Prognostic Tool in Pediatric Oncological Hospice

In the December 1st issue of Pediatric Blood and Cancer is an article that presents the validation of a prognostic tool in pediatric hospice care. The study was produced by a team from the Hospital A.C. Camargo, a large cancer center in São Paulo, Brazil. Their overall survival rate in the treatment of pediatric cancers is just over 75%, roughly on par with those in the United States at approximately 80%. At this institution, a nurse-led, multidisciplinary palliative care team was developed in 1999. Patients are referred to it when 3 pediatric oncologists reach a consensus that a particular patient has no chance for a cure.



In 2005, the same group had developed the prognostic tool to predict 60-day survival of pediatric oncology patients. Factors associated with length of survival were identified through Kaplan–Meier and Cox analysis. Variables with the highest predictive values were identified and relatively scored. The salient variables were:

• Diagnosis, CNS/Solid Tumors being the reference scored as 0, and Leukemias/Lymphomas scored as 1
• Hemoglobin Level, greater than 8.0 g/dl being the reference scored as 0, and less than 8.0 g/dL scored as 1
• Home Care Giver, Mother being the reference scored as 0, and Other scored as 1
• Performance Scale scored by the Home Care Giver (Lansky Play Perfomance Scale for patients aged 0-16 years, Karnofsky Performance Scale for patients older than 16), 100 - 80 being the reference scored as 0, 70 - 20 scored as 1, and 10 - 0 scored as 2.5.

These variables were re-tested in this study, and the variable “number of symptoms” was added for testing, which was found not to change the predictive value. Total scores ranged from 0 – 6.5. The patients were divided into three terciles (2005-study probabilities in parentheses): A, scores 0, 60-day survival probability 80% (84.4%) B, scores 1.0 – 1.5, 60-day survival probability 38% (57.8%) C, scores 2.0 – 6.5, 60-day survival probability 28.5% (15.4%).

Some thoughts:
To think the unthinkable is what we do. Being involved in the care of families with children with terminal cancer is difficult. Having validated notions of where things stand is helpful in gauging the type, timing and frequency of team interventions is grounding. Plus, they provide a focus for communication with these families. I didn’t find much out there lately on prognostic tools for pediatric patients with non-acute, life-threatening/life-limiting diagnoses; there's a fair amount on acute-illness, ICU survival. Semantically, the authors/editors use the term palliative care when it more precise to use the term hospice, which I used in this post for framing. The logician in me didn't like the presentation of data/criteria that used greater-than and less-than, but did not account for values of equal-to. So, I think that there are some theoretical and systematic issues for full generalizability to practice in the Developed World, and to individual national health systems. However, these parameters are easily accessible and make clinical sense. We would do well to replicate and validate such prognostic tools for this population across systems.







Kurashima, A., Latorre, M., & Camargo, B. (2010). A palliative prognostic score for terminally ill children and adolescents with cancer Pediatric Blood & Cancer, 55 (6), 1167-1171 DOI: 10.1002/pbc.22644