC review is not required, or encouraged, though special arrangements can be made if material ethical issues are raised. The rights of staff and preventing coercion are relevant concerns here. A concept of ‘proportionate review’ has been introduced, without necessarily directly resolving these issues. Another broader issue, previously highlighted by Yentis, is that activity becomes regulated according to its title rather than the ethical issues raised 5-7. In addition to changes in process and governance, the prevailing opinion on what is ‘ethical research’ or ‘ethical to publish’ may also change over time. As an example, in 2001 Anaesthesia published a study evaluating a new CE-marked airway device in a series of patients without their consent 8. Submission two years later of an identical study evaluating a different device was rejected on the grounds of being unethical 9, 10. It is unlikely now that any journal would publish such a study without REC approval and patient consent, but at that time the paper rejected on ethical grounds by one journal was soon published in another 11. This illustrates that a request for an ethical ‘opinion’ is just that and may, despite available definitions and practice guidelines, lead to a range of opinions, even contradictory ones. Of note, it remains generally accepted that clinicians may ‘trial’ or simply ‘use’ new CE-marked devices without explicit patient consent, but only provided no data is collected for publication: i.e. that no effort is made to evaluate the device formally and disseminate the results to a wider audience for general benefit. It is interesting, therefore, to reflect on the recent paper by Kinsella et al., which evaluated four CE-marked non-Luer spinal needles in 400 patients and compared their performance with a Luer spinal needle in 100 patients 12. At the outset is should be made clear that the authors, before starting the study, correctly sought advice from the local REC, which judged it to be a service evaluation – a view endorsed by the local R&D department – and not requiring patient consent. The devices were used in accordance with manufacturer's instructions. The non-Luer evaluations had twice as many consultants performing the procedures and excluded patients predicted to be difficult (included in the Luer evaluations). The study was very thorough and careful, using clinically meaningful and important outcome measures. The results show less good performance with non-Luer devices overall: they failed more often, required use of multiple needles more often, and were judged by 22–75% of users to be inferior to the Luer procedure or needle and attracted many more negative comments. Although the results of this timely and clinically relevant study are of importance to the profession, its classification as a service evaluation and the decision not to seek patient consent can be challenged. The argument for the study's being a service evaluation relies in part on the fact that a change in practice to non-Luer devices for spinal anaesthesia was ‘externally mandated with a fixed deadline’. The NPSA Patient Safety Alert on Safer Neuraxial Connectors 13 indicates that hospitals ‘should implement a Purchasing for Safety initiative’ and it is clear that the NPSA's intent was to make all hospitals change practice. Compliance with NPSA guidance is taken seriously by regulators, including those responsible for Commissioning for Quality and Innovation payments and the National Health Service Litigation Authority. Further, the NPSA has encouraged local clinical evaluations recommending that “Trusts in England should continue with their plans to assess new neuraxial devices, to gain some familiarity with the functionality of these products” before determining which device will be adopted locally 14. The argument against classification as a service evaluation has several aspects. First, the NPSA alert states: “continued use of any noncompliant devices after the deadline should be recorded in the organisation's risk register” 13, making this a possibility: one currently adopted by most hospitals in England. Second, the study took place during January–April 2012, before the NPSA Alert's April 2012 deadline. Third, the evaluation has been published 15. Fourth, the evaluations were performed to assist the hospitals’ local decision-making process, collecting data primarily for the organisation, rather than for patient care. One pragmatic test for when ethical approval and patient consent is required is ‘whenever (because of performing the study) you are doing something to the patient you would not normally be doing.’ This necessity is increased when the intervention lacks patient benefit and especially if it exposes the patient to risk. Regarding the decision not to seek patient consent, following on from the regulatory argument above, it might follow that if the evaluation is deemed a service evaluation, patient consent is not required. However, the evaluations were performed for institutional benefit and (presumably) because of concerns that the new devices may not perform as well as Luer devices currently in use. It is difficult to imagine how patients might benefit from being included in the evaluation and the results confirm there was a risk of harm. The concept of patient autonomy means the default should be to seek consent wherever it is possible 16. The strength of the arguments either way might be reinforced by considering the risk of harm, the importance of the issue being explored and other available evaluation methodologies. For this study, the risk of harm is likely to be both infrequent and relatively small (repeat procedures, necessity for a general anaesthetic, rare risk of serious complication 17). However, not all consider these risks trivial: when confusion between Luer and non-Luer devices potentially leading to multiple procedures was recently highlighted, it generated another NPSA rapid response report 18. The matter under investigation is important: there is a dearth of independent clinical evaluations of the non-Luer devices. Finally, manikin studies might have gone some way to avoid the need for this study or reduce its size: manikin evaluation of what is an engineering change can mimic many of the processes involved in evaluating the new devices and has already highlighted issues with two devices that have subsequently been redesigned 19. Manikin studies have the potential to identify and eliminate poorly performing devices without any need to expose patients to risk. In this issue of Anaesthesia, Sharpe et al. also report a clinical evaluation of non-Luer spinal needles after REC, R&D and legal department review waived the requirement for ethical review or ‘explicit patient consent’ 20. The aim was to “evaluate all systems….to enable the Trust to make a decision on which system to purchase”. Although their overall findings appear considerably more positive than Kinsella's, only one of seven devices was judged to be as good as the Luer and almost 30% of users would not be happy to use one device again. The authors comment that procedures were performed only by senior anaesthetists ‘to minimise the risk to patients’ and offer their results but do not state which device they have chosen to purchase, and indeed add: “where possible, Trusts should evaluate these systems themselves”. Importantly, the authors did perform a manikin evaluation before clinical evaluation to try to eliminate any poorly functioning devices. Overall, this paper raises similar but perhaps lesser concerns. The issues raised are more than academic. Around the country, many hospitals have performed or are performing similar, probably smaller, ‘local evaluations’. Should these be subject to REC review? Should patients give consent before involvement? Do the results of Kinsella et al.'s study increase the requirement for these? Do the above answers differ if there is intent to publish the results of the evaluation 15? Do those not intending to publish have a responsibility to publish if they identify concerns with new equipment, and how should they go about this? Should anaesthesia organisations collate the results of these small local evaluations 21? Should Kinsella et al.'s study alter the way we approach the routine evaluation of other new equipment for local adoption whether driven by legislation (e.g. ‘safety’ intravenous cannulae), national policy (e.g. non-invasive cardiac output monitors) or local development (e.g. new airways, cannulae etc)? Definitions and opinions will continue to evolve and differ, and disagreements will continue to occur 22. Perhaps what is most important is that all those involved in evaluating new devices (whether research, service evaluation or clinical audit) observe best practice. Best practice certainly involves ensuring the device is safe before use, being trained in its use, and following the manufacturer's instructions for use. In many circumstances it will also involve discussing with the patient that a new device is being used, and in a minority, seeking REC and R&D approval. The interests of the patient exposed to the device/equipment should always be the prime driver in deciding the best action to take. While there may be a wider responsibility to ‘future patients’ and also an ongoing imperative to evaluate new products, this must not erode standards of ethics, consent or patient care. Finally, it can be argued that the current situation illustrates a failure of both manufacturing governance and current clinical research priorities that non-Luer devices have not been evaluated in a co-ordinated manner in manikin studies and appropriate clinical studies before being introduced into widespread practice 23, 24. Funding for such evaluation remains a major stumbling point. I represent the Royal College of Anaesthetists on the National Patient Safety Agency (NPSA) Safer Neuraxial Devices External Reference Group. The views in this editorial are unrelated to that role and do not represent the views of the group, the NPSA or the College. My department has received free non-Luer regional anaesthesia equipment for evaluation in a manikin research project. I have no financial interest or investments in any such company and never have had, neither has any member of my family that I am aware of.