Prognostic Models Associated with 6-Month Survival of Patients Admitted to Nursing Homes

Abstract

Background: Health status and the needs presented by people admitted to nursing homes make it necessary to contemplate aspects such as prognosis to offer quality palliative care. Objective: To compare the prognostic utility in nursing homes of two prognostic models of 6-month survival based on the Palliative Prognostic Index (PPI) or Palliative Performance Status (PPS) instruments and palliative needs indicators. Methods: A longitudinal prospective observational and analytical cohort study of survival and prognostic models in 88 patients with palliative needs (assessed by the NECPAL-ICO-CCOMS©) from an Andalusian (Spain) nursing home was performed. Sociodemographic and clinical variables were assessed, and 6 months later, in September 2017, survival was checked. Multiple logistic regression analysis was performed using the R-Commander program (version 3.2.2). Results: Two models of the logistic regression analysis met the fit criteria. The two models combined the Surprise Question, the presence of persistent symptoms, and the clinical indicators of severity from the NECPAL tool, in addition to the Charlson Comorbidity Index, and varied only in terms of the latter variable, including the PPI in the first model and the PPS in the second. In the first model, significant associations were identified between 6-month survival and the persistent symptoms variable (OR = 7.78, p = 0.025, 95% CI = 1.45–60.92) and PPI (OR = 1.94, p < 0.001, 95% CI = 1.21). In the second model, 6-month survival was also significantly associated with the persistent symptoms variable (OR = 4.57, p = 0.045, 95% CI = 1.07–22.41) and the PPS (OR = 0.93, p = 0.001, 95% CI = 0.88–0.96). Conclusions: Prognostic models such as ours that include variables commonly included in clinical assessments can help nursing home professionals prioritize and ensure adequate mobilization of palliative care resources, which are very limited in these institutions.