Search tips
Search criteria 


Logo of jrsocmedLink to Publisher's site
J R Soc Med. 2001 September; 94(9): 479–482.
PMCID: PMC1282193

End-of-life issues in AIDS: the research perspective

Patricia A Grady, PhD FAAN, Ann R Knebel, DNSc FAAN, and Anita Draper, MA

In 1999 alone, HIV/AIDS-associated illnesses caused the deaths of about 2.8 million people worldwide, including an estimated 480 000 younger than 15 years1. Some 13.2 million children have lost their mothers or both parents since the epidemic began1. In this paper we describe research initiatives being pursued at the National Institute of Nursing Research, but first we examine existing publications to provide some background information.


In the USA, AIDS-related deaths have been declining since 19962. While these trends are encouraging, they lead to a false impression that, once people have access to advanced antiretroviral therapies, end-of-life issues cease to be of great concern. This is not so. Moreover, large disparities in AIDS deaths between different ethnic groups3 (Figure 1), coupled with knowledge of the importance of cultural competence in approaches to ethnic minority groups, indicate that interventions should be targeted more to the demographics of the epidemic. The most common approach to end-of-life care in the AIDS population has been to apply the interventions (hospice, palliative care) used in cancer patients. We lack information on the best ways to provide care in the unique circumstances of the AIDS epidemic.

Figure 1
AIDS-related deaths in the USA in 1998 according to ethnic background (from Ref. 3)

The hospice philosophy rests on the assumption that dying is not only a natural part of life but also an event that can and should be as meaningful as life's other important events. Thus, hospice focuses on assisting patients in having a peaceful death, in which they are free of pain, are aware of what is happening to them and around them, and have as much control of unfolding events as they want.

With AIDS and many other life-threatening illnesses, medical breakthroughs are expanding the timeframe for palliative care and blurring the definition of end of life. The availability of advanced diagnostic techniques and treatments can sway the care of terminally ill AIDS patients in the direction of curative care, so the end-of-life experience is determined by the available technologies. Palliative care in the United States has yet to achieve wide acceptance within the American medical community, where treatment success is customarily measured in terms of lives saved or, in effect, deaths postponed. An approach that eases a patient's path towards an imminent death faces great difficulty in being recognized as legitimate, much less honourable. This limited view of palliative care stems partly from its ‘low-tech’ nature. Many people mistakenly believe that palliative care comes into play only after medicine has exhausted all the skills and technologies that might bring about a cure. Palliative care and medical research and technology are not mutually exclusive. Far from turning its back on the latest findings and state-of-the-art technology, good-quality palliative care draws on them in the determined pursuit of its own goal—a peaceful death. Yet, under certain circumstances, patients may receive radiotherapy, chemotherapy and surgery as part of their palliative care4.

Delivery of palliative care

The history and current drawbacks of hospice care in the USA have been reviewed by Cassel and Demel (p.433). Coverage under the Medicare hospice benefit includes visits by the hospice team, durable medical equipment, and supplies. Medicare also provides coverage for bereavement counselling for the family and caregivers for up to one year after death of the patient. Medicaid, the federal programme of health coverage for indigent Americans, includes a similar benefit for hospice services.

Data on the use of the Medicare and Medicaid hospice benefits by persons with AIDS are difficult to find. The AIDS diagnosis was not mentioned in a study of Medicare beneficiaries' costs and use of care in the last year of life5. One of the authors of that study suggests that the AIDS diagnosis is invisible in the database because deaths are usually coded according to immediate cause, such as pneumonia or malignant disease (Lunney J, personal communication).

For a patient to become eligible for Medicare coverage of hospice services, a physician must certify that the person is terminally ill and can be expected to live only for another six months or less. Obviously, this determination is not always easy to make with accuracy, and Congress has amended the original Medicare benefit to allow for extensions.

Many people who are eligible for Medicare and Medicaid hospice benefits still do not receive them or receive them only a few days before death. The usual explanation is that physicians tend to be overly optimistic about their patients' prognoses or are reluctant to dash their patients' hopes by delivering the six-months-or-less prognosis necessary for benefit eligibility. Of particular relevance to the delivery of palliative care to patients with AIDS is a series of studies showing that Hispanics and African Americans do not enter hospice programmes in significant numbers in the United States. In a study in New York, African Americans expressed a reluctance to sign advance directives and do-not-resuscitate (DNR) orders for fear that their care would be abandoned6.

The reluctance of patients to sign advance directives or agree to DNR orders may be due in part to the difficult nature of end-of-life conversations. In a study conducted in the north-west United States, persons with AIDS from non-white ethnic groups were less likely than non-Hispanic whites to report that they communicated with their primary care provider about end-of-life preferences7. People with low incomes, injection drug users and women with high-risk sex partners tended to report poor communication. In contrast, patients who had completed a living will or durable power of attorney for healthcare reported a better quality of communication. Satisfaction with care was strongly associated with quality of communication. The authors recommend that, since socioeconomic status and ethnicity are associated with both the occurrence and the quality of end-of-life communication, future interventions to improve communications about end-of-life care should consider the effect of these variables.

In a follow-up analysis, the same group assessed the reasons why patients and physicians do not talk about end-of-life care (Table 1)8. Non-white patients, injection drug users and women were more likely to feel that talking about death would bring it closer. The barriers and facilitators to end-of-life communication between persons with AIDS and their care-providers suggest that effective palliative care will demand end-of-life communications that focus on the needs of individuals as well as specific groups.

Table 1
Barriers and facilitators to communication (from Ref. 8)


In the UK, Guthrie and colleagues evaluated whether or not people with HIV in London were able to die where they wanted9. The investigators posited that, for patients, dying where they planned implies that they are aware of their prognosis, that their symptoms are controlled, and that they are involved in decision-making during the late stages of the disease. Information was extracted from existing databases at St Stephen's Clinic in Central London for the years 1982-1993. Early in the epidemic almost all people with HIV died in the hospital, but in the late 1980s there was an increase in deaths at home and in hospices. By the end of the study period, almost 60% of people were dying where they had planned—but many still chose to die in the hospital. Guthrie et al. conclude that hospitals must continue to provide for the needs of patients dying with HIV/AIDS.

Box 1 Domains of high-quality end-of-life care (from Ref. 10)

  • Adequate pain and symptom management
  • Avoiding inappropriate prolongation of dying
  • Achieving a sense of control
  • Relieving burden
  • Strengthening relationships with loved ones

Just what the needs of dying patients are was evaluated in a qualitative study of patients' perspectives of the elements that comprise good end-of-life care10. Three patient groups—dialysis patients, people with HIV infection and residents of long-term care facilities—identified domains that define high-quality end-of-life care (Box 1). These domains were similar to models derived from experts such as the American Geriatrics Society and the Institute of Medicine of the National Academy of Sciences in the US; however, the investigators highlight important differences between the patient-derived and the expert-derived domains. The patient-derived domains were simpler, more specific, not bound by established concepts such as ‘global quality of life’ and focused on outcomes.

Investigators in the UK also retrospectively examined the experience of a generic hospice providing care to patients with advanced HIV disease11. Only 26 patients were admitted during a 4-year period (0.85% of new admissions), supporting Guthrie's conclusion that hospitals will continue to be a major site of palliative care delivery. Referrals to hospice were primarily from hospitals (70%) rather than practitioners. The median length of stay for all patients referred to hospice was only 19 days—suggesting that, as in the US, referrals to hospice come late in the illness trajectory. Some of the symptoms experienced by the patients, such as weakness, immobility, and weight loss, were not amenable to interventions; others, such as pain, dyspnoea, and confusion, were. Services provided to patients reflected the expertise of the full multidisciplinary team. The authors conclude that generic hospices can offer important services to HIV patients who are at the end of life and that professionals need to consider referring these patients to such hospices.

In a prospective study by the Regional Infectious Disease Unit in Scotland, health service and hospice use by HIV-infected individuals was evaluated for 1992-9312. Most of the 513 patients (72%) had been infected through injection drug use. As expected, patients with lower CD4 counts had more and longer hospitalizations. Resource use in this Scottish centre was remarkably similar to that in other British centres, despite the differences in patient populations (risk groups); population variation was less important in determining resource use than was clinical or immunological stage. Hospice use was twice as high among people who were relatively well (pre-AIDS and CD4 count ≥ 200/μL)—a finding attributed by the authors to increased respite care, the effect of risk group (injection drug users) and socioeconomic status. Seemingly, hospice care can be useful early in the AIDS illness trajectory.

In summary, the research base is limited but we can identify some areas upon which to build future research. At present, good-quality end-of-life care can be defined by five domains identified by patients (see Box 1). The site of delivery of palliative/hospice care (home versus hospital) may be less important than implementation of the philosophy. Generic hospices, which have in the past concentrated on oncology patients, may improve outcomes for individuals dying with AIDS. In addition, hospice care should perhaps not be confined to the time identified as the end of life but should encompass a longer trajectory.


The 1997 report from the Institute of Medicine noted significant gaps in scientific knowledge about end-of-life care and stressed that these gaps ‘require serious attention from biomedical, social science, and health services researchers’13. In this same year, the National Institute of Nursing Research (NINR) co-sponsored, with other Institutes at the National Institutes of Health (NIH), a research workshop on ‘symptoms in terminal illness.’ Afterwards, NINR published a programme announcement, Management of Symptoms at the End of Life, and a request for applications, Research on Care at End of Life, to stimulate research on these issues. The large number of applications for funding led to NINR's designation as the lead institute at the NIH in the drive to promote the science of end-of-life care. Building on these earlier initiatives, a programme announcement, Quality of Life for Individuals at the End of Life, was published by NINR and six other Institutes in August 2000. Applications were invited for studies on end-of-life management of physical and psychological symptoms, communication between patients and providers or patients and families, ethics and clinical decision-making, and caregiver support. The programme announcement specified an interest in research on patients' end-of-life issues across the lifespan, including neonates in intensive-care units, children with AIDS, teenagers with cancer, and young adults with degenerative diseases.

Of some twenty-two research studies and seven training and career development awards now being funded, only a few focus on end-of-life issues in HIV/AIDS. To date we have received few applications on this topic. One study that we co-fund with the National Institute of Mental Health investigates the desire for death among terminally ill patients with AIDS. Another is looking at advance directives in patients with AIDS. We plan to continue building this area of science so that we can provide the best care for patients with HIV/AIDS as well as the data that policy organizations need to make informed decisions.

NINR is pursuing other aspects of end-of-life care and is a major sponsor of a new study by the Institute of Medicine on paediatric end-of-life issues. In addition, an NIH-wide research interest group on end-of-life issues was organized by NINR in October 2000, and one of its first actions was to host an open forum in which the NIH community and the general public could consider future directions for end-of-life research. Four end-of-life experts made presentations on technologies used at the end of life, cultural and ethnic dimensions, palliative care, and ethical issues. The forum was timed to build on momentum generated by a highly publicized documentary series, On Our Own Terms—Moyers on Dying, that was broadcast on public television across the US in September 2000.


For science to inform policy we need to acknowledge the impact of the demographics of the AIDS epidemic on the delivery of end-of-life care and construct intervention studies that are targeted to those individuals. The AIDS illness trajectory needs better explanation so that the definition used to determine hospice payment is not based on arbitrary time-frames, and so that effective palliative care interventions can be applied early in the disease process. We also need to find ways to change the perception held by many healthcare providers that palliative care is less interesting and less challenging than high-technology interventions. By investigating the domains of high-quality end-of-life care we can determine the best approaches to caring for persons with AIDS who are at the end of their lives. Finally, we need to tailor measurement instruments for disparate cultures and agree on intra-national and cross-national measures needed to evaluate end-of-life care. Joined by its NIH partners, NINR is positioned to continue developing these research initiatives to keep pace with emerging understanding of research needs in the area of end-of-life care.


1. UNAIDS. Report on the Global HIV/AIDS Epidemic. UNAIDS, 2000
2. Centers for Disease Control and Prevention. HIV/AIDS Surveillance Rep 1999;11: 1-44
3. Centers for Disease Control and Prevention. Data presented at the 1999 National HIV Prevention Conference, Atlanta, Georgia, Aug 29—Sept 1, 1999
4. Rousseau P. Hospice and palliative care. Dis Mon 1995;41: 779-842 [PubMed]
5. Hogan C, Lynn J, Gabel J, Lunney J, O'Mara A, Wilkinson A. Medicare Beneficiaries' Costs and Use of Care in the Last Year of Life. Report submitted to Medicare Payment Advisory Commission, May 1, 2000
6. Butler RN, Burt R, Foley KM, Morris J, Morrison RS. Palliative medicine: providing care when cure is not possible. Geriatrics 1996;51: 33-44 [PubMed]
7. Curtis JR, Patrick DL, Cardwell E, Greenlee H, Collier AC. The quality of patient—doctor communication about end-of-life care: a study of patients with advanced AIDS and their primary care clinicians. AIDS 1999;13: 1123-31 [PubMed]
8. Curtis JR, Patrick DL, Caldwell ES, Collier AC. Why don't patients and physicians talk about end-of-life care? Arch Intern Med 2000;160: 1690-6 [PubMed]
9. Guthrie B, Nelson M, Gazzard B. Are people with HIV in London able to die where they plan? AIDS Care 1996;8: 709-13 [PubMed]
10. Singer PA, Martin DK, Kelner M. Quality end-of-life care. Patients' perspectives. JAMA 1999;281: 163-8 [PubMed]
11. Gibbs LME, Ellershaw JE, Williams MD. Caring for patients with HIV disease: the experience of a generic hospice. AIDS Care 1997;9: 601-7 [PubMed]
12. Brettle RP, Atkinson FI, Wilcock J, Richardson A, Lewis R, MacCullum L, Bath G. Hospital and hospice resource use of HIV-positive patients in Edinburgh. Int J STD AIDS 1997;8: 234-42 [PubMed]
13. Institute of Medicine. Approaching Death: Improving Care at the End of Life. Washington, DC: IOM, 1997

Articles from Journal of the Royal Society of Medicine are provided here courtesy of Royal Society of Medicine Press