User's Manual

G4-Global System User’s Guide | 135
Table 4 shows that the G4-Global Receiver, on average, reads 13% higher than the
YSI at glucose levels of 40-80 mg/dL and reads 9% lower, on average, than the YSI at
glucose concentrations of 301-400 mg/dL (Mean Percent Difference). The G4-Global
Receiver reads, on average, 25% different than the YSI at glucose levels of
40-80 mg/dL and, on average, 13% different than the YSI readings at glucose levels of
301-400 mg/dL (Mean Absolute Percent Difference). The Median Percent Difference
shows that 50% of the time the G4-Global Receiver reads approximately 3% less than
the YSI System and the Median Absolute Difference shows that 50% of the time the G4-
Global Receiver reads about 13% different then YSI blood glucose readings taken within
5 minutes.
Table 5. Clarke Error Grid Analysis
The Clarke Error Grid was used to determine if G4-Global Receiver results
are “clinically accurate.” The Clarke Error Grid divides a correlation plot into 5
zones, which are described in the table below.
Zone Description
A Clinically Accurate, would lead to correct treatment decisions
B Would lead to benign decisions or no treatment
C Would lead to overcorrection of normal glucose levels
D Would lead to failure to be aware of and treatment of high or low glucose
levels
E Would lead to inaccurate and “dangerous” treatment decisions
Results in zones A and B are considered “clinically acceptable”, in which you would most
likely make accurate treatment decisions or decide not to do treatment with the readings
provided, while results in zones C, D, and E can lead to dangerous treatment errors. The
Clarke Error Grid zones are labeled on a correlation plot in Figure 1. Each YSI result (on
the horizontal axis) is matched up with its corresponding G4-Global Receiver result (on
the vertical axis). A point is recorded where the two readings intersect. These results are
based on 1,827 paired data
points recorded.
Section 11.1
Placeholder for G4-Global Receiver Pivotal Study Results