Article Text

Download PDFPDF

Op GRITROCK ethics; the way of things to come?
  1. Katherine A Clay1,
  2. J D Henning2 and
  3. S Horne3
  1. 1Royal Centre for Defence Medicine, Birmingham, UK
  2. 2Academic Department of Anaesthesia, Pain and Critical Care, Royal Centre for Defence Medicine, Birmingham, UK
  3. 3Academic Department of Military Emergency Medicine, Birmingham, UK
  1. Correspondence to Maj Katherine A Clay, Royal Centre for Defence Medicine, ICT Building, Birmingham Research Park, Vincent Drive, Birmingham B152SQ, UK; kate.clay901{at}gmail.com

Abstract

The Defence Medical Services (DMS) deployed on Op GRITROCK to Sierra Leone in support of the Ebola outbreak. This operation was the first large-scale DMS deployment since operations in Afghanistan ceased at the end of 2014. This type of operation revealed a number of ethical issues and challenges that the DMS had not faced for a long time. The ethical issues identified during the deployment are discussed in this article using the principalism approach of Beauchamp and Childress. Many of these issues were not identified prior to deployment, and troops were not adequately prepared. The article will outline the difficulties of using the principles of autonomy, beneficence, non-maleficence and justice during a public health emergency, and conclude with recommendations for ethical considerations that should be identified and mitigated against for future deployments to a public health emergency.

  • INFECTIOUS DISEASES
  • MEDICAL ETHICS
  • Ebola Virus Disease

Statistics from Altmetric.com

Introduction

It is now widely recognised that the Defence Medical Services (DMS) encounter unique ethical challenges on deployment. High-profile medical involvement in cases such as Abu Graib1 and Baha Musa2 have meant that the development of doctrine has been rapid, and research into ethical principles and perceptions is delivering useful direction.

Classical ‘obvious’ ethical dilemmas are well rehearsed during pre-deployment training within the UK DMS, and a framework for ethical decision-making on an individual patient basis is well described.3 Standard clinical dilemmas such as when to withdraw care on the basis of futility are well understood. Less traditional medical roles, such as the handling of Captured Persons and their medical care, has also been effectively codified. Medical Rules of Eligibility (MRoE) have long been established to assist in the control of access to scarce resources during a primary military mission—when the overriding objectives are to care for service personnel while complying with the International Law of Armed Conflict.

Operation GRITROCK raised a range of new ethical issues—mostly, but not exclusively, common to all of the Ebola Treatment Units (ETUs) in Sierra Leone. Some of these issues are unique to the management of a highly infectious disease, where the interests of the population carry much higher weight than usual compared with the rights of the individual. Others are more related to operating in a novel emergency environment alongside humanitarian organisations and may be more likely to be encountered by the DMS in the future.

This paper aims to highlight some of the most important of these novel challenges and the ethical arguments behind them. It also demonstrates how some of the largest stumbling blocks encountered in terms of media, publicity and relations with patients and other agencies can easily be predicted by an ethical examination of the situation. This shows how proactive ethical consideration, such as during the early planning phase of the operation and regularly thereafter, may be protective of the reputation of the DMS and facilitate wider civilian interactions.

The issues will be framed using the traditional biomedical ethical approach of principalism advocated by Beauchamp and Childress,4 using four key ethical principles: autonomy, beneficence, non-maleficence and justice. Case vignettes taken from patients treated in our unit will be used to give context.

Respect for individual autonomy

It is widely accepted that ‘individuals must be respected as independent moral agents with the ‘right’ to choose how to live their own lives’.4 Normally this means that a patient has considerable say in the treatment they receive. However, in an infectious disease outbreak, the duty to protect other people becomes increasingly important. When the pathogen is a class 4 virus with a case fatality rate (CFR) of approximately 50%, historically the main means of controlling outbreaks has been to cohort those who may be affected, isolating them from the wider community.5 The WHO advised that ‘good outbreak control relies on applying a package of interventions, including case management, surveillance and contact tracing …. safe burials and social mobilisation’.6 All of these may be in conflict with a patient's own wishes.

Quarantine and isolation are key pillars in the battle against Ebola. Quarantine is where at-risk patients without any symptoms are taken from the community and cohorted until they either are confirmed with disease or not. Isolation is the placement of patients with symptoms consistent with Ebola or confirmed disease into an Ebola holding or treatment unit. Quarantine of infected patients needs to be enforced and is not always a voluntary action by the individual. If a patient is detained in the treatment centre, this action may well undermine the individual's autonomy. Training prior to deployment at the Army Medical Services Training Centre provided some preparation for this eventuality; with scenarios provided to challenge the clinician into using their negotiating skills rather than force to ensure the patient remained in the treatment unit and complied with treatment. If a confused or combative patient was admitted into the DMS ETU, staff had the option of padlocking the entrance and exit gates to prevent them leaving—this being more in common with UK legal precedent for the management of patients without capacity. However, there was no means (or plan) to hold a competent patient against their will. This principled respect for autonomy even in the face of the huge risk to the public of non-compliance contrasted somewhat with other forms of legalised coercion within Sierra Leone at the time; safe burial was compulsory, and restrictions on movement were periodically implemented in certain areas7 alongside mandatory temperature screening. Most DMS clinicians have some understanding of medical ethics, but Public Health Ethics are perhaps more poorly understood. Public Health Ethicists understand that there may be a need to restrict individual choice in efforts to prevent disease and promote health—and indeed this may not solely be in the case of an epidemic; the banning of smoking in public places, the addition of fluoride to water and folic acid to flour are all good examples.

The privacy of a patient is impossible to protect during an Ebola outbreak. The patient is removed from their community and isolated for up to many weeks. Contact tracing is undertaken in their community including sexual contacts. If the patient survives and returns to their household, they will be known as an ebola virus disease (EVD) survivor, which can be stigmatising, leading to the patient being unwanted at home or in work. This is in direct contrast to usual UK practice with regard to confidentiality and the patient's right to control how information about their health is used and shared.

Case 1: A 49-year-old Sierra Leonean man with confirmed EVD with diarrhoea and renal failure. Shortly after arriving in the treatment unit he became confused and combative. He did not speak English, and an interpreter was not available. He pulled out two peripheral intravenous cannulas and was making attempts to walk around the unit; he was unsteady on his feet, which concerned the team on duty. He refused attempts to reinsert the cannula or to place a catheter or bowel management system (BMS). The clinical team decided to sedate the patient with haloperidol and diazepam in his best interests to prevent him from harming himself. This also had the effect of preventing him from harming other patients, which was thought to be a possibility while the staff were not in the unit. It also allowed a line to be reinserted and fluids started. The patient was cared for on a mattress placed on the floor.

Case 1 discussion: Respect for autonomy incorporates allowing patients their liberty, privacy and gaining their consent to any proposed intervention. In this case, the patient did not speak English and he was admitted during a night shift when interpreters were not present. Despite attempts to speak to the patient using basic Krio, we could not converse adequately with him so an accurate assessment of his confusional state could not be made. Invasive procedures were carried out without his consent as it was thought to be in the patient's best interests. Had we not been inside the ETU, greater attempts could have been made to find an interpreter or involve relatives in order to reduce the patient's confusion and distress. Involving relatives was not always possible because many of our patients were transferred from other ETUs from around the provinces and it was difficult for relatives to travel these distances either because they were in quarantine themselves or due to financial constraints.

Beneficence

The principle of beneficence requires us ‘to strive to do good where possible’. The public health interventions of contact tracing and isolating at-risk individuals or enforcing safe burial practice may be justified by this principle as a means of protecting the health of the population.

The WHO-published average CFR for EVD is 50%,8 which is a population estimate across three countries and the whole epidemic. The CFR for the DMS ETU was 44% (in 36 patients). The DMS provided a higher level of supportive care with central venous access, monitoring of fluid balance, basic bloods including electrolytes and a wider spectrum of antimicrobials available compared with other ETUs. Other units have published their mortality rates; the Chinese Military report a CFR of 73.13%9 compared with published CFR of 31%10 by the Republic of Sierra Leone Armed Forces. All units provide different protocolised care, but it is difficult to compare their outcomes because the number of patients treated and the case selection for each unit differs. In addition, as the outbreak progressed the disease severity appeared to lessen and CFRs from different time points during the outbreaks may not be comparable. Medicins Sans Frontieres (MSF) reports that in their experience from previous outbreaks ‘good clinical care can reduce overall case fatality rates by between 10 and 15 percent’.11 Available evidence suggests that it is in the best interests of the patient to be treated at a facility with good supportive care, but it is not proven that central venous access provides a better patient outcome than peripheral venous access.

Case 2: A 39-year-old man admitted with a 10-day history of feeling unwell. On admission he was found to be in renal failure and dehydrated. A Foley urinary catheter was inserted along with a central venous catheter (CVC) into the right internal jugular vein. The area around the line continued to ooze for many hours after insertion, which was thought to be secondary to Ebola-induced coagulopathy. He was given four units of fresh frozen plasma, 1 g of tranexamic acid and adrenaline 1:10 000 was infiltrated around the insertion site, which led to control of the bleeding.

Case 2 discussion: The ethical issues raised around CVC insertion within the ETU include consent; it was often difficult to fully consent patients for the procedure due to language barriers, although where possible interpreters were present. Also, CVCs are not widely used in Sierra Leone so it is difficult for the patients to understand the procedure they are being asked to consent for. Other ETUs were using peripheral venous access with comparable mortality rates, and it has not been proven that CVCs improve patient outcome as a study has not compared the two interventions in the same unit with the same patient subpopulation. The benefit of the CVC strategy was a reduction in the high-risk procedure of peripheral cannula insertion or venipuncture. The disadvantage was an increase in bleeding from a central line insertion point compared with that of a peripheral cannula. This left the healthcare worker (HCW) caring for this patient's line site at greater exposure to infected blood.

The DMS ETU placed central venous access into all EVD patients admitted into their unit when in the gastrointestinal phase of their illness. This was in order to allow regular blood sampling to monitor renal function, electrolytes, liver function, clotting and haematology. It would then be possible to correct any deficiencies by providing blood products, clotting factors, fluids and potassium using this line. The central line, combined with a BMS, enabled an accurate fluid balance to be recorded and meant that patients with extreme fatigue and profuse diarrhoea were able to stay in bed, clean and comfortable. The use of central lines and BMSs provided a dual benefit. In using a BMS, the number of nursing visits to perform basic care including washing was markedly reduced and arguably the idea of a rectal device was not very acceptable to some patients. In drawing blood from central lines, a needleless system could be used, reducing the risks to the staff of needlestick injury. This had the associated benefit of less punctures for the patient. In the first 23 patients, >150 peripheral venepunctures, with their associated risk of needlestick injury, were avoided due to the CVC strategy employed at the DMS unit.12 As central lines were placed in many patients early in their illness while they were still relatively well, for some HCWs there was a perception that the effect of the procedure was more to benefit staff rather than the patient; this sometimes led to ethical disquiet.

Non-maleficence

Adhering to the principle of non-maleficence requires that ‘one should avoid doing harm to others’.4 Adherence to this principle of non-maleficence within ETUs across West Africa was undermined by two practices: the use of experimental therapies and the process of cohorting patients suspected of having Ebola (one of the reasons for establishing ETUs).

Experimental therapies

The unprecedented scale of the Ebola outbreak in West Africa meant that many experimental therapies were being used in human trials in the absence of alternative therapies. Professor Solomon, director of the National Institute for Health Research health protection research unit in emerging zoonotic infections, said that ‘it is unusual for drugs to be tried in healthy volunteers first, in case of side-effects. The difference here is the desire for this ‘first in man’ experiment to be for a patient with the disease’.13

Data published in 2014 suggested ion channel blockers such as amiodarone inhibit filoviral cell entry.14 This led Emergency, an Italian non-governmental organisation (NGO) that had been in Sierra Leone since 2001, to trial using amiodarone inside their ETU.15 This experiment had only been carried out in cell lines and its experimental use led to mixed opinions in the team working inside the ETU. UK Med volunteers stopped working in the unit following their concern over the ‘off-license administration of this potentially toxic drug, given to suspect as well as confirmed Ebola patients without their consent’.16 Emergency provided a rebuttal to these comments on their website, asserting ‘it was slanderous to state that Emergency does not ask for patients’ consent … On the contrary, we wonder why no one is asking the African patients their ‘consent’ to be treated in a very different way (‘a second class way’, one could say) compared to the ‘western patients’’.17 In this circumstance, experimental therapy was being given to Sierra Leonean patients for a brief time but it was not clear whether this treatment led to patient harm.

At a meeting in September 2014 reviewing potential Ebola therapies and vaccine, the WHO agreed that:In the absence of proven treatments, participants widely agreed that convalescent whole blood and convalescent plasma, among other experimental interventions, should be considered for use for people with Ebola virus disease. Any such use should be scientifically studied though carefully designed research studies. If, however, convalescent whole blood and convalescent plasma are used for treating people with Ebola virus disease outside research studies, this use should be considered ‘monitored emergency use of unregistered and experimental interventions’, a term coined by WHO in the context of the current Ebola outbreak to refer to an exceptional use of experimental interventions outside clinical trials, and to reflect the urgent need to collect data on their efficacy and safety.18

The DMS treatment centre used convalescent whole blood for consenting HCWs admitted from Connaught Hospital. This was the only group the DMS treated that had an agreed protocol for convalescent blood or plasma.

Case 3: A 21-year-old nursing student presented with a 3-day history of fever, malaise, headache and general lethargy. He was confirmed to be EVD-positive in the unit and a rapid diagnostic test (RDT) was positive for Plasmodium falciparum malaria for which he received oral artemether and lumefantrine. Over the next week, the patient deteriorated, entering the gastrointestinal phase of the illness with worsening renal failure. At this point with informed consent, the patient received two units of convalescent whole blood provided by the referring hospital. A further RDT was positive for P. falciparum malaria, and he received a further course of therapy. This patient had a protracted illness and was discharged from the treatment centre on day 26 to receive ongoing rehabilitation.

Case 3 discussion: There are risks associated with administrating a whole blood transfusion. Known risks include the early complications such as haemolytic reactions, febrile reactions, transfusion-related acute lung injury and reactions secondary to bacterial contamination. The late complications include transmission of infection; for example, viruses (hepatitis A, B, C, HIV, CMV), bacterial (syphilis, salmonella) and parasites (malaria, toxoplasmosis).19 These transmitted infections may be easily treatable as in the case of malaria, but donor blood was not screened for parasitic or bacterial infections. Blood received at our unit had previously been screened for bloodborne viruses but the donor may have HIV or hepatitis and be in the ‘window period’ where diagnostic tests would be negative. A literature review from sub-Saharan Africa20 identified that it is difficult to screen for malaria in blood donors as malaria tests are not sensitive enough to detect low-level parasites. Also, blood is in short supply and therefore donors are not likely to be rejected in countries with high endemicity due to limited supplies. The potential benefits of convalescent blood are not known. With no data to support this unproven therapy, on balance the potential harm may outweigh the benefits.

Cohorting suspected cases

In addition to the issues around autonomy that are raised by cohorting suspected cases, there is also a real risk of harm. At the height of the epidemic, in some areas up to 90% of suspected cases eventually tested positive. What is ethically challenging is that 10% did not. During the time when 90% were likely to be positive, this minority 10% by being quarantined alongside the 90% were at increased risk of contracting the disease compared with the background risk they would have faced at home. Under these circumstances, the ethical ‘weight’ given to this risk of harm was outweighed by the benefits to the population of controlling the outbreak. However, in our ETU, the rate of positivity was probably closer to 50% at the peak, and in all areas had dropped below 10% by the end of the epidemic—a natural consequence of the prevalence of EVD dropping while the rate of other conditions that similarly could meet the case definition (eg, malaria) remaining the same. In early 2015, MSF raised concerns that some individuals had actually contracted the EVD while in ETUs, and this drove a national move to using only dedicated treatment units with a higher degree of separation in the suspected areas. Due to the high level of resourcing for our bed-base compared with other ETUs, the Kerry Town site was actually spared this dilemma; our suspected patients were kept completely isolated with separate rooms and bathrooms, and so the risk of cross-contamination was low. However, there were still opportunities for patients to encounter each other in passing or for contamination to persist in an area and so the risk will always have been higher than zero. This understanding meant that infection control and disinfection between patients was absolutely paramount at all times to keep this risk as low as possible. Other ETUs had to adapt their structure to manage this changing risk profile.21 This is no small undertaking in the midst of an epidemic, but is clearly the correct course of action should the epidemiological profile mean that the risk of harm increases. This highlights that ethical issues may vary considerably during the same deployment, and what was right at the beginning may be unethical towards the end.

Justice

The principle of justice states that ‘people should be treated fairly, although this does not necessarily equate with treating everyone equally’.4 The military treatment centre in Kerry Town was initially a 12-bed facility with the capability of providing a high level of supportive care. The facility's primary mandate was to accept local HCWs from UK-funded treatment facilities and international HCWs. Numbers of cases meeting these criteria were low—a typical outcome of the use of MRoE. These are designed to control access to the facility so that there would always be capacity available to ensure completion of our primary mission and as such are an essential tool in mission planning and execution. However, in the context of an increasing epidemic, the military were not working close to full bed capacity. The usual means to adapt to changes in circumstances on the ground is to have a discretionary component to the MRoE, and this was also the case on GRITROCK. Unfortunately, this discretionary component brings its own difficulties; if the discretion becomes the expected norm, then there is no longer the same control over access to the facility. Many agencies were unaware that there was flexibility in the MRoE.

The MRoE caused unease and friction in several ways. It was a source of intense frustration among the Sierra Leonean doctors who mostly worked in government hospitals that they were not automatically eligible for admission if they were to be infected. This perceived disparity in care for local HCWs contributed to junior doctor strikes in the country in December.21 Significant negative publicity was also drawn to the unit22 by comparison of our MRoE with other nations (although this was based upon incomplete information and was incorrect). Another issue was the definition of HCW—the MRoE-stated doctor or registered nurse, paramedic or laboratory worker. The MRoE had not taken into account that in Sierra Leone much of the Ebola control effort was supported by non-registered HCWs; ambulance drivers, hygienists, student nurses and community health workers. Unsurprisingly, most agencies did not see why their Sierra Leonean doctors should be eligible but not their ambulance workers. Again, the discretionary part of the MRoE allowed these personnel, in the main, to be accepted. It was subsequently made explicit that Sierra Leonean HCWs would be accepted if bed capacity allowed. In reality, most local HCWs referred to the facility were accepted; unfortunately, the damage to the UK military reputation and the loss of goodwill accumulated the local government and other non-governmental agencies took a while to repair.

The MRoE may also have been part of an observed issue for staff of the ETU. While not explicitly about justice, the exact nature of the mission had a significant impact on how personnel felt about their role. Among those military clinical staff who felt that they had deployed to help care for Ebola patients, there was a constant sense of unease that the facility had spare capacity and yet also provided a higher level of care than most patients would get elsewhere. For this group, the decision to use the discretionary part of the MRoE would likely have been perceived very positively. Personnel who felt that they were deployed to bolster the resolve of international volunteers to come and work in Sierra Leone were much more comfortable that the facility was not running at capacity as places were guaranteed to be available for such cases should they become infected. They may have perceived the use of the discretionary parts of the MRoE as threatening to their operational effectiveness. Still others seem to have perceived discretionary cases as unreasonably increasing their own personal risk of contracting the disease because the ‘additional’ discretionary admissions undermined the risk assessment calculations made for the mission.

During the first half of the outbreak, experimental therapies such as ZMAPP, a collection of monoclonal antibodies directed at the Ebola virus, were only available to international HCWs.13 These HCWs were also repatriated to their home country for medical care. With this higher level of medical care and the availability of experimental medication, they had a much greater chance of survival compared with their Sierra Leonean counterparts. A different standard of care was provided to Sierra Leonean patients, which will have increased the mortality in this group and therefore harm was not avoided. It also attracted significant media condemnation as an example of inequality or failure to observe the principle of justice.23

Phua comments that with regard to the principle of justice ‘there should be fairness in the distribution of costs and benefits with respect to individuals as a result of public health interventions’.24 New Africa Analysis25 question how the Department for International Development (DFID) can justify the amount of money spent on the Kerry Town site, which includes the 100-bed Save the Children facility. According to the article, Kerry Town cost £53 million, and the medical supplies and equipment were provided on a separate budget. The author comments that with a total spend of approximately £422 million by DFID there should be an improved medical infrastructure as part of a lasting legacy. A total of £53 million could have built two large government hospitals equal in scale to the Connaught Hospital in Freetown. Instead now that the tents have been taken down at Kerry Town, all that remains are the concrete foundations. More money will be spent in Sierra Leone to help rebuild and improve the existing medical system, but it could be argued that the money spent on the Ebola outbreak could have been spent in a different way to leave a lasting legacy.

Expanding ethics for the DMS

It is apparent from the historical cases of ethical breaches that the reputation of a medical service rests on how ethically it behaves. It is also clear from the above discussion that many of the unique ethical issues around Ebola are relatively predictable, even using a simple autonomy, beneficence, non-maleficence and justice model. What perhaps is less immediately clear is how many of the issues with morale and reputation that impacted on Operation GRITROCK might have been identified and mitigated against before the deployment. It is likely that a robust ethical review undertaken as part of the planning process might pre-empt and mitigate against similar issues in the future. As the British Army's capstone doctrine on operations points out, ‘the moral component requires care and effort to be invested in understanding and protecting a number of standards’.

While many of the ethical issues faced by staff were predictable using the well-taught medical ethical approach outlined above, in many cases they were not easily ‘solvable’ using the quadrant approach normally advocated by the British Military.26 This is because there were at least two other ethical codes at play in this operation, neither of which have been well addressed in the last few years, where the main effort has been combat operations.

First, there are the ethics and principles of humanitarian aid, which indicate a real role for the military, but only as a last resort, to ‘bridge the gap’ until full NGO support is mobilised.27 A lack of understanding of the principles of humanitarianism has been repeatedly cited as a cause of friction between the military and NGOs when they need to operate in close proximity. Although the Oslo guidelines adhere to the last resort approach, they are also pragmatic and recognise that concurrent deployments are likely in future; training and dialogue with NGOs is essential to improve the effectiveness of the DMS in such situations.27

The other code that needs to be better understood is the public health ethic; in such widespread public health disasters, the principalism of Beauchamp and Childress must be replaced by a far more consequentialist code, focused on the end state of preventing the disease. This puts a much higher emphasis on the public collective good and often requires government action to assert the health of the public over individual autonomy.

Recommendations: ethical risks that should be identified and mitigated against prior to deployment

Where staff are deployed into circumstances where respect for autonomy may be difficult to achieve—for example, where coercive practices may be necessary or where patients are likely to lack capacity—these should be fully exercised and discussed during training. The command team understanding the situation and being able to give direction is not enough; staff implementing decisions also need to understand why they are acting in a way that is contrary to the norms they are used to.

Where personnel may feel that there is a double effect from an action (eg, better for them but also safer for us), these should be acknowledged and talked about; allowing personnel to understand why we are acting in ways that may not intuitively seem purely in the patient's best interest.

Where there are risks that the actions of staff may cause harm to patients, these must be explored in depth. An approach to novel therapies or research needs to be determined as early in the deployment as possible, and education of staff should take place so that they understand the risks we are asking our patients to take, and the rationale for them.

The MRoE will always be a source of some contention. Making sure that the MRoE are in the public domain in their entirety where possible would prevent misunderstanding and misrepresentation. Alongside that, a strategy to address the inevitable inequalities that are generated by limiting access would allow proactive media lines; for example, to underscore how the MRoE actually enhance our contribution to the international effort.

Interpretation of the mission influences how individuals deployed weigh the various ethical components, and so absolute clarity about our role will improve morale by reducing friction between groups with different perspectives.

The ethical review needs to be repeated throughout the operation so that changing circumstances and their impact on what we are doing are fully understood. In particular, in the context of infectious diseases the impact of the prevalence on the way we manage the outbreak needs to be re-evaluated regularly.

Conclusions

In summary, an infectious diseases outbreak will always lead to tensions between practices that ensure the best medical care for an individual and those that limit the spread of the disease and protect the health of the population. This is particularly marked when the disease has such high infectivity and mortality rates as seen in Ebola. An in-depth ethical assessment and planning will allow us to be better prepared for some of these challenges in the future. The military deployment to Sierra Leone exposed new public health ethical considerations that the DMS had not faced in previous primarily military campaigns. Preparing for and educating the clinical staff on the complex ethical dilemmas likely to be faced in future public health emergencies will have a significant, positive effect on the morale of clinical staff caring for patients, on patient experience and on the reputation and standing of the DMS.

Acknowledgments

The authors would like to acknowledge the help of Dr Heather Draper in the writing of this article.

References

Footnotes

  • Contributors KAC: planning and writing the article. JDH: review of content. SH: Planning and writing the article, senior review. KAC and SH: responsible for the overall content.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.