|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Originally published as JCO Early Release 10.1200/JCO.2004.06.111 on July 26 2004 © 2004 American Society of Clinical Oncology. Pattern of Care at the End of Life: Does Age Make a Difference in What Happens to Women With Breast Cancer?From Palliative Care Medicine, and the Cancer Nutrition and Rehabilitation Program, Department of Oncology, and the Department of Epidemiology and Biostatistics, McGill University; Division of Clinical Epidemiology, Royal Victoria Hospital; and McGill University Health Centre; Montreal, Canada Address reprint requests to Bruno Gagnon, MD, MSc, Division of Clinical Epidemiology, Royal Victoria Hospital, 687 Pine Ave W, R.4.29, Montreal, Quebec H3A 1A1, Canada; e-mail: bruno.gagnon{at}clinepi.mcgill.ca
PURPOSE: In the last 40 years, palliative care has become the standard of care at the end of life. However, there are limited data about the degree of access to such care at the population level. METHODS: Using administrative databases, a care-oriented profile score was created to describe the care received during the last 6 months of life for 2,291 women who were dying of breast cancer in the province of Quebec, Canada, during the years 1992 to 1998. The care received was described through indicators of care that would reflect a palliative care philosophy. An ordinal score was developed for comparisons among age groups of women using a proportional odds ordinal regression model. RESULTS: We found that only 6.9% of women died at home, while 69.6% of them died in acute care beds. While most women (75%) had few indicators indicating provision of palliative care during the last 6 months of life, younger women (< 50 years) were even less likely (odds ratio, 0.70; 95% CI, 0.54 to 0.90) to receive such care compared with middle aged women (50 to 59 years; serving as the reference group), while older women (> 70 years) were more likely (odds ratio, 1.85; 95% CI, 1.49 to 2.29). CONCLUSION: Our study indicates that a sizeable proportion of women terminally ill from breast cancer do not have access to palliative care—an issue that health care policy makers may wish to explore further.
The WHO1 Expert Committee defined palliative care as the active total care of patients whose disease is not responsive to curative treatment. Control of pain; of other symptoms; and of psychological, social, and spiritual distress is paramount. The goal of palliative care is the achievement of the best possible quality of life for patients and their families.1 Therefore, palliative care affirms life and regards dying as a normal process; it neither hastens nor postpones death; it offers a support system to help patients live as actively as possible until death; and it offers a support system to help the family cope during the patient's illness and in their own bereavement.1 Clearly, palliative care goes beyond focusing on medical issues. The sole control of pain and other symptoms will fall short of the goals of palliative care. Palliative care programs have clearly improved the quality of life of patients with advanced cancer and their families. Reports on the value of hospices,2 home care programs,3 specialized palliative care teams in hospitals,4 and regional programs5 have shown proven benefits to patients and families. Quality of life, the main outcome of the palliative care philosophy, is significantly improved in patients admitted to palliative care units.6 How much has this philosophy of care influenced the care received by all dying patients in a community? Prospective studies, while optimal for this type of question, are very costly and difficult to carry out. Some researchers have used the collected clinical information available in administrative databases to study these clinical questions successfully.7-11 Such studies are feasible at a reasonable cost and allow the study of the whole population. The provision of palliative care is not directly documented in administrative databases. However, certain elements of care that reflect provision of palliative care and that have been identified as potential indicators of good quality of end-of-life cancer care, can be extracted from administrative databases. One feature of good end-of-life care is the possibility of a patient remaining at home and dying in relative comfort and dignity. In a prospective cohort study, Townsend et al showed that 67% of patients "dying of" cancer, given favorable circumstances, would prefer to die at home.12 Hays et al13 found the same wishes in a study performed in a continuing-care retirement community, in which 62% of patients said they would prefer to die in their residence or nursing facility. Earle et al,14 in a study of potential indicators of good end-of-life cancer care, confirmed that home death is preferred, as opposed to hospital death. They also identified other indicators of poor-quality end-of-life cancer care—short interval between last chemotherapy dose and death; frequent emergency department visits; high number of hospital and intensive care unit days near the end of life; low proportion of patients enrolled in hospice; and others. The study time interval was selected to reflect the cancer evolution before death. In general, cancer patients experience a marked decline in their physical function in their last 6 months of life15, 16 associated with increased symptom severity.17 Even if physicians are unable to predict survival,18, 19 they should be able to notice decline in physical function and reorient the care toward a more palliative approach. Therefore, our chosen time frame is the last 6 months of life. In the case of cancer care, it is imperative to distinguish between the terminally ill from cancer, those whose death was preceded by a clinical state characterized by metastatic cancer and slow deterioration before death (dying of cancer), from those who died from complications of cancer treatment at an early stage (died of cancer treatment), or those who were diagnosed previously with cancer but were in remission and died from other illnesses such as chronic obstructive pulmonary disease or heart failure (died with cancer). In the study of the terminally ill, the knowledge of cause of death is insufficient to classify them according to the disease responsible for the needed care. The recorded cause of death may not reflect the underlying pathology that was responsible for the care received during the last months of life. Ascertaining the exact nature of disease causing the terminal state is a prerequisite for the study of patterns of care at the end of life, as patterns vary with different illnesses.20-22 According to the Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatment (SUPPORT) research group, the ideal population for the study of health services at the end of life must meet specific characteristics.23 The people included should have a disease that is common enough to have a significant impact on the health care system: the mortality rate from this disease must be sufficiently high to justify allocation of financial resources for the care patients in their terminal phase; the population should require important decisions to be made regarding appropriateness of care during the process of care such that there is a possibility of moving the patients toward a more care-oriented environment; and the disease process should have fairly stable treatment possibilities in the foreseeable future so as to avoid recommending structures that will be obsolete before they are implemented. Breast cancer, the second leading cause of cancer death in women,24 meets these criteria. When breast cancer spreads beyond regional lymph nodes, it becomes an irreversible progressive fatal illness.25 Breast cancer also affects women of all ages, from the young to the elderly.24 Furthermore, under a medical approach to care, women, particularly young women, would receive an array of curative treatments and procedures almost until death. Older women would likely be treated less aggressively. Hence, the medical or curative model would tend to overtreat the younger women and to accept more easily the dying process in the older women. It is to be expected that a difference would emerge in the access to palliative care—a difference noticeable through the mentioned indicators, between women of different age groups. The overall aim of this study is to define the extent to which women dying of breast cancer had access to palliative care. The specific objectives are to (1) define indicators of a palliative care–oriented approach extractable from the coded information contained in administrative databases of hospitalization and health service billing; (2) incorporate these indicators in a palliative care–oriented profile score; and (3) estimate the influence of age on the care-oriented profile score in the last 6 months of life.
Data Sources We accessed two main sources of data—a hospital discharge database (Maintenance et Exploitation des Données pour l'Étude de la Clientèle Hospitalière [MedEcho]); and a physician fee-for-service billings databases (Régie de l'Assurance Maladie du Québec [RAMQ]). Because health care in the Province of Quebec is publicly funded for all, these databases cover all health care services provided for the whole population of the Province of Quebec. MedEcho contains coded information of all hospital separations. There are data fields for the Medicare number of the subject (NAM), final diagnoses and 14 secondary diagnoses, topography codes serving as specific identification of cancer diagnosis and sites of metastases, dates of admission and discharge, type of bed (acute care, chronic care, or palliative care) during hospitalization, status at discharge (alive or dead), discharge destination (eg, home, other institution), age, sex, regional indicators, and procedures. Type of hospital (acute care v chronic care) is also coded. The RAMQ databases contain, among other things, the NAM (identical to the NAM in MedEcho); a beneficiary file containing all fee for services received (code of act), site of act (eg, at home, in chronic care institution, emergency department, out patient clinic, acute care institution), specialty of physician performing the act (eg, identification code of the physician, date of act); and a physician file containing some information about the physician (age, sex, and licensing information). The linkage of the two databases is done at the ministry level through the NAM, and the merged file is provided with an identification number that is denominalized to respect confidentiality.
Population
Indicators of the Palliative Care–Oriented Profile Score
Other indicators of access to palliative care during the institutionalized care received by each woman for the last 6 months of life was characterized by the number of admissions (< 2) and the total time spent in acute-care beds (excluding time spent in palliative care or chronic care beds), expressed as the percentage of the 6-month period (< 20%). Contiguous admissions were amalgamated and treated as a single admission, and the admission leading to death was excluded. A high percentage of time spent in the hospital and multiple admissions may suggest failure to redirect the care toward a palliative care–oriented approach or a shortage of palliative care beds for patients unable to stay at home. The number of visits to the emergency department was documented as follows: (1) all care received during consecutive days counted as one visit; (2) an emergency department episode of care resulting in an admission did not constitute a visit. Usually women with adequate home support should not need emergency care if an admission is not needed. The specialties of the physicians assuming care and the place of follow-up are important for the quality of care and can serve as a proxy for hospice enrollment. The number of visits to family physicians, medical oncologists, radio-oncologists, medical internists, and surgeons during the periods when the patient was not hospitalized, was used as an indicator of the type of care received; these visits were determined using the health service billing database. All visits that occurred in physicians' offices; at home; and at Centres Local de Santé Communautaire (CLSCs; public community-based health provider units composed of physicians, nurses, and other health care providers with the responsibility to respond to home care needs of a specific and delimitated area) were classified as community care. Visits to the outpatient departments of hospitals were considered hospital-based visits. We considered that the woman's care was assumed by the family physician instead of by a specialist if one of the following criteria were fulfilled: (1) she saw her family physician as often as the treating specialist; (2) the proportion of visits to the family physician was at least 20% of all the specialist visits together. To define the indicator for home visit by the family physician, one single home visit during the last 6 months of life was sufficient. First-line chemotherapy provides palliation of the symptoms experienced by many women with advanced breast cancer, but the use of further chemotherapy becomes increasingly problematic owing to significant side effects associated with limited effectiveness.27 Administrative databases do not record the type or line of chemotherapy administered to patients. As there is a rapid decline in physical function in their last 6 months of life,15 we assumed, because all these women had advanced breast cancer, that chemotherapy given in the last 6 months of life would reflect a failure to provide a palliative care–oriented approach. A palliative care–oriented profile score was created using place of death and characteristics of care during the last 6 months of life. Place of death was assigned a value according to specific sites: (1) home was given a maximum score (score of 5) as it is recognized as the desired place of death; (2) palliative care units received the second highest score (score of 4); (3) chronic care received a slightly lower score (score of 3); and (4) emergency department and acute care beds were given a score of 0 points. However, women dying in acute care beds could receive an additional score (0 to 3 points) corresponding to the number of care-oriented indicators present during the last admission (Table 1). To this a score was added for the level of care received during last 6 months of life, consisting of one point for each of the seven indicators (scores, 0 to 7; Table 1). A comorbidity score was created for each subject using the Romano-Roos Dartmouth-Manitoba modification of the Charlson Comorbidity Index to include all available diagnostic information. For this study, we covered the 2 years preceding death (breast cancer–related diagnoses were excluded).
Statistical Analysis
The palliative care–oriented profile score is only an ordinal noninterval in nature. For example, the relative value assigned to home death is a ranking value and not a quantity. Therefore, the analysis must be applicable to ranked data. A statistical technique, the proportional odds ordinal regression model,28, 29 allows the calculation of a cumulative proportional odds ratio from an ordinal score. Using the 13 possible palliative care–oriented profile scores (0 to 12), an odds ratio for each of the 12 cut points (these cut points represent each step between a score of 0 or > 0; of 0, 1 and > 1; of 0, 1, 2, and > 2; etc) was calculated, and a cumulative proportional odds ratio was calculated for each age group using the 50- to 59-year-old age group as the reference group. For each age group, the consistency of the odds ratios across cut points was verified by inspection, and a test for homogeneity (
From the initial cohort of 28,100 women, 2,291 were classified as dying of breast cancer. Table 2 presents the number and the percentage of women in each age group having each of the indicators for a care-oriented model. The high proportion of hospital death (69.9%), with a very low proportion of home death (6.9%), needs to be noticed.
The average number of admissions during the last 6 months of life was 1.8 (standard deviation [SD], 1.2), with a median of two and a range from 0 to 10. The average number of emergency visits was 0.7 (SD, 1.1), with a range from 0 to 11. The average number of home visits was 1.0 (SD, 3.3), with a range from 0 to 53; 21.3% of women had at least one home visit, and 10% of women had three or more visits. Figure 1 shows the relative percentages of women cared for by general practitioners or by specialists for specific place of death. The proportion of women dying in acute or palliative care beds under the care of specialists was in keeping with the fact that specialists took care of 77.8% of the women. However, of the women who died at home, more than half were under the care of general practitioners (53.5%). Almost all women who were cared for by a general practitioner had at least one home visit (96.5%), whereas only 18.9% of the women cared for primarily by specialists were visited at home.
In Table 3, the influence of age is similar for each of the three indicators and the palliative care–oriented profile score across all age groups. Younger women are less likely (< 40 years: odds ratio, 0.69 [95% CI, 0.54 to 0.90]; 40 to 49 years: odds ratio, 0.70 [95% CI, 0.56 to 0.88] to receive a palliative care–oriented approach of health care services than the women of the reference group (50 to 59 years). In contrast, women 70 years of age and older were more likely to receive a palliative care–oriented model of health care services (odds ratio, 1.85; 95% CI, 1.49 to 2.29). The effect is the same—each of the three indicators demonstrating that the choice of the cut point of these indicators would not have modified the effect of age on the palliative care–oriented profile score. As adjusting the model for presence of comorbidities did not modify the estimates for each age group, we did not report the results.
Using administrative data, this study was able to describe at the population level, the access to palliative care at end of life and the impact of age on these patterns. One striking finding was the very low rate of home deaths. Death at home was not directly identified, but was assumed if the woman did not die in a hospital, in the emergency department, or in a chronic care institution. Although the result was low, our method, if erroneous, would have overestimated the number of home deaths. A 6.9% overall home death rate is the lowest documented in the literature and is approached only by a Japanese study with a reported home death rate of 8.2%.30 The data offer several possible explanations for this low rate of home death—sex (all women), age (a third of women younger than 50 years), low involvement of general practitioners, a predominance of care by specialists through hospital-based outpatient clinics, a high proportion of women needing to go the emergency department repeatedly to receive care, and few home visits—as most of these factors have been associated with a lower probability of dying at home.8, 30-34 While it was not possible to document other factors such as socioeconomic level, presence of family caregivers, or people's wish to die at home, a recent study showed that the most important factor contributing to a low rate of home death is the accessibility and utilization of acute care beds.35 High accessibility, defined as a rate of hospital beds of greater than 3.3 beds per 1,000 beneficiaries, was associated with a lower rate of home death and a higher rate of hospital death. The Province of Quebec, with a rate of 4.4 hospital beds per 1,000 inhabitants,36 would be considered a high-accessibility area, and this high rate may explain the low home death rate. This is the first study attempting to use indicators of care available from administrative databases, other than place of death, to define a more palliative care–oriented approach in the care of a population of women dying of breast cancer. However most of these indicators could be applied with little adjustment to other types of cancer or illnesses. The indicators of the last hospital admission leading to death were created with the objective of distinguishing between terminal admissions due to insufficient access to palliative care beds and medical admissions. The cut point of 7 days was chosen, as physicians were likely to recognize the terminal state of the women. With the possibility of imminent death, even if physicians have difficulties in predicting death,18, 19 they would be more likely to limit their interventions and shift services more toward providing comfort care and family counseling, and limiting consultation to a palliative care specialist. However, prolonged terminal admissions with involvement of at least one consultant other than a palliative care specialist seemed to be the norm. Unfortunately, consultation with specialists in palliative care was especially rare (9.7%). This low rate of palliative care consultation, coupled with a comparably low rate of death in a palliative care beds (15.7% to 16.5%) suggests, probably, not so much a lack of will to refer, but rather a scarcity of palliative care specialists working in cancer centers and acute care hospitals. All these findings suggest that, for the period from 1992 to 1999, sufficient resources for end-of-life care were not in place to meet the needs of the women dying of breast cancer. The indicators suggestive of a palliative care–oriented model during the last 6 months of life were infrequently present, with only a small number of women receiving follow-up by a general practitioner in the community and with home visits. A more medical-oriented model was the norm for this period—women were mostly treated by specialists, attended outpatient clinics, and had repeated emergency department attendances, with or without subsequent admissions in acute care beds. This pattern of care was especially the case for the younger women (< 50 years old), as demonstrated by the results of the ordinal regression models for number of admissions, emergency department visits, and home visits. This impression that a palliative care–oriented approach was not the norm in the care of younger women was confirmed by the results of the modeling of the care-oriented profile score by age. It must be remembered that the palliative care–oriented profile is simply a numerical way of summarizing specific features of care that have been recommended in the literature as being indicative of good end-of-life care.14 It was constructed for the purposes of facilitating statistical analysis. The validity of this model has not been formally tested. The testing of such of an index is beyond the scope of this article. However, the ranking of the places of death makes clinical sense and the value of one for each index of care made it possible to have a balanced weight between place of death and last 6 months of life. The direction of the cumulative odds ratios across age groups suggests that patient-physician relationship regarding treatment decision was influenced by the woman's age in a consistent way. It is impossible to know if this age-related disparity in pattern of care was mainly the result of the physician's attitude or patient's request. Surely a mutual influence probably occurred. However, while a more care-oriented approach could be understandable in the care of the very old women, the presence of a more medical-oriented approach with the younger women so close to death is worrisome, as they are often caring for children and are therefore more likely to experience more intense psychological and existential distress. A more palliative care–oriented approach may better suit their needs at the end of life. We recognize some limitations of this study of end-of-life care using administrative data. It is clear that none of the predefined indicators guarantee good provision of palliative care. For example, one woman could die at home in great distress without proper home care, and another, in an acute surgical unit with good symptom control and support from the nursing staff. However, we attempted to choose indicators that made sense from the present literature. Clearly, further work is needed to validate these indicators of a palliative care–oriented approach, and to create others that will capture additional important elements of care such as homecare services, provision of palliative radiotherapy, and utilization of specialized pain control techniques. Nevertheless, even if this methodology could be further developed requiring adjustments for other diseases, it is already sufficient to document the shortcomings of the present delivery of health care services to the terminally ill. Home death is an important element of good end-of-life care. In our study, this single element was given a relative value of 5, and, as indicated above, the number of women given this value is likely to be slightly overestimated, and may be so for older women. If in fact a bias was introduced by misclassification of place of death of older women, this could only influence the result of our analysis to a marginal degree. By giving the same proportion of home death to older women as in younger women, this will reduce the absolute percentage by only 1.6% (Table 2). Also, the only possible places of death for these older women are chronic care or hospice beds, reducing the care-oriented profile score by 1 or 2 points in this small proportion of older women. We considered the effect of comorbid conditions in our analysis in two ways—in the design phase, through restriction, and in the final analysis, as a confounder. By classifying women who died of comorbid conditions such as stroke, acute myocardial infarction, and congestive heart failure as women who died with breast cancer, and by subsequently excluding this group, we restricted the sample. The inclusion of comorbidity as a confounder in the model did not change the estimate of the effect of age on type of care received by older women included in our study population. This lack of influence of comorbidity on patterns of care at the end of life is in keeping with a previous study that found no influence of comorbidity on age-related patterns of initial treatment of elderly women with nonmetastatic breast cancer.37 This type of study extends the valuable work done qualitatively on the experience of women dying of breast cancer38-42 and reflects a population view of the magnitude of the issue. Our study was successful, using existing administrative databases, in documenting important indicators of the quality of health care services provided at the end of life. Similar methodologies could be used to evaluate other health care delivery systems, to help promote better health care policies for the dying, and to monitor changes over time, especially after the introduction of these new health service policies.
The authors indicated no potential conflicts of interest.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
1. World Health Organization Expert Committee: World Health Organization Techinica Report Series 804: Cancer Pain Relief and Palliative Care. Geneva, Switzerland, World Health Organization, 1990 2. Seale CF: What happens in hospices: A review of research evidence. Soc Sci Med 28:551-559, 1989 3. Higginson IJ, Finlay I, Goodwin DM, et al: Do hospital-based palliative teams improve care for patients or families at the end of life? J Pain Symptom Manage 23:96-106, 2002[CrossRef][Medline]
4. Hearn J, Higginson IJ: Do specialist palliative care teams improve outcomes for cancer patients? A systematic literature review. Palliat Med 12:317-332, 1998
5. Bruera E, Neumann CM, Gagnon B, et al: Edmonton Regional Paliative Care Program: Impact on patterns of terminal care. CMAJ 161:290-293, 1999
6. Cohen SR, Boston P, Mount BM, et al: Changes in quality of life following admission to palliative care units. Palliat Med 15:363-371, 2001 7. Constantini M, Camoirano E, Madeddu L, et al: Palliative home care and place of death among cancer patients: A population-based study. Palliat Med 7:323-331, 1993[Medline]
8. Higginson IJ, Jarman B, Astin P, et al: Do social factors affect where patients die: An analysis of 10 years of cancer deaths in England. J Public Health Med 21:22-28, 1999 9. Sager MA, Easteling DV, Kindig DA, et al: Changes in the location of death after passage of medicare's prospective payment system. N Engl J Med 320:433-439, 1989[Abstract] 10. Catalan-Fernandez JG, Pons-Sureda O, Recober-Martinez A, et al: Dying of cancer: The place of death and family circumstances. Med Care 29:841-852, 1991[CrossRef][Medline]
11. Moinpour CM, Polissar L: Factors affecting place of death of hospice and non-hospice cancer patients. Am J Public Health 79:1549-1551, 1989 12. Townsend J, Frank AO, Fermont D, et al: Termial cancer care and patients' preference for place of death: A prospective study. BMJ 301:415-417, 1990
13. Hays JC, Galanos AN, Palmer TA, et al: Preference for place of death in a continuing care retirement community. Gerontologist 41:123-128, 2001
14. Earle CC, Park ER, Lai B, et al: Identifying potential indicators of the quality of end-of-life cancer care from administrative data. J Clin Oncol 21:1133-1138, 2003 15. Teno JM, Weitzen S, Fennell ML, et al: Dying trajectory in the last year of life: Does cancer trajectory fit other diseases? J Palliat Med 4:457-464, 2001[CrossRef][Medline]
16. Lunney JR, Lynn J, Foley DJ, et al: Patterns of functional decline at the end of life. JAMA 289:2387-2392, 2003 17. Oxford Textbook of Palliative Medicine. New York, NY, Oxford University Press, 1993 18. Vigano A, Dorgan M, Bruera E, et al: The relative accuracy of the clinical estimation of the duration of life for patients with end of life cancer. Cancer 86:170-176, 1999[CrossRef][Medline]
19. Christakis NA, Lamont EB: Extent and determinants of error in doctors' prognoses in terminally ill patients: Prospective cohort study. BMJ 320:469-472, 2000 20. Levenson JW, McCarthy EP, Lynn J, et al: The last six months of life for patients with congestive heart failure. J Am Geriatr Soc 48:S101-S109, 2000 (suppl 5) 21. Lynn J, Ely EW, Zhong Z, et al: Living and dying with chronic obstructive pulmonary disease. J Am Geriatr Soc 48:S91-100, 2000 (suppl 5)[Medline] 22. McCarthy EP, Phillips RS, Zhong Z, et al: Dying with cancer: Patients' function, symptoms, and care preferences as death approaches. J Am Geriatr Soc 48:S110-S121, 2000 (suppl 5)[Medline] 23. Murphy DJ, Knaus WA, Lynn J: Study population in SUPPORT: Patients (as defined by disease categories and mortality projections), surrogates, and physicians. J Clin Epidemiol 43:11S-28S, 1990 24. National Cancer Institute of Canada: Canadian Cancer Statistics. Toronto, Canada, Canadian Cancer Society, Health Canada, 2000 25. Cha CH, Kennedy GD: Metastic breast cancer. Surg Clin North Am 79:1117-1141, 1999[CrossRef][Medline]
26. Mayo NE, Scott SC, Shen N, et al: Waiting time for breast cancer surgery in Quebec. CMAJ 164:1133-1138, 2001 27. Campora E, Gardin G, Gasco M, et al: Metastatic breast cancer patients failing first-line, anthracycline-containing chemotherapy: Is further therapy of benefit? Anticancer Res 19:3429-3432, 1999[Medline] 28. Scott SC, Goldberg MS, Mayo NE: Statistical assessment of ordinal outcomes in comparative studies. J Clin Epidemiol 50:45-55, 1997[CrossRef][Medline] 29. Kleinbaum DG, Klein M: Original Logistic Regression: Statistics for Biology and Health—A Self-Learning Text. New York, NY, Springer, 2002, pp 304-310 30. Tang ST, McCorkle R: Determinants of place of death for terminal cancer patients. Cancer Invest 19:165-180, 2001[CrossRef][Medline]
31. Tolle SW, Rosenfeld AG, Tilden VP, et al: Oregon's low in-hospital death rates: What determines where people die and satisfaction with decisions on place of death? Ann Intern Med 130:681-685, 1999 32. Mor V, Hiris J: Determinants of site of death among hospice cancer patients. J Health Soc Behav 24:375-385, 1983[CrossRef][Medline] 33. De Conno F, Caraceni A, Groff L, et al: Effect of home care on the place of death of advanced cancer patients. Eur J Cancer 32A:1142-1147, 1996 34. Thorne CP, Seamark DA, Lawrence C, et al: The influence of general practitioner community hospitals on the place of death of cancer patients. Palliat Med 8:122-128, 1994[Medline] 35. Pritchard RS, Fisher ES, Teno JM, et al: Influence pf patient preference and local health system characteristics on the place of death. J Am Geriatr Soc 46:1242-1250, 1998[Medline] 36. Agences statistique fédérales du Canada: Nombre de lits d'hôpitaux pour 1000 habitants. Montreal, Canada, Institut de la statistique Quebec, Gouvernement du Canada, Gouvernement du Québec, 2002
37. Newschaffer CJ, Penberthy L, Desch CE, et al: The effect of age and comorbidity in the treatment of elderly women with nonmetastatic breast cancer. Arch Intern Med 156:85-90, 1996 38. Davies M, Sque M: Living on the outside looking in: A theory of living with advanced breast cancer. Int J Palliat Nurs 8:583-590, 2002[Medline]
39. Chiu L: Transcending breast cancer, transcending death: A Taiwanese population. Nurs Sci Q 13:64-72, 2000
40. Sahlberg-Blom E, Ternestedt BM, Johansson JE: Patient participation in decision making at the end of life as seen by a close relative. Nurs Ethics 7:296-313, 2000
41. Hilton BA, Crawford JA, Tarko MA: Men's perspectives on individual and family coping with their wives' breast cancer and chemotherapy. West J Nurs Res 22:438-459, 2000 42. Whelan M: Teaching the value of qualitative research. Nurse Educ 24:4, 1999 Submitted June 24, 2003; accepted May 8, 2004.
Related Editorial
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2004 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
2) was carried out. This analysis was repeated with the comorbidity score as a confounder (as an ordinal variable with possible values of 0, 1, 2, and > 2) to test the influence of comorbidity on age-related patterns of care provided at the end of life. Some of the indices had multiple values (eg, number of admissions), and they were dichotomized at a certain cut point. To exclude the possibility that this choice of the cut point may have artificially introduced the association with age, the same procedure was used to test the influence of age on the number of admissions (using five cut points: 0, 1, 2, 3, > 3), the number of emergency department visits (using four cut points: 0, 1, 2, > 2), and the number of home visits (using six cut points: 0, 1, 2, 3, 4, and > 4, all of which are ranked interval in nature) to ensure that the influence of age was constant across the cut points of these indicators. 
