General introduction 19 because the opportunity to timely prepare patient and family for the end of life is missed out. 3.4 Estimation of life expectancy Estimating the life expectancy of patients can help clinicians to become aware of possible palliative care needs and to determine which treatments are appropriate.62 Clinicians can ask themselves a temporal question: “How long will this patient live?”, which is answered with a number of time units (for example, 10 months). The temporal question is considered a simple way to formulate estimated survival. However, temporal predictions are often too optimistic and not very accurate.71 A probabilistic question, “What is the chance that this patient survives the following hours, days, weeks, months, or year?”, is more accurate, but more difficult to interpret.56, 72 Moreover, clinicians are hesitant to estimate the patient ’s life expectancy, because it may feel definitive and as if one provides a guarantee.56 3.5 Use of the Surprise Question The surprise question, “Would I be surprised if this patient dies within one year?”, is a tool to early identify patients who may benefit from palliative care.73 If the answer to the surprise question is “No, I would not be surprised”, assessment of palliative care needs should be conducted. The surprise question is asked and answered by the clinician. The surprise question is not primarily meant to estimate survival within one year, but it combines a clinician-estimated probability of death during the following year to identify those who are currently in need of palliative care, with a gut feeling.73-75 Time frames other than death within one year (e.g., 6 months, 3 months, 1 month, or 1 week) can be used to determine the urgency of palliative care and what care should be aimed at, for example, supportive care, hospice care or terminal care.76 Systematic reviews and meta-analyses show that the surprise question has a c-statistic for one-year mortality ranging from 0.512 to 0.822 with an accuracy of 75%. However, the surprise question lacks sensitivity and positive predictive value (67% and 37%, respectively).77, 78 The surprise question might perform better when predictors of approaching death are added.78 No studies are conducted to test whether the surprise question adequately identifies patients with palliative care needs. Nonetheless, use of the surprise question is still recommended in clinical guidelines, such as the Netherlands Quality Framework for Palliative Care.2 3.6 Use of tools combining identification methods The Netherlands Quality Framework for Palliative Care advocates that criteria for palliative care case finding should become part of the care process. These criteria include trigger moments, predictors for approaching death, and the surprise question.2, 40 Several screening tools are developed that combine the aforementioned ways of identifying patients with 1
RkJQdWJsaXNoZXIy MTk4NDMw