Searches of PubMed, Embase, and CINAHL yielded this one meta-analysis and two systematic reviews.

Kim, Sang-Hyun, et al. “Accuracy and precision of continuous noninvasive arterial pressure monitoring compared with invasive arterial pressure: a systematic review and meta-analysis.” Anesthesiology 120.5 (2014):1080-1097.
Twenty-eight studies with 919 patients were included in this systematic review. “The overall random-effect pooled bias and SD were -1.6 ± 12.2 mmHg (95% limits of agreement -25.5 to 22.2 mmHg) for systolic arterial pressure, 5.3 ± 8.3 mmHg (-11.0 to 21.6 mmHg) for diastolic arterial pressure, and 3.2 ± 8.4 mmHg (-13.4 to 19.7 mmHg) for mean arterial pressure. In 14 studies focusing on currently commercially available devices, bias and SD were -1.8 ± 12.4 mmHg (-26.2 to 22.5 mmHg) for systolic arterial pressure, 6.0 ± 8.6 mmHg (-10.9 to 22.9 mmHg) for diastolic arterial pressure, and 3.9 ± 8.7 mmHg (-13.1 to 21.0 mmHg) for mean arterial pressure.
CONCLUSIONS: The results from this meta-analysis found that inaccuracy and imprecision of continuous noninvasive arterial pressure monitoring devices are larger than what was defined as acceptable. This may have implications for clinical situations where continuous noninvasive arterial pressure is being used for patient care decisions.”

Ben Sivarajan, V, and DesmondBohn. “Monitoring of standard hemodynamic parameters: heart rate, systemic blood pressure, atrial pressure, pulse oximetry, and end-tidal CO2.” Pediatric critical care medicine 12.4 Suppl (2011):S2-S11.
The conclusion of this systematic review states, “literature would suggest that invasive arterial monitoring is the current standard for monitoring in the setting of shock. The use of heart rate, electrocardiography, and atrial pressure monitoring is advantageous in specific clinical scenarios (postoperative cardiac surgery); however, the evidence for this is based on numerous case series only.”

Chatterjee, Arjun, et al. “Results of a survey of blood pressure monitoring by intensivists in critically ill patients: a preliminary study.” Critical care medicine 38.12 (2010):2335-2338.
“Eight hundred eighty individuals received an invitation to complete the survey and 149 responded. We found that 71% (105 of 149) of intensivists estimated the correct cuff size rather than measuring arm circumference directly. In hypotensive patients, 73% of respondents (108 of 149) reported using noninvasive blood pressure measurement measurements for patient management. In patients on a vasopressor medication, 47% (70 of 149) of respondents reported using noninvasive blood pressure measurement for management.”