Discussion

This case highlights that prognostication, even when backed by scientific studies, is not always accurate. A prior study has shown that doctors overestimate the time a patient has remaining (Cheon et al. Reference Cheon, Agarwal and Popovic2016), but this case showcases an example of experienced palliative care specialists underestimating the prognosis. The uncertainty regarding survival can be distressing to family members who are trying their best to process the condition of a loved one while simultaneously making the necessary arrangements (Hui et al. Reference Hui, Paiva and Del Fabbro2019).

In our case, the patient was determined to be imminently dying based on the low PPS and flattening of nasolabial folds. These findings have been shown to be indicative of death in less than 3 days in 94% of cases (Hui et al. Reference Hui, Hess and Dos Santos2015). Prognostication in cancer can be particularly difficult. Patients who die of diseases related to organ failure such as CHF, COPD, or CKD have markers of the function of the organ that can assist in prognosis and show decline. Cancer patients tend to die of complications, most frequently, sepsis, thrombosis, or arrhythmias without warnings of decline (Bruera et al. Reference Bruera, Chisholm and Dos Santos2015). The nature and onset of those complications cannot be predicted by tumor mass. Another major reason for difficulty in prognosis is the continuous improvement in cancer therapies (Larkin et al. Reference Larkin, Chiarion-Sileni and Gonzalez2019). Prognosis for malignant melanoma has changed greatly in the last decade from medial survival of 11 months in 2013 to 24 months in 2021 (van Not et al. Reference van Not, van den Eertwegh and Haanen2024). This blurring of therapeutic limits is occurring for multiple hematological and solid tumors (Knight et al. Reference Knight, Karapetyan and Kirkwood2023). When our team reconsulted the melanoma oncologists they concluded that the patient might benefit from further treatment. One of the advantages of palliative care teams in acute care facilities is the seamless interaction with specialists that make consultation much less difficult for the patient and family as compared to hospice revocation.

Our patient later clinically improved and revoked comfort care while in the PCU, electing to pursue more treatment offered by the melanoma team. One study showed that around 18% of patients who are discharged from hospice do so to resume disease directed therapy (Russell et al. Reference Russell, Diamond and Lauder2017). Discharge from hospice can be a difficult experience for patients and caregivers (Wu and Volker Reference Wu and Volker2019). Though our patient was able to pursue disease directed therapy after electing comfort care, one study showed that patients who choose to revoke hospice to pursue treatment have higher mortality rates in the next 6 months following revocation (LeSage et al. Reference LeSage, Borgert and Rhee2014). Another study showed that aggressive medical treatment at end of life was associated with worse caregiver bereavement (Wright et al. Reference Wright, Zhang and Ray2008). Improving our existing survival prognostic models might prevent this kind of distress for patients and families.

Physicians who care for patients toward the end of life are experienced in keeping families updated day to day, as clinical status can change quickly and without notice. Current prognostic models can help guide discussion with family members, but physicians need to be aware of the risk of the anchoring bias, in which the medical team only makes decisions based on the initial presented information, rather than reevaluating the situation completely even when things do not go as expected (Rehana and Huda Reference Rehana and Huda2021). Continuously reassessing day to day and keeping families updated is of vital importance. A bias toward palliative care is less common (Erel et al. Reference Erel, Marcus and Dekeyser-Ganz2022), but as this case highlights, is still possible.

A number of studies have shown the need for better communication for families from the medical team at this delicate time. Some of the issues family members had include not receiving an explanation for how long the patient and family could talk and having a feeling of uncertainty caused by vague explanations about future changes (Mori et al. Reference Mori, Morita and Bruera2022). Though the inaccuracies can be distressing, forgoing these survival prognostic models completely is also not the answer. Not giving family members a chance to prepare for the patient’s death could result in a sense of having unfinished business, which leads to higher depression and grief scores (Mori et al. Reference Mori, Morita and Bruera2022).

Uncertainty in prognosis cannot be helped, however there are actions the medical team can take to mitigate the effects of uncertainty for family members. For instance, it is common practice at our institution to use large margins in our temporal and percentage estimations due to this uncertainty (Cheon et al. Reference Cheon, Agarwal and Popovic2016), including phrases like “hours to days” and “days to weeks.” Once patients have elected for comfort care, we do not pursue invasive measures, like labs or imaging, which could potentially help diagnose any complications that come up. Instead, we assure the patient and family members that we will manage any distress that is the consequence of complication with symptom management. Another important strategy is to remind caregivers that because we are imperfect at prognostication (Chen et al. Reference Chen, Chung and Lam2023), any plans or decisions need to be made immediately. Giving the caregivers this opportunity allows them to care for themselves by prioritizing their rest and nutrition so they can care for the patient better as well.