PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of oncologistAlphamed PressThe OncologistContact UsCMESubscriptionsSubmissionsAboutCurrent Issue
 
Oncologist. 2015 July; 20(7): 839–844.
Published online 2015 June 8. doi:  10.1634/theoncologist.2015-0015
PMCID: PMC4492240

Surprise Questions for Survival Prediction in Patients With Advanced Cancer: A Multicenter Prospective Cohort Study

Abstract

Background.

Predicting the short-term survival in cancer patients is an important issue for patients, family, and oncologists. Although the prognostic accuracy of the surprise question has value in 1-year mortality for cancer patients, the prognostic value for short-term survival has not been formally assessed. The primary aim of the present study was to assess the prognostic value of the surprise question for 7-day and 30-day survival in patients with advanced cancer.

Patients and Methods.

The present multicenter prospective cohort study was conducted in Japan from September 2012 through April 2014, involving 16 palliative care units, 19 hospital-based palliative care teams, and 23 home-based palliative care services.

Results.

We recruited 2,425 patients and included 2,361 for analysis: 912 from hospital-based palliative care teams, 895 from hospital palliative care units, and 554 from home-based palliative care services. The sensitivity, specificity, positive predictive value, and negative predictive value of the 7-day survival surprise question were 84.7% (95% confidence interval [CI], 80.7%–88.0%), 68.0% (95% CI, 67.3%–68.5%), 30.3% (95% CI, 28.9%–31.5%), and 96.4% (95% CI, 95.5%–97.2%), respectively. The sensitivity, specificity, positive predictive value, and negative predictive value for the 30-day surprise question were 95.6% (95% CI, 94.4%–96.6%), 37.0% (95% CI, 35.9%–37.9%), 57.6% (95% CI, 56.8%–58.2%), and 90.4% (95% CI, 87.7%–92.6%), respectively.

Conclusion.

Surprise questions are useful for screening patients for short survival. However, the high false-positive rates do not allow clinicians to provide definitive prognosis prediction.

Implications for Practice:

The findings of this study indicate that clinicians can screen patients for 7- or 30-day survival using surprise questions with 90% or more sensitivity. Clinicians cannot provide accurate prognosis estimation, and all patients will not always die within the defined periods. The screened patients can be regarded as the subjects to be prepared for approaching death, and proactive discussion would be useful for such patients.

Keywords: Surprise questions, Prognostic value, 7-day survival, 30-day survival, Advanced cancer

Introduction

Treatment and end-of-life care decisions are often difficult for cancer patients, and both are influenced by the patient’s estimated prognosis [13]. Therefore, predicting the prognosis for cancer patients is an important issue for patients, family, and oncologists [47]. The surprise question, “Would I be surprised if this patient died in the next year?” is an innovative tool to improve end-of-life care [8], and the prognostic value for cancer patients has been shown in several empirical studies [9, 10]. Historically, the surprise question has been used in the National Gold Standards Framework as a trigger to identify those needing proactive palliative care interventions. The surprise question was then incorporated into the Gold Standards Framework Prognostic Indicator Guidance guidelines as a screening tool for patients likely to have a short life expectancy [11, 12].

Most cancer patients deteriorate rapidly in their last month or week [13], and several prognostic prediction tools for short-term survival have been validated, including the Palliative Prognostic Index (PPI) [14], Palliative Prognostic Score (PaP score) [15], PaP Score with Delirium [16], and the Prognosis in Palliative Care Study (PiPS) predictor model [17]. However, although the prognostic accuracy of the surprise question has value in determining the 1-year mortality for cancer patients [911], the prognostic value for short-term survival (days to weeks) has not been formally assessed. If the surprise question for 7- and 30-day survival prediction has accurate prognostic value, this method would be a simple and precise alternative to the PPI, PaP score, and PiPS model. Our aim was to assess the prognostic value of the surprise question for 7-day survival and 30-day survival in cancer patients.

Materials and Methods

The present study was a part of a larger study comparing the prognostic values of the PaP score and the PPI and PiPS models in hospital-based palliative care teams, palliative care units, and home-based palliative care services [18]. This multicenter prospective cohort study was conducted in 58 palliative care services in Japan from September 2012 through April 2014. The participating units included 16 palliative care units, 19 hospital-based palliative care teams, and 23 home-based palliative care services. After a brief instruction workshop, the palliative care physicians evaluated the patients and recorded all outcome measures on the first day of admission and followed up all patients until death or 6 months after enrollment. The present study was conducted in accordance with the ethical standards of the Declaration of Helsinki and the ethical guidelines for epidemiological research presented by the Ministry of Health, Labor and Welfare in Japan. The local institutional review boards of all participating institutions approved the present study.

Patients

Eligible patients were consecutively enrolled in the study if they had been newly referred to the participating institutions during the study period. All institutions were asked to evaluate and collect data for a designated number of patients (20, 40, 60, 80, or 100) according to the size of the institution. The inclusion criteria were age >20 years old, a diagnosis of locally extensive or metastatic cancer (including hematological neoplasms), and admission to a palliative care unit, receipt of care by hospital-based palliative care teams, or receipt of home-based palliative care services.

Measurements

We collected the palliative care physicians’ responses to the surprise question, “Would I be surprised if this patient died in the next 7 days?” (7-day surprise question) and “Would I be surprised if this patient died in the next 30 days?” (30-day surprise question). The response was categorized as yes or no. We also recorded the participants’ demographic and clinical characteristics, including age, sex, site of the primary cancer and metastatic disease, and anticancer treatments (i.e., chemotherapy, hormonal therapy, and radiotherapy).

Statistical Analysis

The data were divided into two independent groups according to the response to each surprise question (“yes, surprised”/“not surprised”). The Kaplan-Meier method was used to generate survival curves in survival days for each group, and the log-rank test was used to compare the survival time between the groups. To assess the prognostic value of each surprise question, we calculated the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy in each of the three settings for all patients. All analyses were conducted using SPSS-J, version 22.0 (IBM Corp., Tokyo, Japan, http://www-01.ibm.com/software/analytics/spss/).

Results

We recruited 2,425 patients for the present study: 964 from the hospital-based palliative care teams, 903 from the hospital palliative care units, and 558 from the home-based palliative care services. Of these, 62 and 2 patients were excluded because of a lack of follow-up data and missing responses to the surprise questions, respectively. We thus included 2,361 patients in the final analysis: 912 from the hospital-based palliative care teams, 895 from the hospital palliative care units, and 554 from the home-based palliative care services. The patient characteristics are summarized in Table 1. Their mean age was 69.1 years, and the digestive organs were the most frequent site of primary cancer, followed by respiratory/intrathoracic cancer.

Table 1.
Patient characteristics

The numbers of patients with the response “not surprised” to the surprise questions were 931 (39.4%) for 7-day survival and 1,852 (78.4%) for 30-day survival (Table 2). The Kaplan-Meier survival curves for both the 7-day and 30-day surprise questions showed that the survival rates differed between the “not surprised” group and the “yes, surprised” group in all palliative care setting (Fig. 1; p < .001). The survival rates also differed in each of the 3 settings (Fig. 1; p < .001).

Table 2.
Response to the surprise question and survival outcomes
Figure 1.
Kaplan-Meier survival curves for 7-day and 30-day surprise questions. Group A is the “yes, surprised” group (blue line), and group B is the “not surprised” group (green line).

The sensitivity, specificity, PPV, NPV, and accuracy of the 7-day and 30-day surprise questions are listed in Table 3. The sensitivity, specificity, PPV, and NPV for the 7-day surprise question were 84.7%, 68.0%, 30.3%, and 96.4%, respectively. The sensitivity, specificity, PPV, and NPV for the 30-day surprise question were 95.6%, 37.0%, 57.6%, and 90.4%, respectively. In the different settings, the sensitivity, specificity, PPV, and NPV of the 7-day surprise question ranged from 82.3% to 92.6%, 59.2% to 79.1%, 19.7% to 34.3%, and 93.5% to 98.7%, respectively. The sensitivity, specificity, PPV, and NPV of the 30-day surprise question in each setting ranged from 90.4% to 99.2%, 19.7% to 55.9%, 50.0% to 62.6%, and 85.1% to 97.4%, respectively.

Table 3.
Prognostic value of the surprise question for 7-day and 30-day survival

Discussion

Findings

To the best of our knowledge, ours is the first study to assess the prognostic value of the 7-day and 30-day surprise question for survival in patients with advanced cancer. One of the most important findings of the present study was that both surprise questions had high sensitivity, with values of more than 90% (30-day survival) and more than 80% (7-day survival). These percentages are higher than those in previous studies using existing prognostic tools. The sensitivity of the PPI for the 3-week prognosis ranged from 74% to 83% [14], and the sensitivity of the PaP scores for the 30-day prognosis ranged from 73% to 92% [19]. The results were similar in all three settings: hospital-based palliative care teams, palliative care units, and home-based palliative care services. These results indicate that surprise questions are useful screening tools to identify patients with shorter survival.

Another important finding of the present study was that the specificity of both surprise questions was relatively low, especially for 30-day survival. This is contrast to the findings of previous studies that demonstrated that surprise questions for a 1-year prognosis prediction had adequate specificity of 70%–90% [9, 10]. One potential interpretation of the low specificity in the present study is that determining which patients will die within a 7-day or 30-day period is much more difficult than determining which patients will die within 1 year. A recent study revealed that the physical signs of impending death had very high specificity for death within days but low sensitivity [20]. To improve the prediction methods for short time intervals, a combination algorism of high sensitivity methods (e.g., surprise questions) and high specificity methods (e.g., specific physical signs) might be promising for testing in future studies. Screening questions alone cannot be used to provide definitive prognosis predictions because of the low specificity and high false-positive rates.

Of interest is that differences in the question format might result in a different predictive profile. A recent study indicated that the probabilistic approach (i.e., what is the approximate probability this patient will be alive, 0%–100%?) was more accurate than the temporal approach (i.e., what is the approximate survival time for this patient?) [21]. Surprise questions are a conceptually unique question format compared with existing tools. Future studies exploring how the difference in question format influences the predictive profile in clinical settings are necessary.

Strengths and Limitations

The strength of the present study was the large number of participants enrolled in three palliative settings, including palliative care units, hospital-based palliative care teams, and home-based palliative care services. One limitation was that 62 of 2,361 patients (2.6%) could not be followed up or data on their date of death were lost. However, we believe that this proportion of patients was sufficiently small to have had no effect on our conclusions. All assessments were performed by palliative care physicians, and ratings by generalists or oncologists might generate different findings.

Conclusion

Surprise questions can be a useful tool to screen patients with short survival; however, the high false-positive rates do not allow clinicians to provide a definitive prognosis.

This article is available for continuing medical education credit at CME.TheOncologist.com.

Acknowledgments

This work was supported in part by the National Cancer Center Research and Development Fund (Grant 25-A-22). All researchers were independent from the funders. This study was performed in the Japan ProVal Study Group (Prognostic scores Validation Study Group). The participating study sites and site investigators were as follows: Satoshi Inoue (Seirei Hospice, Seirei Mikatahara General Hospital), Masayuki Ikenaga (Hospice Children’s Hospice Hospital, Yodogawa Christian Hospital), Yoshihisa Matsumoto (Department of Palliative Medicine, National Cancer Center Hospital East), Mika Baba (Department of Palliative Care, Saito Yukoukai Hospital), Ryuichi Sekine (Department of Pain and Palliative Care, Kameda Medical Center), Takashi Yamaguchi (Department of Palliative Medicine, Kobe University Graduate School of Medicine), Takeshi Hirohashi (Department of Palliative Care, Mitui Memorial Hospital), Tsukasa Tajima (Department of Palliative Medicine, Tohoku University Hospital), Ryohei Tatara (Department of Palliative Medicine, Osaka City General Hospital), Hiroaki Watanabe (Komaki City Hospital), Hiroyuki Otani (Department of Palliative Care Team, and Palliative and Supportive Care, National Kyushu Cancer Center), Chizuko Takigawa (Department of Palliative Care, KKR Sapporo Medical Center), Yoshinobu Matsuda (Department of Psychosomatic Medicine, National Hospital), Hiroka Nagaoka (Center for Palliative and Supportive Care, Tsukuba University Hospital), Masanori Mori (Seirei Hamamatsu General Hospital), Yo Tei (Seirei Hospice, Seirei Mikatahara General Hospital), Shuji Hiramoto (Department of Oncology, Mitsubishi Kyoto Hospital), Akihiko Suga (Department of Palliative Medicine, Shizuoka Saiseikai General Hospital), Takayuki Hisanaga (Tsukuba Medical Center Foundation), Tatsuhiko Ishihara (Palliative Care Department, Okayama Saiseikai General Hospital), Tomoyuki Iwashita (Matsue City Hospital), Keisuke Kaneishi (Department of Palliative Care Unit, JCHO Tokyo Shinjuku Medical Center), Shohei Kawagoe (Aozora Clinic), Toshiyuki Kuriyama (Department of Palliative Medicine, Wakayama Medical University Hospital Oncology Center), Takashi Maeda (Department of Palliative Care, Tokyo Metropolitan Cancer, and Infectious Disease Center, Komagome Hospital), Ichiro Mori (Gratia Hospital Hospice), Nobuhisa Nakajima (Department of Palliative Medicine, Graduate School of Medicine, Tohoku University), Tomohiro Nishi (Kawasaki Comprehensive Care Center, Kawasaki Municipal Ida Hospital), Hiroki Sakurai (Department of Palliative Care, St. Luke's International Hospital, Tokyo), Satofumi Shimoyama (Department of Palliative Care, Aichi Cancer Center Hospital), Takuya Shinjo (Shinjo Clinic, Kobe), Hiroto Shirayama (Iryouhoujinn Takumikai Osaka Kita Homecare Clinic), Takeshi Yamada (Department of Gastrointestinal and Hepato-Billiary-Pancreatic Surgery, Nippon Medical School), Shigeki Ono (Division of Palliative Medicine, Shizuoka Cancer Center Hospital), Taketoshi Ozawa (Megumi Zaitaku Clinic), Ryo Yamamoto (Department of Palliative Medicine, Saku Central Hospital Advanced Care Center), Naoki Yamamoto (Department of Primary Care Service, Shinsei Hospital), Hideki Shishido (Shishido Internal Medicine Clinic), Mie Shimizu (Saiseikai Matsusaka General Hospital), Masanori Kawahara (Soshukai Okabe Clinic), Shigeru Aoki (Sakanoue Family Clinic), Akira Demizu (Demizu Clinic), Masahiro Goshima (Homecare-Clinic Kobe), Keiji Goto (Himawari Zaitaku Clinic), Yasuaki Gyoda (Kanamecyo Home Care clinic), Jun Hamano (Division of Clinical Medicine, Faculty of Medicine, University of Tsukuba), Kotaro Hashimoto (Fukushima Home Palliative Care Clinic), Sen Otomo (Shonan International Village Clinic), Masako Sekimoto (Sekimoto Clinic), Takemi Shibata (Kanwakeakurinikku-Eniwa), Yuka Sugimoto (Sugimoto Homecare Clinic), Mikako Matsunaga (Senri Pain Clinic), Yukihiko Takeda (Hidamari Clinic), Takeshi Sasara (Yuuaikai Nanbu Hospital), Jun Nagayama (Peace Clinic Nakai).

Footnotes

For Further Reading: David Hui, Kelly Kilgore, Linh Nguyen et al. The Accuracy of Probabilistic Versus Temporal Clinician Prediction of Survival for Patients with Advanced Cancer: A Preliminary Report. The Oncologist 2011;16:1642–1648.

Abstract:

Clinicians have limited accuracy in the prediction of patient survival. We assessed the accuracy of probabilistic clinician prediction of survival (CPS) and temporal CPS for advanced cancer patients admitted to our acute palliative care unit, and identified factors associated with CPS accuracy. Eight physicians and 20 nurses provided their estimation of survival on admission by (a) the temporal approach, “What is the approximate survival for this patient (in days)?” and (b) the probabilistic approach, “What is the approximate probability that this patient will be alive (0%–100%)?” for ≥24 hours, 48 hours, 1 week, 2 weeks, 1 month, 3 months, and 6 months. We also collected patient and clinician demographics. Among 151 patients, the median age was 58 years, 95 (63%) were female, and 138 (81%) had solid tumors. The median overall survival time was 12 days. The median temporal CPS was 14 days for physicians and 20 days for nurses. Physicians were more accurate than nurses. A higher accuracy of temporal physician CPS was associated with older patient age. Probabilistic CPS was significantly more accurate than temporal CPS for both physicians and nurses, although this analysis was limited by the different criteria for determining accuracy. With the probabilistic approach, nurses were significantly more accurate at predicting survival at 24 hours and 48 hours, whereas physicians were significantly more accurate at predicting survival at 6 months. The probabilistic approach was associated with high accuracy and has practical implications.

Author Contributions

Conception/Design: Jun Hamano, Tatsuya Morita

Provision of study material or patients: Jun Hamano, Tatsuya Morita

Collection and/or assembly of data: Jun Hamano, Tatsuya Morita, Satoshi Inoue, Masayuki Ikenaga, Yoshihisa Matsumoto, Ryuichi Sekine, Takashi Yamaguchi, Takeshi Hirohashi, Tsukasa Tajima, Ryohei Tatara, Hiroaki Watanabe, Hiroyuki Otani, Chizuko Takigawa, Yoshinobu Matsuda, Hiroka Nagaoka, Masanori Mori, Naoki Yamamoto, Mie Shimizu, Takeshi Sasara, Hiroya Kinoshita

Data analysis and interpretation: Jun Hamano, Tatsuya Morita

Manuscript writing: Jun Hamano, Tatsuya Morita

Final approval of manuscript: Jun Hamano, Tatsuya Morita, Satoshi Inoue, Masayuki Ikenaga, Yoshihisa Matsumoto, Ryuichi Sekine, Takashi Yamaguchi, Takeshi Hirohashi, Tsukasa Tajima, Ryohei Tatara, Hiroaki Watanabe, Hiroyuki Otani, Chizuko Takigawa, Yoshinobu Matsuda, Hiroka Nagaoka, Masanori Mori, Naoki Yamamoto, Mie Shimizu, Takeshi Sasara, Hiroya Kinoshita

Disclosures

Hiroya Kinoshita: The National Cancer Center Research and Development Fund (Grant 25-A-22) (RF). The other authors indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board

References

1. Weeks JC, Cook EF, O’Day SJ, et al. Relationship between cancer patients’ predictions of prognosis and their treatment preferences. JAMA. 1998;279:1709–1714. [PubMed]
2. Matsuyama R, Reddy S, Smith TJ. Why do patients choose chemotherapy near the end of life? A review of the perspective of those facing death from cancer. J Clin Oncol. 2006;24:3490–3496. [PubMed]
3. Finlay E, Casarett D. Making difficult discussions easier: Using prognosis to facilitate transitions to hospice. CA Cancer J Clin. 2009;59:250–263. [PubMed]
4. 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;284:2476–2482. [PubMed]
5. Kutner JS, Steiner JF, Corbett KK, et al. Information needs in terminal illness. Soc Sci Med. 1999;48:1341–1352. [PubMed]
6. Kirk P, Kirk I, Kristjanson LJ. What do patients receiving palliative care for cancer and their families want to be told? A Canadian and Australian qualitative study. BMJ. 2004;328:1343. [PMC free article] [PubMed]
7. 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;23:6240–6248. [PubMed]
8. Pattison M, Romer AL. Improving care through the end of life: Launching a primary care clinic-based program. J Palliat Med. 2001;4:249–254. [PubMed]
9. Moss AH, Lunney JR, Culp S, et al. Prognostic significance of the “surprise” question in cancer patients. J Palliat Med. 2010;13:837–840. [PubMed]
10. Moroni M, Zocchi D, Bolognesi D, et al. The ‘surprise’ question in advanced cancer patients: A prospective study among general practitioners. Palliat Med. 2014;28:959–964. [PubMed]
11. O’Callaghan A, Laking G, Frey R, et al. Can we predict which hospitalised patients are in their last year of life? A prospective cross-sectional study of the Gold Standards Framework Prognostic Indicator Guidance as a screening tool in the acute hospital setting. Palliat Med. 2014;28:1046–1052. [PubMed]
12. Thomas K. The GSF Prognostic Indicator Guidance. Prognostic Indicator Guidance (PIG) 4th ed. The Gold Standards Framework Centre in End of Life Care, October 2011. Available at http://www.goldstandardsframework.org.uk/cd-content/uploads/files/General Files/Prognostic Indicator Guidance October 2011.pdf.
13. Lynn J. Perspectives on care at the close of life. Serving patients who may die soon and their families: The role of hospice and other services. JAMA. 2001;285:925–932. [PubMed]
14. Morita T, Tsunoda J, Inoue S, et al. The Palliative Prognostic Index: A scoring system for survival prediction of terminally ill cancer patients. Support Care Cancer. 1999;7:128–133. [PubMed]
15. Pirovano M, Maltoni M, Nanni O, et al. A new palliative prognostic score: A first step for the staging of terminally ill cancer patients. J Pain Symptom Manage. 1999;17:231–239. [PubMed]
16. Scarpi E, Maltoni M, Miceli R, et al. Survival prediction for terminally ill cancer patients: Revision of the palliative prognostic score with incorporation of delirium. The Oncologist. 2011;16:1793–1799. [PMC free article] [PubMed]
17. Gwilliam B, Keeley V, Todd C, et al. Development of prognosis in palliative care study (PiPS) predictor models to improve prognostication in advanced cancer: Prospective cohort study. BMJ. 2011;343:d4920. [PubMed]
18. Baba M, Maeda I, Morita T, et al. Independent validation of the modified prognosis palliative care study predictor models in three palliative care settings. J Pain Symptom Manage. 2014 [Epub ahead of print] [PubMed]
19. Maltoni M, Scarpi E, Pittureri C, et al. Prospective comparison of prognostic scores in palliative care cancer populations. The Oncologist. 2012;17:446–454. [PMC free article] [PubMed]
20. Hui D, dos Santos R, Chisholm G, et al. Clinical signs of impending death in cancer patients. The Oncologist. 2014;19:681–687. [PMC free article] [PubMed]
21. 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. The Oncologist. 2011;16:1642–1648. [PMC free article] [PubMed]

Articles from The Oncologist are provided here courtesy of AlphaMed Press