Perioperative intermediate care units (termed surgical special care units) have been widely implemented across health systems because they are believed to improve surveillance and management of high-risk surgical patients. Our objective was to conduct a systematic review to investigate the effects of a 3-level model of perioperative care delivery (ie, ward, surgical special care unit, or intensive care unit) compared to a 2-level model of care (ie, ward, intensive care unit) on postoperative outcomes. Our protocol was registered with PROSPERO, the international prospective register of systematic reviews (CRD42015025155). Randomized controlled studies and nonrandomized comparator studies were included. We performed a systematic search of Medline, Cumulative Index to Nursing and Allied Health Literature, Embase, and the Cochrane library (inception - 11/2017). The primary outcome was mortality; secondary outcomes included length of stay and hospital costs. We identified 1995 citations with our search, and 21 studies met eligibility criteria (2 randomized controlled studies and 19 nonrandomized comparator studies; 44,134 patients in total). Surgical special care units were characterized by continuous monitoring (12 studies), the absence of mechanical ventilation (8 studies), nurse-to-patient ratios (range, 1:2-1:4), and number of beds (median: 5; range: 3-33). Thirteen studies reported on mortality. Notable findings included no observed difference in overall in-hospital mortality, but an apparent increase in intensive care unit mortality in a 3-level model of care. This may reflect a decanting of lower acuity patients from the intensive care unit to the surgical special care unit. No significant difference was found in hospital length of stay; however, 2 studies demonstrated reductions in hospital costs with the implementation of a surgical special care unit. Significant clinical and methodological heterogeneity precluded pooled analysis. Given the prevalence of surgical special care units, the results of our review suggest that additional methodologically rigorous investigations are needed to understand the effect of these units on the surgical population.