Cancer versus Non-Cancer Patients Admitted to Hospice: a comparison study.

Context

Malignant disease remains the leading reason for hospice admission, and direct, patient-level comparisons of cancer and non-cancer patients within the same residential hospice are scarce.

Objectives

To compare sociodemographic and clinical characteristics, symptom burden, nutritional support, length of stay and outcome between oncological and non-oncological patients admitted to a single hospice.

Methods

Prospective observational cohort study of 175 consecutive admissions over one year, after removal of duplicate records. Patients were classified as oncological (n=151, 86.3%) or non-oncological (n=24, 13.7%) based on structured case report form fields. Sociodemographic, functional (Karnofsky Performance Status, KPS), nutritional, symptom (Edmonton Symptom Assessment System, ESAS) and cognitive (Memorial Delirium Assessment Scale, MDAS) data were collected at admission (T0); ESAS completion was also examined at 1 week (T7). Categorical variables were compared with the chi-square or Fisher's exact test, continuous variables (reported as median and interquartile range, including ESAS and MDAS scores) with the Mann-Whitney U test; p-values were adjusted for multiple comparisons with the Benjamini-Hochberg procedure (q).

Results

Most non-oncological patients had a neurological disease (21/24), mainly neurodegenerative diseases, amyotrophic lateral sclerosis and dementia. Non-oncological patients were older (median 78.5 vs 72 years, P=0.002, q=0.012) and had worse performance status at admission (median KPS 20 vs 30, P<0.001, q=0.006). Artificial nutrition was more frequent among non-oncological patients (37.5% vs 14.9%, P=0.018), although not after correction for multiple comparisons (q=0.060). MDAS scores were higher in non-oncological patients (median 13 vs 4, P=0.021). ESAS data were missing in about half of oncological and three quarters of non-oncological patients; no ESAS item differed between groups after correction for multiple comparisons, and the non-oncological group's valid ESAS sample fell to a single patient by T7. Length of stay (11 vs 8 days, P=0.518) and in-hospice mortality (80.0% vs 70.8%, P=0.455) did not differ significantly.

Conclusion

In this single-centre series, non-oncological patients, mostly affected by neurological diseases, were admitted older, more functionally compromised and with higher MDAS scores than oncological patients, despite comparable length of stay and mortality. Apparent differences in symptom burden are confounded by substantial, likely informative, missing data and should be interpreted with caution. Whether these differences reflect later referral or the age profile of the underlying diseases cannot be determined from these data.