Multidisciplinary clinical prediction of survival of patients attending Palliative Radiation Oncology and Palliative Care clinics: a prospective cohort study

Introduction

Estimating prognosis in the setting of advanced or incurable cancer has a significant impact on patients and families, regarding their ability to plan their future and make decisions about medical treatment (1). Prognosis influences interventions such as initiation or termination of systemic therapy or radiotherapy, artificial nutrition, setting of care, clinical trial eligibility, referral to specialist services, and accessing hospice (1-4). As such, many patients wish to have their prognosis clearly communicated, allowing them to make critical decisions as they near end of life (1,5-10).

There are many challenges in formulating an accurate clinical prediction of survival (CPS) (11,12). There is a well-established propensity of health care providers (HCPs) to overpredict survival (11,13-25), with some publications outlining CPS exceeding actual survival (AS) by as much as 3- to 5-fold (15). Factors utilized by clinicians to predict lifespan range from clinical experience to lab results, comorbidities, and disease-related parameters (9). CPS is dynamic and its formulation differs between early-stage and advanced settings (20). Factors that are often predictive in the earlier stages of disease (such as primary site or grade), tend to be less important than factors such as symptoms and performance status (PS) in metastatic cancer (21,26-28). Additionally, physician CPS may be more accurate in patients with a very short (<14 days) or very long (>1 year) survival, but clinicians may be less able to consistently identify those with an intermediate prognosis of weeks or months (20,29).

Care for patients with advanced cancer is optimally provided by multidisciplinary teams (MDTs), with each HCP providing specialized skills and knowledge. An MDT-based approach can help patients and families navigate medical treatments, as well as address diverse psychosocial, spiritual, rehabilitative, and dietary needs (30,31). Thus, a team may include oncologists, Palliative Care (PC) physicians, registered nurses, nurse practitioners, registered dietitians, physiotherapists, social workers, and others. Each discipline brings their unique perspective regarding patient prognosis (14,16,32). Additionally, trainees and early career clinicians also rotate through these MDT, gaining experience in formulating CPS (14).

To date, there have been few studies describing physician and non-physician prognostication in palliative settings. A recent study reported the tendency to overpredict survival of patients assessed for palliative radiotherapy (PRT) by oncologists, radiation therapists and allied health professions (32). In another study, a consensus prediction of survival was constructed by a MDT (PC physicians, PC nurses, a social worker, pharmacist, and a pastoral care worker); individual perspectives were not reported (33). Our objectives were to compare CPS performed by multiple disciplines experienced in providing specialist palliative and supportive care, and palliative radiation therapy. We present this article in accordance with the STROBE reporting checklist (available at https://apm.amegroups.com/article/view/10.21037/apm-25-108/rc).

Methods

Patient populations

The Cross Cancer Institute (CCI) is an academic tertiary cancer center in the province of Alberta, Canada, serving a catchment area of approximately 1.5 million people (34). The CCI offers comprehensive oncologic care including consultation, systemic and radiation treatments, education, and research in both inpatient and outpatient settings. Two referral-based clinical services providing care for patients with advanced cancer were included in this analysis:

• The Palliative Radiation Oncology (PRO) program provides expedited access to palliative radiation therapy as well as supportive care (32). The PRO team consists of radiation oncologists, nurse practitioners, registered nurses, pharmacists, radiation therapists, occupational therapists, social workers, and registered dieticians, with other disciplines accessed as required.
• The specialist PC team provides consultation for ambulatory patients with inadequately controlled pain or other symptoms. Assessments by PC physicians, registered nurses and residents were included.

Data collection

The time period of data collection was 06/2010 to 12/2014. In a prospective, confidential, and independent manner, after usual direct in-person clinical assessment of consecutive patients, each HCP provided an estimate of expected survival in days, weeks, months, or years (20). The HCP expressed survival prediction as a single number and unit of time and were not given predetermined ranges of survival or categories to choose from. Along with each prediction, HCP submitted in free text up to 4 factors that influenced their estimate. Factors referring to similar underlying concepts, characteristics, or clinical observations were aggregated. For example, the factor “reasonable performance status” was coded as “performance status”, and similarly “liver and bone metastases” was coded as “extent of disease”. Each prediction was made in the context of a therapeutic clinical encounter; no estimate was provided based solely on medical record review. Patient demographics, CPS, AS, and factors influencing each prediction were collected and anonymized. The individual identity of each HCP was not recorded to preserve anonymity; however, clinical encounters of patients attending the two programs during the study time period occurred with 23 Radiation Oncologists and three PC physicians. The sample size was not predefined but was determined by the number of patients seen within each program over the study period.

Statistical analysis

As a matter of routine care, over the time of the study, some patients were seen twice in one clinic, or in both clinics. Given that the individual clinicians comprising the teams changed from week to week and previous survival predictions were not available at the time of subsequent visits, repeat visits’ predictions were analyzed as independent values.

CPS was first considered ‘correct’ if within 30 days of AS (20,32). However, this criterion is too strict for AS of longer duration. For example, at 1 year of survival, 30 days’ error is only 8.2%. Thus, we also considered that CPS was correct if within 30% of AS, as has also been previously reported (11,15,17,23).

Summary statistics were calculated, including means and standard deviations for continuous variables and frequency and percentages for categorical variables. Overall survival (OS) was measured from the date of clinic visit to the date of death. Kaplan-Meier estimates of median OS and 95% confidence interval (95% CI) were obtained.

Time to event analysis (log-rank test) compared AS, as well as the absolute quantity of overprediction, in days. Cox proportional hazard analyses (univariable and multivariable) evaluated the timespan of overprediction in days and adjusted for any effects of patient age, primary tumor histology, and duration of AS. Variables significant at the P<0.05 level in univariable analyses were entered into the multivariable model. Chi-squared test (test of proportions) tested the dependent variable, prognostic accuracy (yes vs. no), by discipline group. No artificial intelligence (AI) tools were used in the collection or analysis of data, the production of images, or in the writing of the manuscript.

Ethics approval

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the provincial Heath Research Ethics Board of Alberta-Cancer Committee, file HREBA.CC-17-0437. The Institutional Review Board waived informed consent requirements since predictions were analyzed only after the care plan had been established and carried out.

Results

Clinicians assessed 980 patients, of whom 944 have died and form the study population (Table 1). Primary tumor site case mix differed between clinics as expected, influenced by the teams’ respective roles in clinical care. The median OS was 30 days shorter in the PC clinic population than in the palliative radiation setting (Figure 1).

Figure 1 Actual survival by clinic. Palliative Radiation Oncology clinic median survival: 129 days (95% CI: 114–143). Palliative Care clinic median survival: 92 days (95% CI: 75–109) (log-rank test; P<0.001). CI, confidence interval.

Criteria for correct survival prediction were first compared. Regarding the 30-day criterion of correctness, a total of 576/2,776 (20.7%) predictions met this criterion. For the 30% criterion, 628/2,776 (22.6%) predictions met this criterion. Some predictions met both criteria; thus we report as correct predictions which met either one or both criteria for correctness (N=852) (Figure S1). The 30-day criterion mainly is correct at shorter survival times of less than 200 days, whereas the 30% criterion is not very likely to be correct at survival times of less than 100 days but captures more correct predictions on a longer time scale. All further data are presented as correct by both criteria applied.

Survival prediction metrics are provided in Table 2, including the number of predictions, median actual and predicted survival, and % difference between actual and predicted survival.

Correct and incorrect predictions by clinic and discipline are shown in Table 3. Overall, 30.7% of predictions were correct (i.e., met either one or both criteria for correctness); of the remaining, inaccurate, ones, 22.2% were underpredictions and 47.1% were overpredictions. PC physicians and PC registered nurses made correct predictions just over 40% of the time, and this was significantly higher than most other HCP. Pharmacists had a lower frequency of correct predictions (21.3%) than all other HCP. The median number of days of overprediction also varied by HCP, with PC physicians overpredicting by the shortest amount [a median of 47 days (95% CI: 39–55)] compared with all other HCP. Pharmacists demonstrated the longest overprediction [a median of 161 days (95% CI: 135–187 days)].

The number of days of overprediction was subjected to univariable and multivariable analysis (Table 4). Compared with PC physicians, almost all assessors were significantly more likely to overpredict at the univariable level. The greatest overpredictions were by medical students, pharmacists, and Allied Health Professionals. Differences among the assessor groups persisted after adjustment for patient gender, primary tumor site, and duration of AS. The exception was the PC residents being not significantly different from PC physicians; however, this was the assessor group who provided the smallest number of predictions.

Patients surviving 14 days or less from consultation were reviewed (N=47). Nine of 47 (19.1%) were predicted to survive 6 months or greater and all but one of these optimistic CPS was made by a Radiation Oncologist. None was receiving systemic therapy at the time of the clinic visit.

Factors used to make predictions

Combining all parameters mentioned as either the first or second ranked factor influencing prediction, disease extent (29.1%), primary histology (23.0%), PS (17.4%), and age (6.3%) were most frequently cited. These four parameters accounted collectively for 75.8% of factors submitted. The remaining 24.2% included comorbidities, treatment response, symptoms, attitude/appearance, family support, weight loss, lab values and other. Each discipline reported a different pattern of factors used in CPS; however, the factors chosen did not differ significantly between correct or incorrect predictions (data not shown). For HCP groups who made at least 130 predictions, the top three factors cited in making their CPS is illustrated in Figure 2. PS and disease extent were the most frequently used factors by PC physicians, whereas PC nurses mainly cited disease extent and used a wider range of additional factors (Figure 2A). In the MDT within the PRO program, a wide variation of factors was reported (Figure 2B). Disease extent was the most consistently used factor across all disciplines. PS was used mainly by Radiation Oncologists and registered nurses, who were less likely to rely on primary disease histology than other members of the team.

Figure 2 Proportions of the top three factors used in constructing survival predictions by different assessor groups, for groups providing at least 130 predictions. (A) Specialist palliative care clinic. (B) Palliative radiation oncology clinic. Black bars represent factor 1 i.e., the topmost factor used by the healthcare professionals in survival prediction. The sum of all the black bars within each healthcare professional group is equal to 1.0 (i.e., 100%). Grey bars represent the second most identified factor, and white bars represent the third most identified factor. “Not used” indicates a factor was not provided. Factor 1, factor 2 and factor 3 were all different across assessor categories (P<0.001, Chi-squared test). MD, physician; perf, performance; RN, registered nurse; supp, support; Tx, treatment.

Discussion

In this largest single study of physician and non-physician prognostication in advanced cancer, we evaluated two groups of clinicians experienced in providing PRT and PC. The programs included a wide case mix of patients with AS ranging from short days to years. Our results confirm the propensity for overprediction, which was present to varying degrees in all disciplines.

For prognostication, CPS is simple, practical, prompt and convenient, and remains the standard of care against which new methods should be measured (1). The mental process of CPS is complex and involves consciously or unconsciously considering various factors (35,36). These include illness trajectory, PS, symptom burden, psychological indications, and lab values (35,36). Variability in the weighting given to each parameter can result in heterogeneous and inaccurate estimations (18). Overestimates of survival can be related to the rapid onset of complications in advanced stage disease, while underestimates can result from improvement of tumour-related pathologies (e.g., bowel obstruction) or reversible benign complications (37). Accuracy of CPS may also be impacted by initiation, response, progression and discontinuation of successive lines of systemic treatment (38).

Clinical predictions of survival can be divided into two types: temporal (a patient will live a certain amount of time, such as in our study), and probabilistic (a patient has a percentage chance of survival at a certain time point) (18). In general, temporal predictions are less accurate than probabilistic, with significant literature, including this study, confirming a tendency towards overprediction (11,13-16,18,20-22,24-25,32). However, temporal predictions are more commonly discussed (18), and easier to explain to patients and families (5,8).

It is essential to define what constitutes an ‘accurate’ prediction, which varies between studies and makes comparisons between studies challenging. Some define a CPS as correct if within 30 days of AS (32) and others use CPS within 30% of AS (11). Our data suggest that the 30 days’ criterion is too strict for survival times exceeding ~200 days, since to be accurate for someone expected to live 1 year, a CI of 0.92–1.08 would be required. However, the 30% criterion is perhaps too strict for a shorter prognosis; to be accurate for someone expected to survive for 10 days, CPS would need to fall within 3 days. Thus, we considered a clinical prediction to be accurate if it was either within 30 days or within 30% of AS. By combining the two criteria we can encompass the broad survival spectrum in our patient population.

Twenty-five years ago, Christakis et al. reported that the accuracy of CPS is influenced by the duration of the doctor-patient relationship, with greater accuracy associated with longer duration (39). This echoed Oxenham et al., who described mild improvements in CPS accuracy over subsequent visits, attributed to increased familiarity of the physician with the patient (40). However, more recent studies have not confirmed this finding. Gwilliam et al. reported that accuracy was not affected by length of time that the clinician had known the patient (16). In a recent study, Radiation Oncologists provided 373 predictions for 172 patients, half of whom made more than one visit to the clinic over the study period. CPS did not show any significant changes over subsequent visit dates (10). There was no significant difference in the accuracy of CPS between all visits, including comparisons between subsequent visits and between every visit and the first visit (10). Another study reported CPS for 262 patients at their first outpatient PC consultation, finding no significant difference in accuracy by whether the physician had previously met the patient (36). Interestingly, physicians who are not directly involved in patient management may provide more accurate prognostication than those actively engaged in the patients’ care (41).

Seniority, training, and professional experience have also been previously reported as influencing CPS (3,7,20,27,42). However, a 2016 systematic review of 42 studies including over 12,000 predictions was unable to confirm a significant correlation between years of experience and accuracy (15). In fact, in a study of CPS in patients receiving PRT, Benson et al. found no association between physician experience and prediction accuracy (43). This was also reported by a group from Brazil, investigating CPS at the first outpatient visit with PC. Thirty-five doctors participated: 20 lead physicians and 15 residents/fellows. Mean AS was 124 days and overall accuracy was 32%. There were no significant differences in accuracy by lead physician versus resident; fewer than or more than 10 years of professional experience; or physician age older or younger than 35 years (36).

Additionally, the 2016 systematic review by White et al. could not identify a particular clinician group that was more consistently accurate than others (15). Oncologists usually meet patients at the start of anticancer treatment—often initial diagnosis—and follow them over a long period. Thus, they regularly observe changes in the patients’ condition (44). By contrast, PC physicians do not become involved until end-of-life is acknowledged (44). Urahama et al. therefore proposed that oncologists would be more accurate in survival prediction than PC physicians, who only observe a patient over a short period of time (44). On the other hand, physicians are also generally more accurate in predicting survival the closer the patient is to death, likely related to increasing symptom burden and decline in physical function (45). Once anticipated survival is short, care is often transitioned from Oncology to the PC team. Our two cohorts’ median survivals confirm this trend: the median AS was 92 days for patients assessed by PC versus 129 days in patients evaluated for radiotherapy.

In a prospective study from Japan, accuracy was compared between oncologists who predicted survival on referral to hospice, and PC physicians who independently predicted survival upon admission to hospice (44). One hundred and one patients were enrolled between 2014 and 2017. Overall mean survival time was 28.4 days (range, 1–167 days). The difference between CPS and AS by oncologists and PC physicians was 31.2 days (95% CI: 26.7–35.6) and 15.6 days (95% CI: 12.4–18.8), respectively (P<0.0001). Oncologists were accurate 21.8% of the time, underpredicted in 10.9% and overpredicted in 67.3%. PC physicians were accurate in 37.6%, underpredicted in 11.9% and overpredicted in 50.5% (44). One potential explanation for this difference was that PC physicians’ predictions were assisted by prognostic tools (Palliative Prognostic Score and Palliative Prognostic Index) while oncologists’ predictions were not (44). The authors also suggested that oncologists are not as accurate since they do not follow patients through end-of-life to death; rather patients are often transferred to acute care settings or hospice (44).

In our study, PC physicians were correct most often among all HCP and had the least degree of overprediction. They were also more often correct than a previous report, in which they were accurate 35% of the time, underpredicted in 20% and overpredicted 45% of the time (with accuracy defined as CPS within 30% of AS) (11). Their proficiency may relate to their specific roles at the end of life; for example, they are required to assess when patients meet survival-based criteria for hospice admission, discuss goals of care, and when to most appropriately stop treatment. We noted that PC physicians in this study used a distinctive pattern of factors in defining their prediction. They were the only group that listed weight loss as an important factor in addition to PS, disease extent and lab values. This clinic has a history of publication on the prognostic value of weight loss and other nutritional assessments (46), which may have increased their reliance on such information. The Palliative Performance Scale is also used in standard care in this clinic, and this tool is well known to be a significant predictive factor in palliative patients (47,48). Interestingly, whether a prediction was accurate or inaccurate, participating assessors used the same factors in determining CPS. This echoes a previous CPS study of Radiation Oncologists who also submitted factors on which they based their estimate; no statistically significant correlation was identified between those factors and accuracy (43).

A 2022 study asked nurses and physicians to determine whether patients would survive longer or shorter than the duration estimated by a prediction model previously validated by the same group (38). Seventy-three patients with metastatic cancers were enrolled [2016–2017] with an AS of 9.2 months (range, 0.2–60.4 months). The accuracy of nursing predictions was 61.6% compared to 60.3% for physicians (P=0.85). Among incorrect predictions, 56.1% were overestimated and 43.9% underestimated, but data by profession was not reported (38). From the physician’s perspective, common reasons for inaccuracy included uncertainty regarding whether patients were receiving systemic therapy, the variability of response to systemic therapy, and the timing of initiating hospice (38). Nurses reported holistically evaluating the patient rather than considering particular data elements to inform their predictions (38). Additionally, once a patient is imminently dying (prognosis of less than 48 hours), nursing predictions have been reported as more accurate than physicians (18).

The few published studies incorporating residents’ CPS report inconsistent results. Christakis et al. found that the accuracy of CPS is influenced by clinician (lack of) knowledge and training status (39). Twomey et al., however, reported that junior doctors were more accurate than consultants (49). PC residents comprised approximately half of the physicians from Taiwan in an international study on prognostication in patients admitted to PC units (45). Accuracy of resident predictions was not separately reported although the authors outlined that all received supervision from attending physicians (45); as such, an impact of that collaboration on CPS accuracy could not be excluded. In another study set in a PC inpatient unit, CPS accuracy of attending physicians, residents and registered nurses was assessed (50). Prognosis was recorded for 115 patients within 3 days of admission and then repeated weekly until discharge or death (50). Average AS was 11.5±12 days. There were no significant differences in predicted survival or rate of optimistic, accurate or pessimistic predictions between different staff (50).

We hypothesized that some of factors used to formulate predictions would take minimal clinical experience to evaluate, such as age, and therefore be equivalently assessed by trainees, residents, physicians and non-physicians. We were also curious as to whether factors important in evaluating survival could be more expertly evaluated by non-physicians. For example, an experienced hospital-based registered dietician can more accurately assess oral intake than Oncologists or PC physicians. Indeed, both age and oral intake have been significantly associated with accuracy in a recent report (51).

Given that factors impacting survival may in fact not require extensive clinical knowledge to assess, we encouraged trainee participation. While the number of predictions made by trainees was relatively small, these typically were less likely to be correct and more likely to overpredict compared with specialist physicians. The exception was the residents in PC, who were not significantly different from PC consultants. However, this was the smallest assessor group, and the wide confidence interval suggests limited reliability. No clear guidance exists on how trainees can be taught to perform this task (15). Trainee prognostication was recently identified specifically as an area for future research (41).

Notwithstanding the fact that the members of the MDT are working collaboratively, our results suggest that for a given patient, different HCPs are likely formulating treatment recommendations based on widely disparate survival predictions. This underpins a need for dynamic, collaborative patient-centered team communication. There was a trend across all disciplines to overpredict more frequently than underpredict. CPS that is too optimistic can result in delayed referral to primary or specialist PC, or offering interventions with a higher risk of adverse events than would otherwise be preferable. A clearly communicated and accurate CPS means clinicians can have more realistic discussions about what interventions are likely to be beneficial (5,8,52,53). Most patients wish to stay at home at the end of life (54), and prognostication can help avoid unnecessary hospital admissions and medical care that is inappropriately aggressive (5).

One method utilized to improve clinicians’ ability to prognosticate is the incorporation of objective factors, such as lab results (46,55). Interestingly, a 2011 survey of Canadian PC physicians showed that most respondents both disagreed with the idea that they overpredict and reported not routinely using clinical tools to help prognosticate (56).

Our results must be interpreted in the context of both the strengths and limitations of the study. Strengths include the large cohort of consecutive real-world patients utilizing the entire workflow of palliative and supportive services at a tertiary cancer center. Predictions were made by experienced team members providing routine clinical care. Two different published methods of determining accuracy were utilized. The sample size was sufficient to reveal clear differences in the prognostication skills among most MDT disciplines; however, numbers of some allied HCPs and trainees were limited. Demographic and personal data were not collected on participating assessors to preserve anonymity, so it is not possible to comment on individual clinicians’ characteristics. Survival has generally improved for patients with advanced cancer since these data were collected and the sample may not be entirely representative of current outcomes.

Conclusions

While HCPs experienced in providing supportive cancer care vary in their accuracy in predicting survival, all have a propensity to overpredict. ESMO Clinical Practice Guidelines on prognostication in patients in the last month of life recommend that clinicians consider input from multiple professionals to supplement their own CPS (2). Further research is needed to determine how individuals working within multi-disciplinary teams can incorporate their specialized skills and experience to prognosticate in a collaborative manner to optimally support patient-centered care.

Acknowledgments

The authors would like to gratefully acknowledge all Palliative Care and Symptom Control and Palliative Radiation Oncology program team members at the Cross Cancer Institute for their participation. No AI tools were used in the collection or analysis of data, the production of images, or in the writing of the manuscript. This study was presented in part at the Multinational Association for Supportive Care in Cancer annual meeting in June 2024 and the Canadian Society of Palliative Medicine annual meeting in September 2024.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://apm.amegroups.com/article/view/10.21037/apm-25-108/rc

Data Sharing Statement: Available at https://apm.amegroups.com/article/view/10.21037/apm-25-108/dss

Peer Review File: Available at https://apm.amegroups.com/article/view/10.21037/apm-25-108/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://apm.amegroups.com/article/view/10.21037/apm-25-108/coif). V.E.B. has received grants from the Canadian Institutes of Health Research and the Alberta Cancer Foundation, and has consulted for Pfizer and Nestle Health Sciences, unrelated to the current study. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the provincial Heath Research Ethics Board of Alberta – Cancer Committee, file HREBA.CC-17-0437, and individual consent for this analysis was waived.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Hui D, Mo L, Paiva CE. The Importance of Prognostication: Impact of Prognostic Predictions, Disclosures, Awareness, and Acceptance on Patient Outcomes. Curr Treat Options Oncol 2021;22:12. [Crossref] [PubMed]
2. Stone P, Buckle P, Dolan R, et al. Prognostic evaluation in patients with advanced cancer in the last months of life: ESMO Clinical Practice Guideline. ESMO Open 2023;8:101195. [Crossref] [PubMed]
3. Glare P, Sinclair C, Downing M, et al. Predicting survival in patients with advanced disease. Eur J Cancer 2008;44:1146-56. [Crossref] [PubMed]
4. Weeks JC, Cook EF, O'Day SJ, et al. Relationship between cancer patients' predictions of prognosis and their treatment preferences. JAMA 1998;279:1709-14. [Crossref] [PubMed]
5. Oosterveld-Vlug MG, Heins MJ, Boddaert MSA, et al. Evaluating quality of care at the end of life and setting best practice performance standards: a population-based observational study using linked routinely collected administrative databases. BMC Palliat Care 2022;21:51. [Crossref] [PubMed]
6. Hui D, Paiva CE, Del Fabbro EG, et al. Prognostication in advanced cancer: update and directions for future research. Support Care Cancer 2019;27:1973-84. [Crossref] [PubMed]
7. LeBlanc TW, Marron JM, Ganai S, et al. Prognostication and Communication in Oncology. J Oncol Pract 2019;15:208-15. [Crossref] [PubMed]
8. Mackillop WJ. The importance of prognosis in cancer medicine. In: Sobin LH, Gospodarowicz MK, O'Sullivan B, et al. editors. TNM Online. Hoboken: Wiley; 2014.
9. Steinhauser KE, Christakis NA, Clipp EC, et al. Factors considered important at the end of life by patients, family, physicians, and other care providers. JAMA 2000;284:2476-82. [Crossref] [PubMed]
10. Razvi Y, Chan S, Zhang L, et al. Are we better a decade later in the accuracy of survival prediction by palliative radiation oncologists? Ann Palliat Med 2019;8:150-8. [Crossref] [PubMed]
11. Amano K, Maeda I, Shimoyama S, et al. The Accuracy of Physicians' Clinical Predictions of Survival in Patients With Advanced Cancer. J Pain Symptom Manage 2015;50:139-46.e1. [Crossref] [PubMed]
12. Christakis NA, Iwashyna TJ. Attitude and self-reported practice regarding prognostication in a national sample of internists. Arch Intern Med 1998;158:2389-95. [Crossref] [PubMed]
13. Mandelli S, Riva E, Tettamanti M, et al. How palliative care professionals deal with predicting life expectancy at the end of life: predictors and accuracy. Support Care Cancer 2021;29:2093-103. [Crossref] [PubMed]
14. Tavares T, Oliveira M, Gonçalves J, et al. Predicting prognosis in patients with advanced cancer: A prospective study. Palliat Med 2018;32:413-6. [Crossref] [PubMed]
15. White N, Reid F, Harris A, et al. A Systematic Review of Predictions of Survival in Palliative Care: How Accurate Are Clinicians and Who Are the Experts? PLoS One 2016;11:e0161407. [Crossref] [PubMed]
16. Gwilliam B, Keeley V, Todd C, et al. Prognosticating in patients with advanced cancer—observational study comparing the accuracy of clinicians’and patients’ estimates of survival. Ann Oncol 2013;24:482-488. [Crossref] [PubMed]
17. Kiely BE, Martin AJ, Tattersall MH, et al. The median informs the message: accuracy of individualized scenarios for survival time based on oncologists' estimates. J Clin Oncol 2013;31:3565-71. [Crossref] [PubMed]
18. 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;16:1642-8. [Crossref] [PubMed]
19. Maltoni M, Caraceni A, Brunelli C, et al. Prognostic factors in advanced cancer patients: evidence-based clinical recommendations--a study by the Steering Committee of the European Association for Palliative Care. J Clin Oncol 2005;23:6240-8. [Crossref] [PubMed]
20. Glare P, Virik K, Jones M, et al. A systematic review of physicians' survival predictions in terminally ill cancer patients. BMJ 2003;327:195-8. [Crossref] [PubMed]
21. Lamont EB, Christakis NA. Complexities in prognostication in advanced cancer: "to help them live their lives the way they want to". JAMA 2003;290:98-104. [Crossref] [PubMed]
22. Higginson IJ, Costantini M. Accuracy of prognosis estimates by four palliative care teams: a prospective cohort study. BMC Palliat Care 2002;1:1. [Crossref] [PubMed]
23. Llobera J, Esteva M, Rifa J, et al. Terminal cancer. duration and prediction of survival time. Eur J Cancer 2000;36:2036-43. [Crossref] [PubMed]
24. Mackillop WJ, Quirt CF. Measuring the accuracy of prognostic judgments in oncology. J Clin Epidemiol 1997;50:21-9. [Crossref] [PubMed]
25. Bruera E, Miller MJ, Kuehn N, et al. Estimate of survival of patients admitted to a palliative care unit: a prospective study. J Pain Symptom Manage 1992;7:82-6. [Crossref] [PubMed]
26. Laird BJ, Kaasa S, McMillan DC, et al. Prognostic factors in patients with advanced cancer: a comparison of clinicopathological factors and the development of an inflammation-based prognostic system. Clin Cancer Res 2013;19:5456-64. [Crossref] [PubMed]
27. Glare P. Prognostic Factors in Terminal Cancer. In: TNM Online. Hoboken: Wiley; 2006. Available online: https://onlinelibrary.wiley.com/doi/abs/10.1002/0471463736.tnmp48
28. Maltoni M, Pirovano M, Scarpi E, et al. Prediction of survival of patients terminally ill with cancer. Results of an Italian prospective multicentric study. Cancer 1995;75:2613-22. [Crossref] [PubMed]
29. Orlovic M, Droney J, Vickerstaff V, et al. Accuracy of clinical predictions of prognosis at the end-of-life: evidence from routinely collected data in urgent care records. BMC Palliat Care 2023;22:51. [Crossref] [PubMed]
30. Scott B. Multidisciplinary team approach in cancer care: a review of the latest advancements. EMJ Oncol 2021;9:2-13.
31. Blackwood O, Deb R. Multidisciplinary team approach in breast cancer care: Benefits and challenges. Indian J Pathol Microbiol 2020;63:S105-12. [Crossref] [PubMed]
32. Fairchild A, Debenham B, Danielson B, et al. Comparative multidisciplinary prediction of survival in patients with advanced cancer. Support Care Cancer 2014;22:611-7. [Crossref] [PubMed]
33. Mendis R, Soo WK, Zannino D, et al. Multidisciplinary Prognostication Using the Palliative Prognostic Score in an Australian Cancer Center. Palliat Care 2015;9:7-14. [Crossref] [PubMed]
34. Watanabe SM, Faily V, Mawani A, et al. Frequency, Timing, and Predictors of Palliative Care Consultation in Patients with Advanced Cancer at a Tertiary Cancer Center: Secondary Analysis of Routinely Collected Health Data. Oncologist 2020;25:722-8. [Crossref] [PubMed]
35. Yoong SQ, Bhowmik P, Kapparath S, et al. Palliative prognostic scores for survival prediction of cancer patients: a systematic review and meta-analysis. J Natl Cancer Inst 2024;116:829-57. [Crossref] [PubMed]
36. Preto DD, Pantano NP, Paiva BSR, et al. Clinical Prediction of Survival in Advanced Cancer Outpatients: Do Experienced Physicians and With Prior Patient Evaluation Make More Accurate Predictions? J Pain Symptom Manage 2020;59:e7-e10. [Crossref] [PubMed]
37. Maltoni M, Nanni O, Demi S, et al. Clinical prediction of survival is more accurate than the Karnofsky performance status in estimating life span of terminally ill cancer patients. Eur J Cancer 1994;30A:764-6. [Crossref] [PubMed]
38. Kao J, Zucker A, Urso M, et al. Improving Survival Prognostication in Patients With Metastatic Cancer Through Clinical Judgment. Anticancer Res 2022;42:1397-401. [Crossref] [PubMed]
39. Christakis NA, Lamont EB. Extent and determinants of error in doctors' prognoses in terminally ill patients: prospective cohort study. BMJ 2000;320:469-72. [Crossref] [PubMed]
40. Oxenham D, Cornbleet MA. Accuracy of prediction of survival by different professional groups in a hospice. Palliat Med 1998;12:117-8. [Crossref] [PubMed]
41. Surendran S, Kuriakose J, Ponnamthodi P, et al. Temporal, Probabilistic and Surprise Question Approaches of Clinician Predicted Survival in Advanced Hepatopancreaticobiliary Cancers: A Prospective Cohort Study. Indian J Palliat Care 2025;31:241-9. [Crossref] [PubMed]
42. Hauser CA, Stockler MR, Tattersall MH. Prognostic factors in patients with recently diagnosed incurable cancer: a systematic review. Support Care Cancer 2006;14:999-1011. [Crossref] [PubMed]
43. Benson KRK, Aggarwal S, Carter JN, et al. Predicting Survival for Patients With Metastatic Disease. Int J Radiat Oncol Biol Phys 2020;106:52-60. [Crossref] [PubMed]
44. Urahama N, Sono J, Yoshinaga K. Comparison of the accuracy and characteristics of the prognostic prediction of survival of identical terminally ill cancer patients by oncologists and palliative care physicians. Jpn J Clin Oncol 2018;48:695-8. [Crossref] [PubMed]
45. Lee ES, Hiratsuka Y, Suh SY, et al. Clinicians' Prediction of Survival and Prognostic Confidence in Patients with Advanced Cancer in Three East Asian Countries. J Palliat Med 2023;26:790-7. [Crossref] [PubMed]
46. Tarumi Y, Watanabe SM, Lau F, et al. Evaluation of the Palliative Prognostic Score (PaP) and routinely collected clinical data in prognostication of survival for patients referred to a palliative care consultation service in an acute care hospital. J Pain Symptom Manage 2011;42:419-31. [Crossref] [PubMed]
47. Jang RW, Caraiscos VB, Swami N, et al. Simple prognostic model for patients with advanced cancer based on performance status. J Oncol Pract 2014;10:e335-41. [Crossref] [PubMed]
48. Casarett DJ, Farrington S, Craig T, et al. The art versus science of predicting prognosis: can a prognostic index predict short-term mortality better than experienced nurses do? J Palliat Med 2012;15:703-8. [Crossref] [PubMed]
49. Twomey F, O'Leary N, O'Brien T. Prediction of patient survival by healthcare professionals in a specialist palliative care inpatient unit: a prospective study. Am J Hosp Palliat Care 2008;25:139-45. [Crossref] [PubMed]
50. Chen PY, Huang CH, Peng JK, et al. Prediction Accuracy Between Terminally Ill Patients' Survival Length and the Estimations Made From Different Medical Staff, a Prospective Cohort Study. Am J Hosp Palliat Care 2023;40:440-6. [Crossref] [PubMed]
51. Hiratsuka Y, Suh S, Yoon S, et al. Factors related to accurate clinicians’ prediction of survival: An international multicenter study in East Asia. Support Care Cancer 2024;32:490. [Crossref] [PubMed]
52. Li D, Prigerson HG, Kang J, et al. Impact of Radiation Therapy on Aggressive Care and Quality of Life Near Death. J Pain Symptom Manage 2017;53:25-32. [Crossref] [PubMed]
53. Pearlman RA I. Inaccurate predictions of life expectancy. Dilemmas and opportunities. Arch Intern Med 1988;148:2537-8. [Crossref] [PubMed]
54. Cai J, Zhang L, Guerriere D, et al. Where Do Cancer Patients in Receipt of Home-Based Palliative Care Prefer to Die and What Are the Determinants of a Preference for a Home Death? Int J Environ Res Public Health 2020;18:235. [Crossref] [PubMed]
55. Tam A, Scarpi E, Maltoni MC, et al. A Systematic Review of Prognostic Factors in Patients with Cancer Receiving Palliative Radiotherapy: Evidence-Based Recommendations. Cancers (Basel) 2024;16:1654. [Crossref] [PubMed]
56. Corkum M, Viola R, Veenema C, et al. Prognosticating in palliative care: a survey of Canadian palliative care physicians. J Palliat Care 2011;27:89-97. [Crossref] [PubMed]