Physicians often rely on surrogate decision-makers (SDMs) to make important decisions on behalf of critically ill patients during times of incapacity. It is uncertain whether targeted interventions to improve surrogate decision-making in the intensive care unit (ICU) reduce nonbeneficial treatment and improve SDM comprehension, satisfaction, and psychological morbidity. To perform a systematic review and meta-analysis of randomized clinical trials (RCTs) to determine the association of such interventions with patient- and family-centered outcomes and resource use. A search was conducted of MEDLINE, Embase, and other relevant databases for potentially relevant studies from inception through May 30, 2018. Randomized clinical trials studying interventions that were targeted at SDMs or family members of critically ill adults in the ICU were included. Key search terms included surrogate or substitute decision-maker, critically ill, randomized controlled trials, and their respective related terms. This study followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines. Two independent, blinded reviewers independently screened citations and extracted data. Random effects models with inverse variance weighting were used to pool outcomes data when possible and otherwise present findings qualitatively. Outcomes of interest were divided into 3 categories: (1) patient-related clinical outcomes (mortality, length of stay [LOS], duration of life-sustaining therapies), (2) SDM and family-related outcomes (comprehension, major change in goals of care, incident psychological comorbidities [posttraumatic stress disorder, anxiety, depression], and satisfaction with care), and (3) use of resources (cost of care and health care resource use). Of 3735 studies screened, 13 RCTs were included, comprising a total of 10 453 patients. Interventions were categorized as health care professional led (n = 6), ethics consultation (n = 3), palliative care consultation (n = 2), and media (n = 1 pamphlet and 1 video). No association with mortality was observed (risk ratio, 1.03; 95% CI, 0.98-1.08; P = .22). Intensive care unit LOS was significantly shorter among patients who died (mean difference, -2.11 days; 95% CI, -4.16 to -0.07; P = .04), but not in the overall population (mean difference, -0.79 days; 95% CI, -2.33 to 0.76 days; P = .32). There was no consistent difference in SDM-related outcomes, including satisfaction with care or perceived quality of care (n = 6 studies) and incident psychological comorbidities (depression: ratio of means, -0.11; 95% CI, -0.29 to 0.08; P = .26; anxiety: ratio of means, -0.08; 95% CI, -0.25 to 0.08; P = .31; or posttraumatic stress disorder: ratio of means: -0.04; 95% CI, -0.21 to 0.13; P = .65). Among 6 trials reporting effects on health care resource use, only 1 nurse-led intervention observed a significant reduction in costs ($75 850 control vs $51 060 intervention; P = .04). Systematic interventions aimed at improving surrogate decision-making for critically ill adults may reduce ICU LOS among patients who die in the ICU, without influencing overall mortality. Better understanding of the complex processes related to surrogate decision-making is needed.