Inter-Rater Reliability of the Phase of Illness Tool in Pediatric Palliative Care

• Published in: American journal of hospice & palliative medicine 2020-10, Vol.37 (10), p.837-843
• Main Authors: Burke, Kimberley, Coombes, Lucy H., Petruckevitch, Ann, Anderson, Anna-Karenia
• Format: Article
• Language: English
• Subjects: Adult; Child; Hospice and Palliative Care Nursing; Humans; Palliative Care; Pediatrics; Prospective Studies; Reproducibility of Results
• ISSN: 1049-9091; 1938-2715
• DOI: 10.1177/1049909120912674

Background: Phase of Illness is used to describe the stages of a patient’s illness in the palliative care setting. Categorization is based on individual needs, family circumstances, and the adequacy of a care plan. Substantial (κ = .67) and moderate (κ = .52) inter-rater reliability is demonstrated when categorizing adults; however, there is a lack of similar studies in pediatrics. Objective: To test the inter-rater reliability of health-care professionals when assigning pediatric palliative care patients to a Phase of Illness. Furthermore, to obtain user views on phase definitions, ease of assignment, feasibility and acceptability of use. Method: A prospective cohort study in which up to 9 health-care professionals’ independently allocated 80 pediatric patients to a Phase of Illness and reported on their experiences. This study took place between June and November 2017. Results: Professionals achieved a moderate level of agreement (κ = 0.50). Kappa values per phase were as follows: stable = 0.63 (substantial), unstable = 0.26 (fair), deteriorating = 0.45 (moderate), and dying = 0.43 (moderate). For the majority of allocations, professionals report that the phase definitions described patients very well (76.1%), and they found it easy to assign patients (73.5%). However, the unstable phase caused the most uncertainty. Conclusion: The results of this study suggest Phase of Illness is a moderately reliable, acceptable, and feasible tool for use in pediatric palliative care. Current results are similar to those found in some adult studies. However, in a quarter of cases, users report some uncertainty in the application of the tool, and further study is warranted to explore whether suggested refinements improve its psychometric properties.