Patients with metastatic cancer require substantial heath care resources. Palliative care has been increasingly recognized for improvement of quality of life and reducing healthcare costs. Here, we examined the effect of prior palliative care encounters on the total hospital charges (TOTCHG) and length of stay (LOS) during the subsequent hospitalization.
We used National Inpatient Sample (NIS) 2014 to extract data for patients non-electively hospitalized with corresponding ICD9 code of previous palliative care visit (ICD9 code V667) and metastatic cancer. NIS is a nationally representative survey of hospitalizations conducted by Heathcare Cost and Utilization project. It represents 20% of all hospital data in the US. Univariate regression screening (threshold P > 0.1) and hybrid selection were used to create multivariate regression models. Relationship between TOTCHG and previous palliative care encounter as well as LOS and previous palliative care encounter were analyzed by using established models.
A total number of 136591 patients admitted non-electively with metastatic cancer was identified among which 24736 had been coded for previous palliative care encounter. Teaching hospital admission, rural hospital admission, self-pay, increased age and increased Charlson score were associated with higher rate of previous palliative encounter. The multivariate regression model for LOS and previous palliative care visit were adjusted for survival outcome, number of procedures during hospitalization, number of previous chronic conditions, and number of the diagnosis during hospitalization. The model for TOTCHG and previous palliative care visit were adjusted for survival outcome, number of procedures and length of stay. We found that previous palliative care encounter was associated with both lower total hospital charge (P
Prior palliative care visit has been associated with decreased length of stay and total hospital charges. Future studies are needed to determine if early outpatient palliative care encounter will especially benefit patients with certain tumor types.
Clinical trial identification
Legal entity responsible for the study
All authors have declared no conflicts of interest.