Pilot study on the use of a multimorbidity index in patients receiving home parenteral nutrition

TL;DR

This pilot study explores whether a multimorbidity index can predict resource use and outcomes in patients receiving home parenteral nutrition.

Contribution

The study evaluates the use of the Cumulative Illness Rating Scale (CIRS) as a multimorbidity index in home parenteral nutrition patients.

Findings

Higher CIRS scores were associated with increased PN formula changes and hospital length of stay.

Regression analysis showed a significant interaction between hospital LOS and PN formula changes in patients with CIRS scores ≥17.

The study suggests the need for further validation of the multimorbidity index in this patient population.

Abstract

Home parenteral nutrition (HPN) patients often have multiple comorbidities. A validated multimorbidity index (MMI) could help determine resource needs and risks with HPN. We evaluated MMI scoring to determine if it predicted HPN resource use and outcomes.

We performed a secondary analysis of 60 HPN patients from a previous study based on Cumulative Illness Rating Scale (CIRS) scoring. We examined three variables: PN formula changes, hospitalizations, and hospital length of stay (LOS). Density plots were produced to select a CIRS score cutoff value. Spearman correlations among the three variables of interest were computed. The data were then subjected to Wilcoxon rank sum tests and negative binomial regression models to determine if the measured variables differed significantly between the groups.

In 60 HPN patients, CIRS scores ranged from 9 to 25 with a mean ± SD of 17.0 ± 3.85. Patients with CIRS scores ≥ 17 had higher rates of the three variables than those with CIRS scores < 17 (PN formula changes = 367 vs 297, hospitalizations = 19 vs 12, and total hospital LOS days = 122 vs 100). Although these raw data did not differ significantly, negative binomial regression analysis indicated that the interaction of total hospital LOS and PN formula changes was significantly higher in patients with a CIRS score ≥ 17 than a CIRS score < 17.

CIRS data from HPN patients showed a trend wherein higher multimorbidity scores were associated with PN changes and LOS. This approach requires further study and validation.

Figures