Table 3.
Comparison of percentage risk allocated to each category between error grids.
| Percentage values falling into each risk category | |||||||
|---|---|---|---|---|---|---|---|
| Error Grid | A (total) | A (false negative) | B | C | D | E | C+ |
| CEG | 63.2 | 3.7 | 29.6 | 0.0 | 7.4 | 0.0 | 7.4 |
| PEG | 77.0 | 5.8 | 21.0 | 2.0 | 0.1 | 0.0 | 2.0 |
| SEG | 65.0 | ?? | 23.6 | 11.2 | 0.2 | 0.0 | 11.4 |
| HEG (all) | 88.3 | 0.0 | 7.0 | 4.7 | 0.0 | NA | 4.7 |
| HEG (<4mmol/L) | 63.6 | 0.0 | 21.3 | 15.2 | 0.0 | NA | 15.2 |
HEG, Hypoglycaemia Error Grid. The HEG allocates a slightly lower percentage of values to categories of moderate and higher risk (C+) than the average over the charts due to the allocation of low risk to missed hyperglycaemia. However, this is reversed when the HEG is used only with pairs where at least one value is <4mmol/L and offers a practical risk of 15.2% of values categorised as C+. The CEG and PEG (and likely SEG) classify several false negatives for hypoglycaemia as no risk, while the HEG classifies all false negatives as at least low risk. It is not possible to quantify false negatives assessed by the SEG due to the limited way with which it can be interacted. ?? - not possible to compute, NA, not applicable.