Meta-analysis highlights importance of communication in DNR decision-making

  • Becker C & al.
  • JAMA Netw Open
  • 5 Jun 2019

  • International Clinical Digest
Access to the full content of this site is available only to registered healthcare professionals. Access to the full content of this site is available only to registered healthcare professionals.


  • Communication regarding CPR risks/outcomes and do-not-resuscitate (DNR) preferences might attenuate unnecessary medical interventions in the emergency department (ED).

Why this matters

  • Proactively link communication around CPR/code status with decisions about other emergency care to improve patient understanding, clarify achievable outcomes to help patients make informed DNR decisions.
  • "[I]deally, discussions should enable the clinician and patient to achieve a shared understanding of what outcomes that particular patient values and what outcomes are clinically achievable for them," Gavin D. Perkins, MD, FRCP, FFICM, writes in a related editorial.

Key results

  • 15 trials (11 quantitative synthesis, 5 secondary [patient knowledge] endpoint).
  • Pooled: significant association between communication interventions, lower preference for CPR 53.6% vs 38.6% usual care (risk ratio [RR], 0.70 [95% CI, 0.63-0.78; P<.001>
  • Stratification by intervention type: video-based education (RR, 0.56; 95% CI, 0.48-0.64) vs other education types (RR, 1.03; 95% CI, 0.87-1.22) linked to lower CPR preference (between-group heterogeneity, P<.001>
  • Secondary: while stratification by age did not demonstrate significant differences between younger vs older patients (P=.48), in studies with low bias risk, a stronger association noted between communication interventions and higher patient knowledge vs higher bias risk trials (standardized mean difference, 0.60; 95% CI, 0.43-0.77 vs 0.28; 95% CI, −0.10 to 0.67; between-group heterogeneity, P=.14).

Study design

  • Meta-review assessing association between communication interventions and patient DNR preference based on knowledge of CPR risks, outcomes.
  • Funding: None disclosed.


  • Limited number of studies.
  • Limited generalizability.
  • Type 2 error risk.