Cancer Pain Trajectories in the Last Two Years of Life

生命最後兩年的癌症疼痛軌跡

Jia-Wen Guo(郭佳雯) ； Djin Tay ； Sharon M. Weinstein ； Ann Lyons ； Katherine A. Sward

《醫療資訊雜誌》 Volume 32 Issue 1 (2023/03) Pp. 41-58

Abstracts

Pain is a prevalent consequence of cancer that may be associated with the progression of the disease. Understanding pain trajectory patterns can support end of life pain management, a crucial part of palliative care. This secondary data analysis study identifies 2 pain trajectory patterns 2 years before death- stable with mild pain, and elevating to moderate pain, by analyzing 989 deceased cancer patients' data from electronic health records using a longitudinal machine learning algorithm. Age at death, sex, comorbidities, bone cancer diagnosis, and receiving cancer treatments within 3 months before death were significant predictors of pain trajectories based on the decision tree, random forest, and logistic models. The findings of this study can inform timely patient and clinician communications to improve pain management and end-of-life palliative care plans.

Keywords

Cancer pain ； Palliative care ； End of life care ； K-means cluster

Parallel abstracts

疼痛是癌症的常見症狀，其與疾病的進展是可能相關的。瞭解癌痛的軌跡變化可支持安寧緩和療法與癌末疼痛管理。本次級資料研究係以縱貫機器學習算法分析989名已故癌症患者的電子病歷記錄，發現其死亡前兩年有兩種疼痛軌跡變化：「穩定的輕度疼痛」和「加劇為中度疼痛」。根據決策樹、隨機森林和邏吉斯回歸分析，這兩種疼痛軌跡變化的顯著預測因子包括：死亡年齡、性別、合併症、骨癌診斷，以及是否在死亡前三個月內接受癌症治療。本研究結果可促進臨床醫病溝通以及時改善癌症疼痛管理，也可應用於面安寧緩和治療的計畫。

Parallel keywords

癌症疼痛 ； 安寧緩和 ； 臨終關懷 ； K-means