Share this article on:

Ethical Considerations for Ventricular Assist Device Support: A 10-Point Model

Petrucci, Ralph J.*; Benish, Lynne A.†; Carrow, Barbara L.‡; Prato, Lisa§; Hankins, Shelley R.¶; Eisen, Howard J.¶; Entwistle, John W.∥

doi: 10.1097/MAT.0b013e3182223269
Adult Circulatory Support

The potential for long-term support on a ventricular assist device (VAD) in the bridge-to-transplant (BTT) and destination therapy (DT) settings has created unprecedented ethical challenges for patients and caregivers. Concerns include the patient's adaptation to life on a device and the ethical, clinical, and practical issues associated with living on mechanical support. On the basis of our experience treating 175 consecutive VAD patients, we have developed a model to address the ethical and psychosocial needs of patients undergoing VAD implantation. Patient preparation for VAD implantation encompasses three phases: 1) initial information regarding the physical events involved in implantation, risks and benefits of current device technology, and the use of VAD as a rescue device; 2) preimplant preparation including completion of advance directives specific to BTT/DT, competency determination, and identifying a patient spokesperson, multidisciplinary consultants, and cultural preferences regarding device withdrawal; and 3) VAD-specific end-of-life issues including plans for device replacement and palliative care with hospice or device withdrawal. This three-phase 10-point model addresses the ethical and psychosocial issues that should be discussed with patients undergoing VAD support.

From the *Department of Psychiatry, Drexel University College of Medicine/Hahnemann University Hospital; †Temple University Hospital, Philadelphia, Pennsylvania; ‡Thoratec Corporation, Pleasanton, California; §The Cardiac Center, Children's Hospital of Philadelphia; ¶Division of Cardiology, Department of Medicine, and ∥Department of Cardiothoracic Surgery, Drexel University College of Medicine/Hahnemann University Hospital, Philadelphia, Pennsylvania.

Submitted for consideration November 2010; accepted for publication in revised form April 2011.

Disclosure: Ralph J. Petrucci was a neuropsychological consultant for Thoratec Corporation during this investigation. Barbara L. Carrow is employed by Thoratec Corporation.

Reprint Requests: Ralph J. Petrucci, EdD, Department of Psychiatry, Drexel University College of Medicine, 245 N. 15th Street, Philadelphia, PA 19102. Email: Ralph.Petrucci@drexelmed.edu.

Ventricular assist devices (VADs) were initially used as a bridge to transplant (BTT), largely owing to the limitations of the technology. Rates of device infection, device failure, and stroke were high with the early devices, and the pneumatic devices were cumbersome. As the technology improved, 2-year survival rates of 24% were obtained in a group of VAD patients ineligible for transplantation who received electrical device implants. This opened the door for these devices to be used as destination therapy (DT) in transplant-ineligible patients, an option usually considered in patients who were too old for transplantation. Despite approval for this indication in the United States and Europe,1,2 utilization rates were low, primarily because survival beyond 2 years was poor. Currently, miniaturized pumps that provide continuous flow have demonstrated significant improvements in durability and long-term patient outcomes and, thus, are being implanted in large numbers for both the BTT and DT indications.3–5

Patients on a VAD face issues related to their unique medical characteristics, the length of time they may be supported on the device, and the increased potential for unexpected outcomes with this technology. These issues are creating challenges related to patient selection for device implantation, management of device-related complications, and withdrawal of care.6–8 As VAD centers accumulate more patients on long-term support,9,10 the potential increases for ethical and psychosocial issues to arise. Although predicting all the potential areas of conflict may be overwhelming, proper preimplant preparation may ameliorate many of these issues before they arise. Although many ethical standards have been developed for use of other lifesaving technologies and treatment of serious medical conditions such as stroke and cancer, these standards are not necessarily applicable to VAD patients. Improvements in this technology have led to heightened expectations from VAD patients and their families and a reassessment of terms such as “futile care.” The improved durability of VADs has created a situation in which “end of life” may require the active withdrawal of device support rather than “letting nature take its course.” A comprehensive set of best practice guidelines for VAD therapy could address many of the ethical concerns in treating these patients.

We have developed a framework for the discussion of potential ethical and psychosocial issues based on experience attained over nearly 20 years with VAD patients at a single institution. A total of 175 VADs were implanted during this period (72% male, age range: 16–75 years, 61% ischemic), with 64 undergoing emergent implantation either for cardiogenic shock or after cardiotomy. Outcomes in this subgroup were poor, with only five patients eventually receiving transplants and five recovering cardiac function. The remaining 54 expired within 45 days from infection, hemorrhage, or multiorgan failure, with almost all deaths preceded by withdrawal of support. In this group, 11 developed conflicts because family members could not agree on when mechanical support should be withdrawn or on who should make the final decisions, because of their feelings of culpability. Few of these patients had healthcare directives (HCDs). Only three families of the nonsurvivors produced an HCD and only one patient who underwent transplantation completed an HCD later. Of the remaining 111 patients who presented for urgent or elective VAD implantation, 22 patients (20%) identified a healthcare spokesperson (16 spouses, five siblings, and one parent) after developing advance directives, living wills, and/or power of attorney (AD/LW/POA) documents. Of eight patients who received implants for DT (four currently alive), only one has AD/LW/POA. The remaining three have identified a spokesperson but have not formally documented that decision.

Surveying the variability within the VAD-dependent group leads one to recognize the need for a systematic process to guide patients and families through both urgent and elective implants. We have developed the guidelines presented herein on the basis of our broad range of experiences with this diverse group of VAD patients and have modified them as new issues have arisen.

Back to Top | Article Outline

Guidelines

Our 10-point model for preparing patients for VAD implantation encompasses three phases, including the initial information (informed consent), preimplant preparation (advance directives and competency determination), and discussion of VAD-specific end-of-life preferences (Table 1). Each area usually requires initial explanation to the patient and family and repetition over time to ensure that they most effectively retain the information provided. The timing of the information presented will differ with the clinical status of the patient as there may not be enough time before implantation to discuss these issues with critically ill patients.

Back to Top | Article Outline

The Initial Information Phase

The initial information phase begins with a series of meetings between the family and the surgeon, cardiologist, or nurse coordinator. These discussions may be the patient's first exposure to the concepts and realities of VAD implantation and management. These exchanges provide information for the beginning of the consent process. This contact is critical because most patients are unfamiliar with VADs and many have unrealistic preconceived notions regarding VAD support. The accurate communication of information in this phase may help prevent problems later.

The initial information phase encompasses the first four of the 10 areas in our model. The physical events leading up to this type of surgical intervention should be explained, including the physiological developments that have occurred in the patient to justify consideration for mechanical device implantation and the implications of elective versus emergent implantation. The patient should be informed about the device technology currently available, including the risks, benefits, and possible outcomes of implantation. The possibility of device failure and need for device replacement must be addressed. Specific devices available at the institution should be explained, along with the implications of physiology and body size on device selection, the potential for VAD complications, and the possible need for additional surgical procedures.

The patient should also be advised that the VAD is a rescue device and not a “cure” for heart failure and that VAD implantation has relatively high mortality and morbidity rates compared with conventional cardiac surgery. Available data from trials using the device to be implanted can be presented.10–13 The potential outcomes of ongoing medical treatment versus the institution of mechanical support can be compared, using evidence-based medicine where possible. Models that predict outcomes in congestive heart failure, such as the Seattle Heart Failure Model, can be presented.14 The patient and family should be informed about the differences between VAD use as a bridge to recovery, decision, or transplantation versus VAD as a destination device, especially if the classification of the patient is subject to change at a future time. Caregivers should clearly explain to the patient his/her indication and what steps, if any, he or she must take to become or remain a candidate for heart transplantation. Patients should also be given realistic expectations for recovery from the operation supported by the center's survival data, rate of neurologic events, patients' average length of hospital stay, and the percent of BTT patients undergoing transplantation in a given timeframe. Idealistic or unrealistic expectations of the patient or family should be addressed prospectively to minimize the incidence of dissatisfaction later.

This phase should also address timing of the surgery, expected recovery time, and a rough timeframe for transplantation in the case of a BTT patient. The clinician should note that care plans can change at any time based on the patient's physical, hemodynamic, social, or neurologic needs. The patient should be advised that the plan may be modified based on variables such as age, cancer, pregnancy, or the time necessary to demonstrate modification of undesirable behavior, such as a 6-month trial and establishment of a contract for abstinence from substance abuse.

Back to Top | Article Outline

The Preimplant Preparation Phase

Preimplant preparation includes discussing the psychological and social aspects of life with a VAD and encompasses the next four points in our 10-point model. A physician, nurse coordinator, psychiatrist, psychologist, social worker, or other member of the VAD team who has the appropriate level of training and knowledge regarding VAD support should direct this discussion. Building on the initial informed consent, patients should prepare advance directives, a living will, and/or a healthcare power of attorney before device implantation to avoid issues that may arise after implantation if significant complications occur, such as serious neurologic injury. Advance directives are more important for VAD patients than for most other heart failure patients because most patients and families do not inherently understand the issues that may arise on VAD support. For instance, they may not realize that the blood pressure may be supported for an extraordinarily long time even in an otherwise nonsurvivable situation. Advance directives should be discussed specifically in relationship to VAD placement and to the patient's status as BTT or DT before implant. Possible conflicts among potential surrogate decision makers should be clarified at this time. Resolving such conflicts may help protect the patient and family and the physicians and the institution from contention later. The major risks of the procedure, such as changes in care or outcome related to infection, bleeding, or neurologic events, and the potential impact of changing a patient's designation from BTT to DT should be discussed when completing advance care planning.

Preimplant preparation also requires determining the patient's competency or capacity to understand the information provided and the consequences of his or her decision. The patient and family also need to understand that the determination of capacity may change after implantation has occurred. Neurologic or psychiatric evaluations may be necessary at this point, and the institution's risk management group may become involved when complications or significant concerns arise. Developing advance care planning may help to reduce conflicts. The patient should appoint a spokesperson or surrogate, so that communications between the medical staff and the family and/or friends do not become fragmented. The patient should be assured that caregivers will provide frequent updates to his or her spokesperson or health advocate.

Before device implantation, the team should discuss with the patient and family options that may be considered if the patient has a significant untoward event around the time of implantation, including need for machine withdrawal, emergent pump replacement, or do not resuscitate/do not intubate orders. If the clinical situation justifies it, additional heroic measures such as extracorporeal membrane oxygenation may be discussed. The patient should be involved in discussions of the eventual possibility of removing mechanical life support. The patient's cultural preferences for device withdrawal (presence of family or clergy, music or reading of religious scriptures, etc.) can be addressed. The patient's preferences should be communicated to the family or spokesperson if the patient permits it. The hospital and VAD team should make plans to accommodate the patient's social, religious, and cultural preferences if possible. Finally, the patient and family should be advised that the institution's ethics committee may be consulted if conflicts arise between family expectations and the patient's progress that cannot be resolved through the usual means.

Back to Top | Article Outline

The VAD-Specific End-of-Life Phase

VAD-specific end-of-life preferences should be discussed with the patient and family if they have not been addressed during the preimplant phase. This phase encompasses the final two points of our 10-point model and includes decisions about VAD withdrawal, hospice, palliative care, and other end-of-life scenarios that may occur after implantation, such as developing end-stage cancer, dementia, device failure, or another critical medical event. This discussion is potentially the most important because nearly all patients undergoing VAD implantation are facing imminent mortality and are being offered a treatment option with which they are not familiar to decrease this risk. Despite VAD implantation, end-of-life issues may arise if a catastrophic complication occurs related to the operation or device or if the device begins to fail.12 This discussion should entail the patient's preferences for or against aggressive management (if it is an option) and palliative or comfort care. This last phase may include more time for making decisions and considering more information because it is outside the realm of surgical complications, and these events may develop over time rather than acutely. In addition, the patient will have adjusted to life on a VAD and may have a different outlook than he or she had before implantation. The VAD team should try to elicit the patient's desires regarding remaining in hospital, going home, or receiving palliative care in a hospice facility. An acceptable device withdrawal process that meets the needs of the patient and family should be formulated according to the earlier discussion of the patient's religious or cultural beliefs. Ongoing sharing of information throughout the process of VAD consideration helps to further facilitate a palliative care discussion.13

Back to Top | Article Outline

Case Scenarios

The importance of the concepts covered by this protocol can be seen by examining the real-life issues faced by patients undergoing VAD implantation and their medical providers. We present four patient scenarios highlighting the influence of unusual psychosocial variables that may disrupt and delay care in VAD patients. Accompanying each case is a list of the points in our protocol that could be used to prevent or ameliorate the issue. These examples clearly demonstrate that not all possible scenarios can be anticipated but that planning may help mitigate the ensuing turmoil.

A divorced 50-year-old man presented for transplant evaluation nonemergently and exhibited abusive verbal behaviors with staff, borderline medical compliance, and suboptimal family support. A VAD was chosen to give him time to demonstrate compliance after inotropic management failed. Before implantation, his sister was identified as having POA. Within 6 months after DT implant, the sister withdrew from her supporting role, leaving the man's 17-year-old daughter as the primary support and decision maker. Over the next year, the patient underwent a transplant with the support of his then 18-year-old daughter. Knowledge of family dynamics allowed a surrogate decision maker to be named, and the patient was able to receive an organ transplant. He has subsequently done well (see points 3–6 in Table 1).

A single 42-year-old man presenting for urgent consideration for transplantation had poor social support. He had been homeless in the recent past and was now living with a cousin. He had been hospitalized multiple times because of heart failure, and assessing his compliance with medical advice was impossible owing to geographic distance and the need for frequent hospitalizations. Because he was estranged from his adult daughter, a brother and a cousin were identified as his support. Advance directives were completed, and he underwent implantation as DT. Soon after implant, family members became minimally involved in his care. The patient died from multisystem organ failure after a prolonged hospitalization in which no decision maker gave any significant input. The presence of advance directives, however, gave the team the necessary information to care for the patient according to his wishes (see points 2–7, 9, and 10 in Table 1).

An 80-year-old man with end-stage heart failure presented with progressive deterioration despite inotropic support. The options of DT and hospice were presented. The team raised concerns about his support because his wife was elderly and their adult children lived far away. Preimplant conversations included discussions on developing a backup plan in case the wife was unable to assume full caregiver responsibilities. After the patient underwent implantation, his hospitalization and time in rehabilitation were more prolonged than the family had expected, leading to conflict and anxiety. As the patient began to require additional help with daily care, his children were able to make alternate care plans to avert a major problem (see points 3, 5, 6, and 8 in Table 1).

A 50-year-old woman with an elderly husband presented for a nonemergent transplant evaluation. Her husband was identified as having POA. An adopted 22-year-old daughter was identified who could provide additional support. A VAD was implanted as BTT, but during the patient's immediate postoperative course, a debilitating anoxic neurologic event occurred, necessitating a long medical and rehabilitation course. After the patient was discharged to home, her husband developed dementia, became incompetent, and moved out of state to be near his family for further care. The adopted daughter withdrew involvement from the patient, leaving her homeless and with no support. The patient then became a ward of the state and required placement in an institution. In this case, the dissolution of family support was completely unexpected and was worsened by the inability of the patient to live independently (see points 4–6, 8, and 9 in Table 1).

Back to Top | Article Outline

Discussion

Because only 18% of our patient population had AD/LW/POA in place when they were hospitalized, families or other support persons were often left without specific knowledge of the patient's wishes. Further, none of the patients with AD/LW/POA had considered the complex issues associated with VAD support at the time they signed the documents, so these papers were necessarily vague regarding the specifics of VAD support and the subsequent decisions that are sometimes required. With a priori direction lacking from these patients, it is timely and necessary to add VAD-specific psychosocial and ethical considerations to discussion of the medical and surgical issues. The timing of these discussions must be modified based on the clinical situation.

For nonemergent procedures, nurse coordinators or mental health providers can initiate discussions on informed consent, advance directives, and end-of-life issues during the preoperative period. These team members should also emphasize the need to establish a spokesperson and to clearly identify that person to other members of the family and medical team. With emergent procedures, it is often the surgeon or cardiologist who first seeks to elicit the family's wishes when the patient cannot make decisions or name a spokesperson. To address these situations, we have proposed a patient-centered process to prepare patients and family members for VAD implantation. This process will help reduce the potential for ambiguity and misunderstanding of the meaning of artificial support and withdrawal. This process can be helpful in both the elective and emergent settings, although the timeline for addressing the issues may need to be individualized. The aim is to avoid conflict between the patient and/or family and medical caregivers by clearly discussing the aims, expectations, and realities of VAD support in advance. Similarly, this helps families avoid making difficult decisions hastily during a time of stress by addressing these issues under controlled and planned circumstances. The process is ever changing and dynamic, and the model is not a pro forma solution for the delivery of service but rather a guide to the areas that should be addressed.

VAD-specific discussions are important as patients undergoing VAD implantation are different from other heart failure and cardiac surgical patients in numerous ways, as are the problems they may face. Because the medical and the psychosocial needs of VAD patients create a unique mechanical variant of the cardiac disease spectrum, establishing a comprehensive framework for discussion will improve communication among the medical care providers and the patient and will help avoid future conflicts. Life on a VAD produces uncertainty for the patient and family. Care providers, including those involved with palliative planning, must be able to address the special needs of VAD patients to help mitigate this. Thorough preimplant discussions regarding these issues are part of this process.

This 10-point patient-centered model ensures that patients and their families receive all the information necessary to be prepared for a life-altering therapy and that critical areas of discussion are not unintentionally overlooked during the implantation process. The timing of the discussions will vary greatly with the urgency of the implantation. The medical team should ensure that the patient and family receive a clear, cohesive message without contradictory information. Because members of the VAD team are familiar with the pre- and postsurgical needs of implant patients, team members should lead these discussions. Doubts and fears result from the prospect of an implant, and patients find it difficult to adapt to their new medical condition and the changes in their care plan that may result. Dedicated VAD healthcare professionals can help minimize patients' and families' confusion through effective communication. “Helpful” input from other care providers should be minimized to decrease the patient's and family's confusion. The timing of the discussions regarding VAD support must be flexible because of the nature of the disease process. Ideally, when the need for VAD implantation can be anticipated, the initial information and preimplant preparation phases should begin early enough for the patient to absorb and process the information before the surgery. For many patients, this can take several days or longer. Addressing VAD-specific end-of-life issues such as device withdrawal before implantation may be difficult, but discussing advance directives preoperatively is imperative. Often, the VAD-specific end-of-life discussions can be delayed until the patient is doing well and has been discharged to home. At this point, the patient can address serious issues such as catastrophic device failure or other major complications without also being distracted by fear of the impending surgery.

In the older DT population enjoying an extension of life, further life-changing events such as retirement, spousal death, neurologic events, or dementia require care modifications and team flexibility. Addressing these potential issues preoperatively may help avoid later confusion and conflicts that produce a medical limbo in which the patient becomes a burden with an end-point known as “destination to nowhere,”15 where the patient is dependent on a device and has no real chance of a meaningful life. In some cases, a patient may decline implantation because of the tenuous nature of his/her support.

Often, new issues arising from these discussions can be addressed directly by members of the VAD team. Visual or graphic descriptions of the device can be helpful. After a patient's candidacy for a VAD has been ascertained, direct communication between the prospective VAD patient and a present or former VAD patient can be instructive and gratifying to the patient and family. These ambassadors may be extremely helpful in countering the unrealistic fears that a patient may have regarding life on mechanical support. However, care must be taken to avoid presenting a prospective VAD patient with a view that may be unrealistic because of age, comorbidities, or other problems. As the numbers of VAD sites and patients receiving implants grow, an organized and systematic approach to informing the patient and family about VAD expectations is advised.16 Clearly, although the sequence of our 10-point model is not appropriate for every VAD implant patient, the various points need to be addressed at some time in all VAD patients. In an emergency, saving the life of the patient must take precedence as long as the patient or family has the minimum information to make an informed decision. However, even during emergent implants, the concept should be presented that the VAD is a rescue device because many family members do not understand the gravity of these situations and may have unrealistic expectations. The team can address remaining issues once the critical situation has been resolved. This will help to increase the comfort level of the patient and family for life on the device and may help avert problems or misunderstandings in the future. Repeating the preoperative discussions with the patient and family at a later date may be useful because retention and understanding of complex medical information are poor under emergency circumstances. Such repetition may seem unnecessary at the time but may help to avoid confusion and conflict in the future. Because of the complexities involved with VAD implantation and care, we recommend that VAD teams repeatedly encourage patients and their families to sign AD/LW/POAs and provide the documents to facilitate this process. However, despite encouragement, adoption of AD/LW/POA remains low in this patient population. We hope that at minimum, patients will have a conversation with their loved ones regarding their wishes that the family can rely on if a difficult decision must be made. As this process is evolving, we have recently developed a written set of guidelines for our VAD patients including a written request for them to develop advance directives. We anticipate that such changes will further improve the process and decrease the number of recurrent issues that arise, although unanticipated issues may still occur.

Back to Top | Article Outline

Conclusions

VAD implantation, whether elective or emergent, has the potential to create unprecedented medical and psychosocial dilemmas. Ethical policies addressing these dilemmas have lagged behind the technology. Our three-phase 10-point process can assist VAD implant centers in developing a comprehensive plan for addressing problems unique to this therapy. The model is designed to assist in addressing issues that should be discussed before and after the initiation of VAD support.

The structure of this patient-centered protocol is likely to reduce confusion and potential ethical issues, solidify expectations, and allow for a more transparent implantation process. The increasing numbers of patients with end-stage heart failure, and the shortage of available organs, necessitate developing an organized, ethically sensitive approach toward VAD implantation.

Back to Top | Article Outline

Acknowledgment

The authors thank James N. Kirkpatrick, MD, Assistant Professor of Medicine, Hospital of the University of Pennsylvania and Associate Fellow, Center for Bioethics, University of Pennsylvania, for his thoughtful comments and direction.

Back to Top | Article Outline

References

1. Medicare National Coverage Determinations Manual, Chapter 1, Part 1 (Section 20.9). Coverage determinations (rev. 129, 12-08-10), 2010. Available at: www.cms.gov/manuals/downloads/ncd103c1_Part1.pdf. Accessed May 17, 2011.
2. Rose EA, Moskowitz AJ, Packer M, et al: The REMATCH Trial: Rationale, design, and end points. Randomized evaluation of mechanical assistance for the treatment of congestive heart failure. Ann Thorac Surg 67: 723–730, 1999.
3. Lietz K, Miller L: Will left-ventricular assist device therapy replace heart transplantation in the foreseeable future? Curr Opin Cardiol 20: 132–137, 2005.
4. Stevenson LW, Kormos RL: Mechanical cardiac support: Current applications and future trial design. J Am Coll Cardiol 37: 340–370, 2001.
5. McCarthy PM, Lamm RD, Sade RM: Medical ethics collides with public policy: LVAD for a patient with leukemia. Ann Thorac Surg 80: 793–798, 2005.
6. Dudzinski DM: Ethical guidelines for destination therapy. Ann Thorac Surg 81: 1185–1188, 2006.
7. Felker GM, Rogers JC: Same bridge, new destinations: Rethinking paradigms for mechanical cardiac support in heart failure. J Am Coll Cardiol 47: 930–932, 2006.
8. Kirkpatrick JN, Fedson SE, Verdino R: Ethical dilemmas in device treatment for advanced heart failure. Curr Opin Support Palliat Care 1: 267–273, 2007.
9. Mueller PS, Swetz KM, Freeman ML, et al: Ethical analysis of withdrawing ventricular assist device support. Mayo Clinic Proc 85: 791–797, 2010.
10. Dembitsky WP, Tector AJ, Park A, et al: Left ventricular assist device performance with long-term circulatory support: Lessons from the REMATCH trial. Ann Thorac Surg 78: 2123–2130, 2004.
11. Rogers JG, Butler J, Lansman SL, et al: Chronic mechanical circulatory support for inotrope-dependent heart failure patients who are not transplant candidates. Results of the INTrEPID trial. J Am Coll Cardiol 50: 741–747, 2007.
12. Miller LW, Pagani FD, Russell SD, et al: Use of continuous flow device in patients awaiting heart transplantation. N Engl J Med 357: 885–896, 2007.
13. Pagani FD, Miller LW, Russell SD, et al: Extended mechanical circulatory support with a continuous-flow rotary left ventricular assist device. J Am Coll Cardiol 54: 322–324, 2009.
14. Levy WC, Mozaffarian D, Linker DT, et al: The Seattle Heart Failure Model: Prediction of survival in heart failure. Circulation 113: 1424–1433, 2006.
15. Bramstedt K: Destination nowhere: A potential dilemma with ventricular assist devices. ASAIO J 54: 1–2, 2008.
16. Rizzieri AC, Verheijde JL, Rady MY, McGregor JL: Ethical challenges with left ventricular assist device as a destination therapy. Philos Ethics Humanit Med 3: 20–35, 2008.
Copyright © 2011 by the American Society for Artificial Internal Organs