Criminalising doctorsBMJ 2018; 360 doi: https://doi.org/10.1136/bmj.k479 (Published 01 February 2018) Cite this as: BMJ 2018;360:k479
All rapid responses
The quote of Don Berwick that ”Fear is toxic to both safety and improvement, and health systems must abandon blame as a tool” mentioned in Ladher and Godlee’s editorial indeed holds true and the point repeatedly makes itself clear (1). One of the all-time most spectacular Danish criminal law cases involving physician performance was concluded last month by the Danish Supreme Court (Case 225/2017, dated 28 March 2018) and holds remarkable parallels to the Jack Adcock case in UK, although very different conclusions were reached, with emphasis on the institutional constraints under which doctors work. The Danish ‘civilian law system’ (with core principles codified into a referable system of laws as a primary source) includes a specific ‘Law on Authorization of Health Professionals’ covering medical doctors, nurses, etc. Apart from a general requirement for health professional diligence stated in the law’s para 17 (commonly employed when patient complaints are evaluated by the so-called national ‘Disciplinary Board’ sometimes resulting in written reprimands) there is a Para 75 imposing criminal responsibility in terms of fines or imprisonment in case of ’grosser or repeated negligence’.
In brief, the case was as follows (2). A 66-year old male was admitted to the hospital emergency unit in the late evening during August 2013. For months, he had suffered from abdominal pain and was waiting for an abdominal CT for which he had been referred. He also took insulin for diabetes. Nonetheless, the emergency admission mostly was due to ‘social matters’. There had been a worsening of abdominal aching before his admission, but when reaching hospital the aching was the same as during the last months and the patient was administratively ‘triaged green’ (non-acute). Within the emergency unit, the patient was assessed by a young junior doctor ‘A’ who took the medical history and examined him. She noted that the patient self-administered insulin for his diabetes. As the admission test panel (blood pressure etc.) did not fulfil ward requirements for a glucose level test in all diabetic patients, ‘A’ asked the nurse staff to do the glucose test but did not document this request in the medical charts. During the court case, nurse staff witnesses could not recall this order but the order’s authenticity was not disbelieved. According to the record, however, ‘A’ prescribed further standard laboratory tests and noted that, for the time being, the patient should ‘continue his usual treatment’. Among others, this would imply giving insulin according to blood glucose levels. The patient was moved to the medical ward and ‘A’ continued other duties and waited for the assessment of the patient made by senior doctor ‘B’ from the surgical ward. Afterwards, there is a lack of clarity on what happened and medical records documentation is scant. The next morning the patient was comatose. It appeared that he was severely hypoglycaemic and could not regain consciousness. According to the records, during the morning, B had assessed the patient using records information. He did not notice the lack of blood glucose info but concluded that there was not any need for acute surgery. After a couple of weeks, the patient unfortunately died: autopsy findings suggested hypoglycaemia being a major contributing factor.
The Danish Patient Authority and the prosecuting authority took ‘A’ and ‘B’ to the district court as a criminal law case with reference to the abovementioned para 75. They argued that no explicit plan was made for blood glucose testing, and there was no reaction to missing blood glucose tests. By reference to witness statements on ward routines as well as ward regulations in force at the time, the district court found that ‘A’ had had reasons for believing that the blood test had been performed and that nurse staff would notify her about any deviating test results. Furthermore, the court emphasised the circumstance that the admission mostly was socially caused. Additionally, it was considered that ‘B’ had acted in accordance with usual ward routines and “could not be expected to consider all non-surgical matters” during the out of hours duty. Both were found not guilty. The case afterwards was taken to the District High Court by the prosecuting authority. Again, senior surgeon ‘B’ was found not guilty, but with reference to expert statements made by the Danish Medico-Legal Council, ‘A’ was convicted of ’grosser or repeated negligence’ (penalty of 5,000 DKr.) as the connection between abdominal discomforts, diabetes, and hypoglycaemia “should be considered basic medical knowledge”. ‘A’ appealed the ruling to the Danish Supreme Court. The court’s minority (3 judges) were in favour of upholding the District High Court ruling, but in line with the district court, the majority (4 judges) found ‘A’ not guilty. Presupposing that ‘A’ ordered the glucose test, the Supreme Court, like the district court, put weight on actual local routines and ward regulations as well as the subjective perspective of ‘A’ that she could duly “expect that nurse staff notified her in case of deviating test results and also could expect further blood testing be done before the patient’s meal-related dosing a few hours later”.
The case has given rise to a heated debate in the Danish public and political sphere and intensified fear of litigation, defensive medicine, and superfluous documentation requirements among medical doctors. There has been expressed no doubt that the patient received a treatment far below any acceptable standards. However, much debate has concerned whether one or a few front-liners in the healthcare sector should be held disciplinary accountable for “the failings of an overstretched and understaffed system” (cf. (1)). The Danish Supreme Court majority ruled against this latter conception. Anyway, even if fear of litigation and blame now seems to have peaked, it remains unclear to what extent the ‘system’ is going to learn anything from the avoidable death of a 66-year-old patient with diabetes and abdominal discomfort admitted to a modern Danish emergency care unit.
1. Ladher N, Godlee F. Criminalising doctors. BMJ 2018;360:k479. doi: https://doi.org/10.1136/bmj.k479
2. Birkeland S. [Lægegerning og strafansvar]. Dagens Medicin 2018 March 5. https://dagensmedicin.dk/laegegerning-og-strafansvar/
Competing interests: No competing interests
A proposal for the Implementation of Adcock Never Events in Education
Never Events are serious, largely preventable patient safety incidents that should not occur if the available preventative measures have been implemented. Examples are wrong site surgery, incorrect route of administration for chemotherapy and retained foreign bodies after surgery.
The culture and recognition of never events is embedded into NHS culture and Clinical Commissioning Groups monitor these events and report them publicly. When a never event occurs, it suggests that there maybe a potential for patient harm recurring within that Trust and a rapid root cause analysis is undertaken with shared learning within the organisation but also to the wider healthcare community.
The rulings on Dr Bawa-Garba and Nurse Amaro have shifted the medical and nursing profession to reconsider factors that led to the devastating outcome for the Adcock family as well as on the staff involved. The tragic death of Jack Adcock needs to lead to change in the NHS by empowering the public and professionals to understand risks to patient care when educational standards are not met.
There are three areas of risk that are apparent in the events that led to the tragedy that should have never happened.
1. Lack of an educational induction
Doctors in training rotate through many specialities and changeover days carry inherent risk. This risk is mitigated by staggering changeover dates and by ensuring senior supervision of new clinicians within a department. HEE has recognised that Induction is important to ensure that trainees who rotate through the specialities are ready to deliver the basic care needed within that speciality. The risk is further mitigated by ensuring that induction encompasses departmental guidelines and the management of commonly encountered emergency and life-threatening conditions. Further mitigation includes ensuring doctors in a new department are not placed to deliver emergency care within the first few days without close supervision.
2. Non-phased return to work
The Colleges and HEE advocate adherence to return to work guidelines. Doctors may take career breaks for many reasons including maternity leave and after illness and carer’s leave. Any trainee returning to work after a career break should initially be supernumerary so that they can be supported in their return to work. This requires discussion about individual training needs and support as well as mentorship from a nominated colleague arranged prior to the return. Risk maybe mitigated by ensuring that trainers and senior clinicians are present and supporting care when these individuals are initially on call as often the confidence to make clinical decisions maybe altered. To be involved in such a death on the first day returning from maternity leave is a concern for many trainees returning to work and is paralleled with the guilt of leaving a young child at home.
3. Lack of full clinical staffing
Clinicians and healthcare managers are aware daily that there are medical and nursing rota gaps. There are pay caps applied nationally and a workforce morale that has pushed healthcare workers away from healthcare as a career and it is difficult to fill these gaps. At present, there are safe staffing numbers for nursing, but no guidelines exist for medical safe staffing ratios. There is pressure on delivery of service and income and therefore mitigation is established to ensure that a service can continue to provide patient care. Safe staffing numbers need to be established rapidly to ensure that services are being delivered safely as the rota gaps continue to increase. Patients should be aware of safe staffing numbers for medical and nursing and these should be clearly displayed for all areas and any risks identified.
Any of these three factors compromise patient care unless mitigated and all three together cause concern for patient safety.
What is apparent from Jack’s death, is that there was system failure not only in the educational requirements but also in management, clinical, nursing and IT services. The concern remains within the professions is that this maybe replicated today in any NHS organisation
Reporting of never events has catapulted the understanding of risk mitigation to ensure that these incidents never happen and has embedded excellence in clinical standards to prevent these from happening. Implementation of never events that compromise patient safety when educational needs are not met may also accelerate the understanding that these educational processes are inherent to patient safety.
Educational Never Events
1. Trainees should never be delivering service within a department without an induction and clarification of supervision.
2. Trainees should never be delivering service without an individualised learning agreement when returning from a career break.
3. Trainees should only cover rota gaps if the risks are identified and risks signed off and held by Senior management.
Failure to deliver against these standards should be reported to the Director of Medical Education and these should be publicly available on an annual basis.
The GMC, NMC, HEE, NHSI and NHS Employers need to work collaboratively to prevent further tragedies from occurring and the rapid implementation of three new never events would mitigate a degree of risk.
This may improve the morale in the workforce who need reassurance that they are working in an environment that values them and embeds a culture of learning and not of blame.
Competing interests: No competing interests
The UK government has ordered an urgent review of the law on criminal negligence but the criminalisation of medical mishap is happening in many other countries. In France, for example, a nurse from Saint Vincent de Paul Hospital in Paris was placed in police custody for two days over Christmas 2008 following the death of a 3 year old boy she was treating. There was no suggestion she had acted deliberately to harm the child but her care was not considered good enough. It was another eight years before the courts also condemned the senior health manager, chief pharmacist and hospital authority. In another case, a doctor in Montpellier committed suicide after being peremptorily removed from work by a director whose own management was already cause for concern. In a familiar pattern, neither the French Haute Authority de Santé (in charge of quality and security of care) nor professional organisations of peers stood out against these events as they should have.
What makes the Bawa-Garba case different is the unprecedented professional outcry. This might prove to be a turning point; from vindictiveness to an open culture of learning, even one that recognizes the well-intentioned erring clinician as a “second victim” (1) of medical error. However, the GMC continues to march out of step. It has promised a period of “reflection” in an attempt at damage limitation, but also, absurdly, proposed that doctors should make formal records each time they feel under-resourced or unsupported – which is much of the time in many health systems.
The often quoted airline industry serves as an inspiring beacon, but is not a simple transferrable blueprint.(2,3) Other work, such as Reason’s model of organizational failures and Charles Vincent’s ALARM method, did not originate from aviation.(4) Safety culture is paramount in scheduled flying. Aircraft do not take off without a full crew on board, a completed engineering checklist and a full complement of ground and air traffic controllers, and this lesson transfers relatively well to elective procedures. However, it cannot be applied in uncontrolled situations: busy, overcrowded clinics, or hospitals which are routinely under-resourced and understaffed, and where morbidity/mortality and emergencies are everyday occurrences. In these conditions, clinicians have a much harder job juggling risks, yet they are held personally liable in a way pilots and air traffic controllers are not. Sokol, a medical ethicist and barrister, claimed that a late change of surgeon may invalidate a patient’s consent(5); but why should this be an issue when a change of pilot is not? Yet again, without clear justification, individual doctors and nurses are held accountable in a way that is unknown in other more successfully improving safety-critical areas. The airline comparison takes us only so far. Better intellectual, ethical and legal reasoning needs to be applied to this difficult situation.
1 Wu AW. Medical error: the second victim: the doctor who makes the mistake needs help too. BMJ2000;320:726.
2 Ladher N, Godlee F. Criminalising doctors. BMJ 2018;360:k479.
3 Amalberti R, Auroy Y, Berwick D, Barach P. Five system barriers to achieving ultrasafe health care.
Ann Intern Med 2005;142:756-64.
4 Vincent C, Taylor-Adams S, Chapman EJ et al. How to investigate and analyse clinical incidents:
clinical risk unit and association of litigation and risk management protocol. BMJ 2000;320:777-81.
5 Sokol D. Who will operate on you? BMJ 2016;355:i5447.
Competing interests: No competing interests
I emailed the GMC asking how I could formally express my lack of confidence in them and did not receive the courtesy of a reply
Competing interests: No competing interests
Instructions to authors wisely warn "Please anonymise the patient’s details as much as possible, eg, specific ages, ethnicity, occupations".(http://casereports.bmj.com/site/about/guidelines.xhtml) This is also ICMJE 's policy: "Nonessential identifying details should be omitted."(http://www.icmje.org/recommendations/browse/roles-and-responsibilities/p...)
However, an editorial subtitles with the name and the surname of a child with Down’s Syndrome.(1)
Republication (also referred to as “replacement”) must be considered.(http://www.icmje.org/recommendations/browse/roles-and-responsibilities/p...) For Elsevier such cases are among "Criteria for emergency takedown."(https://www.elsevier.com/editors/perk/corrections-to-the-record)
1 Ladher N, Godlee F. What must we learn from XXXX's death? BMJ 2018;360:k479
Competing interests: No competing interests
NHS England estimates there are approximately 37,000 deaths each year from sepsis. This means that during the seven-year period 2011-2017, there were somewhere around 259,000 deaths from sepsis in England. Only one of these deaths, that of Jack Adcock in Leicester in 2011, has resulted in the conviction of health professionals for manslaughter (Dr Hadiza Bawa-Garba and Nurse Isabel Amaro). Sepsis can be difficult to diagnose and there are often delays and omissions in its diagnosis and treatment that contribute to the high death rate from the condition. Indeed, even the former Chair of the General Medical Council, Professor Sir Graham Catto, has admitted failing to diagnose sepsis in a timely manner, an error that contributed to a patient’s death.
Because of the problems in the diagnosis and treatment of sepsis, there have been numerous initiatives to improve its management in both primary care and hospital settings. Details of one of the most recent of these initiatives were published by NHS England in September 2017. Given the scale of mortality from sepsis and the many delays and errors so often seen in its management, why then were Dr Bawa-Garba and Ms Amaro convicted of gross negligence manslaughter? Was their management of Jack Adcock so far out of line from the management of other cases of sepsis that resulted in death that they were justly convicted? Or were they involved in just one of many cases where sub-optimal management of sepsis contributed to death? NICE guidance NG51 and Quality Standards QS161 have only recently set out the expectations of what best practice in sepsis care should look like – several years after the convictions of Dr Bawa-Garba and Ms Amaro.
We need an objective review of sepsis deaths to identify the contribution of sub-optimal management to the death and identify lessons for the future in a non-judgmental manner, and not the prosecution of health professionals, if we are to improve clinical outcomes for patients with sepsis.
Professor Azeem Majeed, Professor of Primary Care, Imperial College London, London W6 8RP
Dr Paul Morgan, Consultant Intensivist and Sepsis Lead, Cardiff and Vale University Health Board, University Hospital of Wales, Cardiff, CF14 4XW
1. NHS England. Improving outcomes for patients with sepsis A cross system action plan. December 2015. https://www.england.nhs.uk/wp-content/uploads/2015/08/Sepsis-Action-Plan...
2. NHS National Patient Safety Agency. Medical Error. August 2005. http://www.nrls.npsa.nhs.uk/resources/?EntryId45=61579
3. NHS England. Sepsis guidance implementation advice for adults. September 2017. https://www.england.nhs.uk/publication/sepsis-guidance-implementation-ad...
4. Sepsis: recognition, diagnosis and early management. NICE guidance NG51. National Institute for Health and Care Excellence, July 2016. https://www.nice.org.uk/guidance/ng51
5. Sepsis. Quality Standard 161. National Institute for Health and Care Excellence, September 2017. https://www.nice.org.uk/guidance/qs161
Competing interests: PM is a Lead Volunteer (Wales) for the UK Sepsis Trust. AM declares no conflict of interest.
Criminalising doctors BMJ 2018: 360:k479
Are skin colour, facial appearance and gender the elephant in the room? An informed personal enquiry.
There are two further considerations that are important to me and relevant to Dr. Bawa-Garba’s particular situation that are not being discussed in the BMA report or in this week’s BMJ magazine, regarding gender and colour. As a practised meditator and someone skilled in reflection in regard to the nature of my own thoughts and actions, I have some understanding that can be usefully shared.
Here are two questions, the first easier to answer and the second maybe can never be answered fully:
1. How many junior doctors in the UK are both female and of darker skin colour? I think that they are a small minority group.
2. Is it coincidence that Dr. Bava-Garba, one of this minority, was targeted as scapegoat?
It is useful to ask this second question, although it may never have an adequate answer. Here are my reflections:
A. On gender:
I have a gender bias.
Social studies do indicate that the opinions of males tend to be valued more highly than opinions voiced by females, and men are listened to more carefully. I have looked at this in myself and found it to be true, to the extent that I have a tendency to undervalue my own thoughts and contributions.
I counter this by being aware of this tendency in myself all the time, and resisting it. This is becoming second nature.
B. On skin colour:
I am biased regarding skin colour.
I recommend everyone who is ‘white’ to read this book: ‘Being White in the helping professions’ by Judy Ryde, which outlines her insightful and extensive research (pub. Jessica Kingsley, 2009 and digitally 2011).
To counter my own bias requires a deliberate act of will. When a person of darker skin colour walks into my consulting room, I mentally say ‘hello’ to my own bias. Then as the person sits down in front of me, I make a deliberate point of acknowledging and seeing them as the person they are.
Competing interests: No competing interests
Your editorial raises a range of worrying features about this case of which perhaps the most important is the perceived scapegoating of an individual doctor for deficiencies in health care that are primarily systemic. But a major additional worry is the chasm of apparent disagreement between doctors on the one hand and the GMC and the courts on the other about behaviour that should be judged to be ‘fundamentally incompatible with being a doctor’. According to the judge who acceded to the GMC’s request that Dr Bawa-Garba should be permanently struck off the medical register, the temporary medical incompetence that led to her patient’s death did demonstrate such behaviour and this was why the court overturned the Medical Practitioners Tribunal Service (MPTS) sentence of one year’s suspension from the medical register and replaced it with permanent erasure.
As the judge explained (http://www.bailii.org/ew/cases/EWHC/Admin/2018/76.html), there are three components whereby the GMC’s overarching duty of protecting the public is achieved. The first is by protecting the health, safety and well-being of the public; the second is by maintaining public confidence in the medical profession; and the third is by maintaining proper professional standards and conduct. The GMC wanted Dr Bawa-Garba struck off on all three grounds but the court clearly disagreed that striking her off was necessary to achieve the first of those objectives, protecting the health safety and well-being of the public. As your editorial points out, the court stated unequivocally that ‘Dr. Bawa-Garba, before and after the tragic events, was a competent, above average doctor’ (para 51 of the judgment). Competent above average doctors do not constitute threats to the health safety and well-being of the public.
But the court did agree with the GMC that the second and third components of its protective duties, maintaining public confidence and maintaining professional standards, did require erasure rather than suspension. Why? Because according to the judgment Dr Bawa-Garba’s behavior during her twelve hour shift was ‘fundamentally incompatible with being a doctor’ - a view apparently shared by the judge who originally handed out her suspended sentence of two years in prison, telling her that her medical career was over; and a view clearly shared by the GMC which asked for Dr Bawa-Garba’s erasure. I imagine that the vast majority of doctors would find it simply incomprehensible that a judge should state both that ‘Dr. Bawa-Garba, before and after the tragic events, was a competent, above average doctor’ and that her temporary medical incompetence during that fatal and tragic twelve hour double shift made her permanently unfit to be a doctor. Yet that is precisely what the court did rule, on the grounds that her ‘truly exceptionally bad failings, causing very serious harm to a patient’ and leading to a verdict of gross negligence manslaughter and a suspended two year prison sentence should be regarded as behaviour ‘fundamentally incompatible with being a doctor’. In my own (strictly personal) view the medical profession should collectively combine to support Dr Bawa-Garba in trying to have this judgment overturned (quite independently of the initial gross negligence manslaughter judgment) thus enabling her medical career to be restored to her, as implicitly recommended by the Tribunal.
Whether or not this happens, given the widespread medical disquiet it has provoked, the GMC should do some serious soul searching about its judgment, supported by the High Court, that the reversible temporary though very serious and fatal medical failings demonstrated by Dr Bawa-Gaba during that tragic twelve hour shift constitute behaviour that is ‘fundamentally incompatible with being a doctor’ and which therefore requires permanent cessation of a doctor’s medical career. The GMC’s recent email to all doctors concerning this case states that it appealed against the Tribunal’s sentence of a year’s suspension on the grounds, upheld by the court, that the Tribunal unlawfully ‘failed to respect the criminal court’s judgment’. Close reading of the judgment does not reveal precisely in what ways the tribunal’s sentence of one year’s suspension from medical practice followed by further review by the tribunal failed to respect the criminal court’s judgment. The sections of the MPTS sanctions guidance https://www.mpts-uk.org/DC4198_Sanctions_Guidance___July_2016.pdf_699722... cited by the judge to support the court’s verdict are explicitly only advisory, being preceded by the words ‘103 Any of the following factors being present may indicate erasure is appropriate (this list is not exhaustive)’ (my emphasis). The two sub sections in section 103 cited as underpinning the judgment state: ‘a A particularly serious departure from the principles set out in Good medical practice where the behaviour is fundamentally incompatible with being a doctor’ and ‘c Doing serious harm to others (patients or otherwise), either deliberately or through incompetence and particularly where there is a continuing risk to patients (see further guidance below at paragraphs 123–126 regarding failure to provide an acceptable level of treatment or care)’ (again my emphasis).
Clearly some sorts of behaviour by a doctor are sufficiently appalling to be ‘fundamentally incompatible with being a doctor’ - murder and rape are obvious examples but so too are a variety of other convictable offences that would clearly undermine public confidence in the medical profession and its standards if the doctor were allowed to continue to practise medicine. The guidance document of the MTPS gives a ‘not exhaustive’ list of such offences. But I suspect few doctors would accept that Dr Bawa-Garba’s failings during that tragically fatal twelve hour shift fitted the definition in section 103a of behaviour that was ‘fundamentally incompatible with being a doctor’ thus rendering her permanently unfit for medical practice. And the thrust of section 103c is surely the clause ‘particularly where there is a continuing risk to patients’ - which the court had already clearly rejected in declaring Dr Bawa-Garba ‘a competent above average doctor’. The medical outrage that this case has provoked demonstrates that - unlike the GMC and the courts dealing with this case - large numbers of doctors would have agreed with the Tribunal’s assessment that trust in the medical profession and its professional standards of "a fully informed and reasonable member of the public" would not be undermined if Dr Bawa-Garba had been suspended rather than struck off. On the contrary it seems clear that - in the absence of additional heinous behaviour - if a doctor’s medical incompetence is remediable then informed and reasonable members of the public would think it was in the public interest for it to be remedied, rather than permanently depriving the public of the services of a doctor.
The GMC has told doctors in its email that it is now working closely with the Royal Colleges, BMA and other organisations to address the concerns expressed by the profession. Let me propose that one way forward would be for the GMC, when it reviews the cases referred to it, including all criminal convictions of doctors, to separate out cases of medical incompetence from the rest. If a case of medical incompetence does not involve any additional heinous behaviour that is ‘incompatible with being a doctor’ - and a conviction should not in itself be necessarily taken to be proof of such behaviour (as even the court made clear) - then the GMC’s objective should be first to protect patients by temporarily withdrawing the doctor’s right to practise, and then to require and facilitate remedial training with the aim of restoring the doctor to medical practice as soon as he or she is once again competent to undertake it. That approach would achieve the GMC’s three protective functions noted by the judge. But in the absence of additional heinous behaviour, erasure from the medical register for medical incompetence - even very serious incompetence - should be a last resort used only if the medical incompetence proves irremediable. That was clearly not the case for Dr Bawa-Garba.
Emeritus Professor of Medical Ethics
Imperial College London
Competing interests: No competing interests
The paediatrician was on her first day back after maternity leave without induction and working 2 doctors jobs without adequate supervision or breaks outside of the European working time directive. She was convicted of manslaughter by a non medical jury who are unlikely to understand the medical aspects of the case fully or understand the impossible and dangerous working conditions she was faced with. The death was caused by the hospital and its safety systems, not simply the doctor on duty. The failings of the doctor were only a part of the picture of an overstretched hospital and NHS. This is in no way an excuse or acceptable but there is a total lack of political responsibility for allowing a department to get this short staffed. Criminalising individual doctors will discourage staff from entering and staying in the NHS, and this will impact and make pressures worse. Doctors continuously pick up the slack in failing rotas when colleagues are off. The manslaughter charge should have been given permission for appeal on the grounds that this was the wrong person to blame. The GMC should look back at the case itself, not the charge, and critically examine from those original facts to determine if the doctor should really be erased from the medical register. Choosing to erase a doctor’s registration in order to prevent undermining the public’s confidence in the profession, may have inadvertently undermined the entire medical profession/NHS itself.
Competing interests: No competing interests