According to a study in the November 17 issue of JAMA.

However, current evidence to aid the usage of AEDs in hospitals has been mixed and limited by single-center studies. Also, these devices may be much less effective or potentially harmful when found in hospitals where only one 1 in 5 hospitalized patients have preliminary cardiac arrest rhythms that react to defibrillation. Before the widespread dissemination of AEDs in hospitals, it therefore becomes critical to show that AED use improves survival, the authors compose. Using data from the National Registry of Cardiopulmonary Resuscitation, Paul S.Qualitative variables were weighed against the usage of the chi-square test or Fisher’s test, quantitative variables with the use of Student’s t-test or a non-parametric test, and survival data with the use of a stratified log-rank check. All these comparisons were adjusted for stratification factors. All checks were two-sided, with a P worth of less than 0.05 thought to indicate statistical significance. Data are presented with 95 percent confidence intervals, calculated by using standard methods predicated on a binomial distribution. All analyses had been performed by using Stata software, version 10.22 A Cox proportional-hazards model was used to estimate the hazard ratios. Hazard ratios indicating the consequences of prognostic elements on the risk of loss of life were calculated and are demonstrated in a forest plot.23 The interaction test was used to assess the heterogeneity of treatment effects for subgroup analyses.24 Analysis of the QLQ-C30 questionnaires was performed relative to the EORTC guidelines.25 The preplanned analysis centered on the scales that are usually most affected in patients with pancreatic cancer: the Global Health Status and Standard of living scale and scales for fatigue, pain, physical functioning, emotional functioning, and role functioning.26 The other QLQ-C30 domains were only examined in an exploratory manner.