The application of the palliative prognostic index in predicting the life expectancy of patients in palliative care : a systematic review and meta-analysis

Article

LIU, Ying | SU, Lin | WANG, Yuting | LIU, Shuang | DONG, Birong

BACKGROUND: The palliative prognostic index (PPI) is a commonly used tool to predict the life expectancy in palliative care patients. However, there is no universal cutoff, and the accuracy of different cutoffs varies. Therefore, we conducted this meta-analysis to explore the validity and accuracy of different PPI scores for different survival time in palliative care setting. METHODS: PubMed, Embase, Cochrane, Scopus and Chinese CNKI databases were searched to identify studies using the PPI as a prognostic tool to predict survival time in palliative care. We calculated pooled hazard ratios (HRs) with corresponding 95% confidence intervals (CIs), and subgroup analyses were also conducted by different cutoffs. After extracting data, we estimated the pooled sensitivity, specificity, positive predictive value, and negative predictive value. RESULTS: We identified 15 studies with 7455 assessments. Seven of these studies were synthesized for a combined HR. The pooled HR was 1.94 (95% CI 1.54-2.44) when cutoffs were 2 and 4, and 2.34 (95% CI 1.50-3.66) when cutoffs were 4 and 6. Of all the studies, 13 studies reported their accuracy, of which four studies were assessed by meta-analysis. The sensitivity of the PPI for 3-week survival ranged from 51 to 92% and specificity ranged from 60.0 to 94.0%, respectively. The sensitivity and specificity of the PPI for 6-week survival were from 46.0 to 89.1% and from 51.7 to 84.4%, respectively. The pooled sensitivity and specificity of the PPI for 3-week survival were 68% (6 as cutoff) and 76% (6 as cutoff), respectively. As for 6-week survival prediction, the pooled sensitivity and specificity were 68% (4 as cutoff) and 82% (4 as cutoff), respectively. CONCLUSION: The PPI is a useful prognosticator of life expectancy of patients in palliative care, especially for patients with short survival time. However, there were no universal cutoff, and the predicted life span varies. Our data eliminated that using 4 and 6 as cutoffs can better predict the patients' survival time for 3 or 6 weeks. Due to small number of studies and poor qualities of them, result may alter as more studies with better quality are enrolled in the future.

http://dx.doi.org/10.1007/s40520-018-0928-7

Voir la revue «Aging clinical and experimental research»

Autres numéros de la revue «Aging clinical and experimental research»

• Description
• Sujet/Public
• Outil
• Outil d'anticipation
• Gestion