Emergency Department Barthel Index Score Predicts Heart Failure Mortality
Researchers used adjusted and unadjusted logistic regression models to examine the relationship between Barthel Index scores and 30-day mortality and used C statistics to assess prognostic value.
Functional status as assessed by the Barthel Index score of patients visiting the emergency department (ED) with acute heart failure functions as a strong all-cause 30-day mortality predictor whereas change in Barthel Index score from baseline to ED visit did not add prognostic value, according to study results published in the Annals of Emergency Medicine.
This study was designed to examine the prognostic value of functional status as measured by the Barthel Index at time of ED visit and acute functional decline (as measured by difference between baseline Barthel Index score and score at ED visit) to predict the 30-day mortality risk for patients with acute heart failure. Participants (mean age, 80.2±10.2 years; 56% women) were Spanish patients from 4 cohorts of the registry for Acute Heart Failure in Emergency Departments who had both baseline and ED visit Barthel Index scores available (N=9098). Associations between Barthel Index scores and 30-day mortality were assessed using adjusted and unadjusted logistic regression models, and C statistics were used to evaluate prognostic value.
The median Barthel Index score was 90 (interquartile range 60-100) at baseline and 70 (interquartile range 45-90) at time of ED visit. Acute functional decline, defined as ≥5-point decrease between scores at baseline and at ED visit, was observed in 53.4% of patients (n=5771). Within 30 days, 9.9% of patients died (n=905). A steep gradient in 30-day mortality was seen across 10-point interval categories in both baseline and ED visit scores (linear trend P <.001 for each case). For example, a Barthel Index score of 50-55 as compared with a score of 100 doubled the mortality risk at both baseline and time of ED visit. Although both baseline and ED visit Barthel Index scores functioned well as predictors of mortality, the latter provided greater insight regardless of the Barthel Index score categorization (P <.001), and a logistic model that simultaneously fit the scores at both time points as continuous 30-day mortality predictors showed that baseline Barthel Index scores did not add prognostic value (P =.65).
Limitations included a degree of variability in Barthel Index scores taken by different physicians, a failure to record Barthel Index items separately, self-reported scores and a potential for recall bias, and a lack of data on do-not-resuscitate orders.
Study investigators concluded that these findings show that "functional status assessed by the [Barthel Index] score at the ED visit is a strong predictor for all-cause 30-day mortality in acute heart failure patients," and that a "systematic collection of functional status at the ED visit should be recommended for all patients with acute heart failure who are attending the ED."
Rossello X, Miró Ò, Llorens P, et al. Effect of Barthel Index on the risk of thirty-day mortality in patients with acute heart failure attending the emergency department: a cohort study of nine thousand ninety-eight patients from the Epidemiology of Acute Heart Failure in Emergency Departments Registry [published online January 19, 2019]. Ann Emerg Med. doi: 10.1016/j.annemergmed.2018.12.009