Vol. 47 No. 1 January 2014
Philippa H. Hawley, BMed, FRCPC Department of Palliative Care British Columbia Cancer Agency Vancouver, British Columbia, Canada E-mail: [email protected] http://dx.doi.org/10.1016/j.jpainsymman.2013.10.009
Disclosures and Acknowledgments The author expresses her appreciation to Dr. Shalini Nayar for editorial assistance.
References 1. World Health Organization. Cancer pain relief. Geneva: WHO, 1986. 2. World Health Organization. Definition of palliative care. 2013. Available from http://www.who.int/cancer/ palliative/definition/en/. Accessed November 20, 2013. 3. Smith T, Temin S, Alesi E, et al. American Society of Clinical Oncology provisional clinical opinion: the integration of palliative care into standard oncology care. J Clin Oncol 2012;30:880e887. 4. Temel J, Greer J, Muzikansky M, et al. Early palliative care for patients with metastatic non-small cell lung cancer. N Engl J Med 2010;363:733e742. Available from http://www.nejm.org/doi/pdf/10.1056/ NEJMoa1000678. Accessed November 20, 2013. 5. Canadian Hospice Palliative Care Association and Quality End-of-life Care Coalition of Canada. ‘‘The way forward integration initiative’’ resources: synthesis of recommendations from national reports on hospice palliative care (May 2012); The palliative approach: improving care for Canadians with lifelimiting illnesses (August 2012); Integrating a palliative approach into the management of chronic, life-threatening diseases: who, how and when? (December 2012); Cost-effectiveness of palliative care: a review of the literature (December 2012). All documents available from http://www.hpcintegration.ca/ resources. Accessed November 20, 2013. 6. Biasco G, Tanzi S, Bruera E. Early palliative care: how? J Pall Med 2013;16:466e470. 7. Bruera E, Hui D. Integrating supportive and palliative care in the trajectory of cancer: establishing goals and models of care. J Clin Oncol 2010; 28:4013e4017. 8. Bhang T. Creating a climate for healing: a visual model for goals of care discussions. [Letter]. J Palliat Med 2013;16:718. 9. Victoria Hospice Society. Medical care of the dying, 4th ed. Victoria, BC: Victoria Hospice Society, 2006:15e19. 10. Morrison RS. Research priorities in geriatric palliative care. J Palliat Med 2013;16:726e729.
Letters
e5
Serial Prognostication: A New Look at an Old Tool To the Editor: We read the article by Arai et al. with interest.1 The authors conducted a retrospective study to examine the association between a change in the Palliative Prognostic Index (PPI) and survival in 374 cancer patients admitted to a palliative care unit in Japan. In multivariate analysis, they found a highly significant association between PPI change and survival (hazard ratio 6.6 per point increase in PPI; 95% CI, 4.9e9), independent of baseline PPI scores. This study suggests that the PPI is not only a prognostic tool but one that is sensitive to change, and importantly, the degree of change has prognostic utility. Some additional information would help readers appreciate the significance of their findings. The PPI comprises five variables: the Palliative Performance Scale score, oral intake, dyspnea, delirium, and edema.2 It would be useful to know the breakdown of PPI scores on admission and at follow-up. Did some of the PPI variables change more than others? Were some of the variables stable over this time period? Based on the data provided, the magnitude of PPI change appeared to be small (median ¼ 0; interquartile range, 0e0.57). A better understanding of how the PPI evolves over time may facilitate future research in this area. How the PPI data were collected also could have a major impact on the interpretation of study findings. Specifically, more information on the physician(s) who collected the data, and how each variable was assessed, would be crucial. How was dyspnea assessed especially in the context of delirium? Finally, because the PPI was assessed in the presence of other physiologic changes (e.g., death rattle), a prospective study adjusting for other prognostic variables would be needed. The timing of the second data point raises some important questions. It would be helpful to learn how the authors decided that the PPI should be repeated five to seven days later, and if any exploratory analysis was done to examine the optimal timing related to a change in the PPI. We wonder if a shorter interval would have similar discriminatory power
e6
for survival.3 By selecting a relatively ‘‘long’’ interval, many patients have either died or been discharged, which could potentially affect the generalizability of the study findings. For instance, the lower prognostic significance of baseline PPI in this cohort may be explained by the fact that all included patients lived for at least five days. Many important decisions in the palliative care unit, such as treatment choices and discharge planning, are dependent on an accurate prognosis.4 The study by Arai et al. raises some important questions on PPI changes. Future research should examine how serial prognostication could be refined, and more importantly, applied in the patient care setting to facilitate clinical decision making. Elizabeth Phan, DO David Hui, MD, MSc, FRCPC Department of Palliative Care and Rehabilitation Medicine The University of Texas M. D. Anderson Cancer Center Houston, Texas, USA E-mail: [email protected] http://dx.doi.org/10.1016/j.jpainsymman.2013.10.006
References 1. Arai Y, Okajima Y, Kotani K, et al. Prognostication based on the change in the palliative prognostic index for patients with terminal cancer. J Pain Symptom Manage 2013 July 22; [Epub ahead of print]. 2. Morita T, Tsunoda J, Inoue S, Chihara S. The Palliative Prognostic Index: a scoring system for survival prediction of terminally ill cancer patients. Support Care Cancer 1999;7:128e133. 3. Mori M, Parsons HA, De La Cruz M, et al. Changes in symptoms and inpatient mortality: a study in advanced cancer patients admitted to an acute palliative care unit in a comprehensive cancer center. J Palliat Med 2011;14:1034e1041. 4. Hui D, Kilgore K, Nguyen L, et al. The accuracy of probabilistic versus temporal clinician prediction of survival for patients with advanced cancer: a preliminary report. Oncologist 2011; 16:1642e1648.
Authors’ Reply to Phan and Hui To the Editor: We wish to express our appreciation to Phan and Hui for their insightful comments. We
Letters
Vol. 47 No. 1 January 2014
examined the association between a change in the Palliative Prognostic Index (PPI) and survival in cancer patients admitted to a palliative care unit in Japan.1 The average length of stay was 26.7 days, and 83.6% of the patients died in our unit during the study. The PPI change per day (DPPI) was significantly and independently predictive for death within three weeks (adjusted hazard ratio: 6.6; 95% CI: 4.9e9.0). A previous study by Lunney et al. demonstrated that the cancer decedents were highly functional early in their final year but markedly more disabled three months before death.2 Thus, the results of our study are not surprising, although the hazard ratio found in the study appeared to be relatively high.2 Phan and Hui also suggested that the changes in the respective PPI components might be useful. Changes in specific PPI components were not noted in our study. Because the overall change in DPPI was small, a separate analysis of the respective PPI components seems to have little value, at least in our patient population. Phan and Hui turned their attention to the presence of dyspnea at rest, especially among patients with delirium. The symptom was evaluated based on one of the complaints of the patients in our study. If the evaluation was incomplete, the presence of the symptom was assessed based on the medical records and discussions with family members or the nursing staff as per the protocol of the original study of the development and validation of the PPI.3 Bennett et al. have reported that survival correlated with the mean ‘‘weekly’’ change of the modified Barthel Index.4 Godfrey et al. reported that patients with a Barthel Index that decreased by 10 or more per week were significantly more likely to have a short prognosis.5 Based on these prior ideas,4,5 we decided to evaluate the PPI change in the five to seven days after admission. We did not examine which period would be the most effective for predicting the prognosis in the experimental protocol before conducting our study. Such a study to determine the most appropriate timing to evaluate the PPI change may merit further investigation. Our study simply provided a challenging point of view on the use of the PPI. Thus, the inclusion of the views of Phan and Hui
Philippa H. Hawley, BMed, FRCPC Department of Palliative Care British Columbia Cancer Agency Vancouver, British Columbia, Canada E-mail: [email protected] http://dx.doi.org/10.1016/j.jpainsymman.2013.10.009
Disclosures and Acknowledgments The author expresses her appreciation to Dr. Shalini Nayar for editorial assistance.
References 1. World Health Organization. Cancer pain relief. Geneva: WHO, 1986. 2. World Health Organization. Definition of palliative care. 2013. Available from http://www.who.int/cancer/ palliative/definition/en/. Accessed November 20, 2013. 3. Smith T, Temin S, Alesi E, et al. American Society of Clinical Oncology provisional clinical opinion: the integration of palliative care into standard oncology care. J Clin Oncol 2012;30:880e887. 4. Temel J, Greer J, Muzikansky M, et al. Early palliative care for patients with metastatic non-small cell lung cancer. N Engl J Med 2010;363:733e742. Available from http://www.nejm.org/doi/pdf/10.1056/ NEJMoa1000678. Accessed November 20, 2013. 5. Canadian Hospice Palliative Care Association and Quality End-of-life Care Coalition of Canada. ‘‘The way forward integration initiative’’ resources: synthesis of recommendations from national reports on hospice palliative care (May 2012); The palliative approach: improving care for Canadians with lifelimiting illnesses (August 2012); Integrating a palliative approach into the management of chronic, life-threatening diseases: who, how and when? (December 2012); Cost-effectiveness of palliative care: a review of the literature (December 2012). All documents available from http://www.hpcintegration.ca/ resources. Accessed November 20, 2013. 6. Biasco G, Tanzi S, Bruera E. Early palliative care: how? J Pall Med 2013;16:466e470. 7. Bruera E, Hui D. Integrating supportive and palliative care in the trajectory of cancer: establishing goals and models of care. J Clin Oncol 2010; 28:4013e4017. 8. Bhang T. Creating a climate for healing: a visual model for goals of care discussions. [Letter]. J Palliat Med 2013;16:718. 9. Victoria Hospice Society. Medical care of the dying, 4th ed. Victoria, BC: Victoria Hospice Society, 2006:15e19. 10. Morrison RS. Research priorities in geriatric palliative care. J Palliat Med 2013;16:726e729.
Letters
e5
Serial Prognostication: A New Look at an Old Tool To the Editor: We read the article by Arai et al. with interest.1 The authors conducted a retrospective study to examine the association between a change in the Palliative Prognostic Index (PPI) and survival in 374 cancer patients admitted to a palliative care unit in Japan. In multivariate analysis, they found a highly significant association between PPI change and survival (hazard ratio 6.6 per point increase in PPI; 95% CI, 4.9e9), independent of baseline PPI scores. This study suggests that the PPI is not only a prognostic tool but one that is sensitive to change, and importantly, the degree of change has prognostic utility. Some additional information would help readers appreciate the significance of their findings. The PPI comprises five variables: the Palliative Performance Scale score, oral intake, dyspnea, delirium, and edema.2 It would be useful to know the breakdown of PPI scores on admission and at follow-up. Did some of the PPI variables change more than others? Were some of the variables stable over this time period? Based on the data provided, the magnitude of PPI change appeared to be small (median ¼ 0; interquartile range, 0e0.57). A better understanding of how the PPI evolves over time may facilitate future research in this area. How the PPI data were collected also could have a major impact on the interpretation of study findings. Specifically, more information on the physician(s) who collected the data, and how each variable was assessed, would be crucial. How was dyspnea assessed especially in the context of delirium? Finally, because the PPI was assessed in the presence of other physiologic changes (e.g., death rattle), a prospective study adjusting for other prognostic variables would be needed. The timing of the second data point raises some important questions. It would be helpful to learn how the authors decided that the PPI should be repeated five to seven days later, and if any exploratory analysis was done to examine the optimal timing related to a change in the PPI. We wonder if a shorter interval would have similar discriminatory power
e6
for survival.3 By selecting a relatively ‘‘long’’ interval, many patients have either died or been discharged, which could potentially affect the generalizability of the study findings. For instance, the lower prognostic significance of baseline PPI in this cohort may be explained by the fact that all included patients lived for at least five days. Many important decisions in the palliative care unit, such as treatment choices and discharge planning, are dependent on an accurate prognosis.4 The study by Arai et al. raises some important questions on PPI changes. Future research should examine how serial prognostication could be refined, and more importantly, applied in the patient care setting to facilitate clinical decision making. Elizabeth Phan, DO David Hui, MD, MSc, FRCPC Department of Palliative Care and Rehabilitation Medicine The University of Texas M. D. Anderson Cancer Center Houston, Texas, USA E-mail: [email protected] http://dx.doi.org/10.1016/j.jpainsymman.2013.10.006
References 1. Arai Y, Okajima Y, Kotani K, et al. Prognostication based on the change in the palliative prognostic index for patients with terminal cancer. J Pain Symptom Manage 2013 July 22; [Epub ahead of print]. 2. Morita T, Tsunoda J, Inoue S, Chihara S. The Palliative Prognostic Index: a scoring system for survival prediction of terminally ill cancer patients. Support Care Cancer 1999;7:128e133. 3. Mori M, Parsons HA, De La Cruz M, et al. Changes in symptoms and inpatient mortality: a study in advanced cancer patients admitted to an acute palliative care unit in a comprehensive cancer center. J Palliat Med 2011;14:1034e1041. 4. Hui D, Kilgore K, Nguyen L, et al. The accuracy of probabilistic versus temporal clinician prediction of survival for patients with advanced cancer: a preliminary report. Oncologist 2011; 16:1642e1648.
Authors’ Reply to Phan and Hui To the Editor: We wish to express our appreciation to Phan and Hui for their insightful comments. We
Letters
Vol. 47 No. 1 January 2014
examined the association between a change in the Palliative Prognostic Index (PPI) and survival in cancer patients admitted to a palliative care unit in Japan.1 The average length of stay was 26.7 days, and 83.6% of the patients died in our unit during the study. The PPI change per day (DPPI) was significantly and independently predictive for death within three weeks (adjusted hazard ratio: 6.6; 95% CI: 4.9e9.0). A previous study by Lunney et al. demonstrated that the cancer decedents were highly functional early in their final year but markedly more disabled three months before death.2 Thus, the results of our study are not surprising, although the hazard ratio found in the study appeared to be relatively high.2 Phan and Hui also suggested that the changes in the respective PPI components might be useful. Changes in specific PPI components were not noted in our study. Because the overall change in DPPI was small, a separate analysis of the respective PPI components seems to have little value, at least in our patient population. Phan and Hui turned their attention to the presence of dyspnea at rest, especially among patients with delirium. The symptom was evaluated based on one of the complaints of the patients in our study. If the evaluation was incomplete, the presence of the symptom was assessed based on the medical records and discussions with family members or the nursing staff as per the protocol of the original study of the development and validation of the PPI.3 Bennett et al. have reported that survival correlated with the mean ‘‘weekly’’ change of the modified Barthel Index.4 Godfrey et al. reported that patients with a Barthel Index that decreased by 10 or more per week were significantly more likely to have a short prognosis.5 Based on these prior ideas,4,5 we decided to evaluate the PPI change in the five to seven days after admission. We did not examine which period would be the most effective for predicting the prognosis in the experimental protocol before conducting our study. Such a study to determine the most appropriate timing to evaluate the PPI change may merit further investigation. Our study simply provided a challenging point of view on the use of the PPI. Thus, the inclusion of the views of Phan and Hui
