Accuracy of Clinicians’ Prediction of Survival and Prognostic Factors Indicative of Survival: a Systematic Review

Full Article

M Zhou, L Holden, N Lao, H Lam, L Zeng, E Chow

Hong Kong J Radiol 2013;16:168-82

DOI: 10.12809/hkjr1313179

Objectives: To review the literature and examine the accuracy of clinicians’ prediction of survival as well as prognostic factors determined to be predictive of shorter survival in terminal cancer patients.

Methods: A literature search was conducted on MEDLINE (1 January 2000 to 29 July 2012), Embase (1 January 2000 to 22 July 2012), and Cochrane Database of Systematic Reviews (1 January 2005 to July 2012). Reference sections of relevant reviews were also examined for relevant articles. All studies examining the accuracy of prediction of survival and prognostic factors indicative of survival in patients with terminal cancer were selected. Descriptive statistics summarised the extracted data.

Results: A total of 85 studies published from 1972 to July 2012 with a study cohort of 30 to 6066 patients were identified. Clinicians’ prediction of survival correlated with patient’s actual survival, but the predictions tended to be too optimistic. The ability of varying health care professionals to estimate survival was contradictory among different studies. Some studies noticed those with more experience with terminal cancer patients were better able to predict an accurate estimation, whereas others concluded that there was no difference. The estimations were also more accurate during short-term time ranges such as the ‘horizon effect’. Only a few assessment tools to assist in predicting the remaining duration of survival in patients were validated. A variety of prognostic factors between studies were identified, but the factors were not validated nor any instruments created.

Conclusion: Demographic information and clinical symptoms can assist in determining the remaining duration of survival of terminal cancer patients. Assessment tools should be convenient and not a burden for the patient as the goal of palliative care is to maintain or improve the quality of life as much as possible. Even with the application of instruments to formulate a prediction, error cannot be completely eliminated. Physicians should warn the patient and their family of the uncertainty of the predictions.

 

中文摘要

臨床醫生預測生存期的準確度及預後因素:系統性綜述

M Zhou, L Holden, N Lao, H Lam, L Zeng, E Chow

目的:回顧文獻並調查臨床醫生對晚期癌症患者生存期的預測準確度及較短生存期的預後因素。

方法:文獻檢索MEDLINE(2000年1月1日至2012年7月29日)、Embase(2000年1月1日至2012年7月22日)和Cochrane系統評價數據庫(2005年1月1日至2012年7月)的資料,並一併回顧所選取文獻中的參考文獻資料。選取所有調查晚期癌症患者生存期的預測準確度及預後因素的研究。描述性統計總結提取的數據。

結果:共發現於1972年到2012年7月期間發表,病人數目從30至6066名不等的85項隊列研究。將臨床醫生預測的生存期與病人實際生存期做相關性分析,發現預測往往過於樂觀。不同醫療保健專家對不同研究中生存期的估計存在矛盾。一些研究發現在晚期癌症患者方面經驗更豐富的臨床醫生能更準確地預測生存期,而另外一些研究得出的結論是兩者並無差異。此外,在短時間範圍內的估計也會更準確。用作幫助預測患者生存的剩餘期限的評估工具中,只有少數得到驗證。研究中找出了各種預後因素,但這些因素並未得到證實,也沒有創立任何預測工具。

結論:人口統計學的資料和臨床症狀可協助確定晚期癌症患者的剩餘生存時間。由於姑息治療的目的是盡可能維持或改善病人生活質素,所以評估工具應以方便為主,而不是加重病人的負擔。即使應用預測工具來預測病人生存期,亦不可能完全沒有錯誤。醫生應提醒患者及其家屬預測生存期的不確定性。