Examining the Predictive Capabilities of Re-Referral Rates in Palliative Care From the Palliative Care Outcomes Collaboration Program Scoring System

OBJECTIVES: Used in palliative care clinical monitoring, the Palliative Care Outcomes Collaboration Program (PCOC), affords additional insight to patient-level outcomes. The objective of this analysis was to assess the relationship between clinical assessment scores within PCOC in palliative care (PC) services and the likelihood of re-referral post-discharge.

METHODS: Between 2020-2022, PCOC data was collected on 100 Palliative Care (PC) patients across a single site in Ireland. Observation scores ranged from one observation per patient to six within that timeframe, with 15% of patients having data from admission to discharge available. Within this dataset, Problem Severity Score (PSS), Symptom Assessment Score (SAS), Resource Utilization Groups — Activities of Daily Living (RUG_ADL), and Australia-modified Karnofsky Performance Status Scale (AKPS) scores were collected, and personal characteristics. Descriptive analysis and a series of inferential analysis were undertaken to establish if patient PCOC scores can predict their likelihood of re-referral.

RESULTS: Descriptive statistics revealed re-referred patients were marginally younger, spent less days in care, and had less severe PSS and SAS scores. Several regression models using various sub-scales lacked predictive power for re-referral. However, as a composite RUG-ADL & AKPS score increased, the likelihood of the patient being re-referred decreased, meaning that patients in deteriorated states at admission were less likely to be re-referred after discharge. Further, as RUG-ADL score increases, i.e. the patient state deteriorates, the likelihood of them being re-referred decreases by approximately 14% (P=0.029).

CONCLUSIONS: This was the first study in Ireland assessing the predictive capability of PCOC in identifying risks of re-referral. Findings were inconclusive given several limitations. There is evidence to suggest that factors potentially influencing re-referral to PC services include the number of PCOC scores taken, time spent in care, and RUG-ADL/AKPS. Further studies with larger cohorts, improvements to data capture and more structured follow-up timepoints could yield more definitive conclusions.