Table 2.
Meta-analysis of POCT devices and subgroups
| Analysis group | Number of studies | Number of patients | Bias (g/l) | Standard deviation of bias | Between-study heterogeneity in bias (I2) | 95% LOA (g/l) |
|---|---|---|---|---|---|---|
| Pulse Co-Ox | 25 | 1110 | 2.3 | 1.15 | 0.57 | (−25.2,29.8) |
| Pulse Co-Ox (funded) | 8 | 423 | 2.1 | 1.17 | 0.98 | (−31.0,35.1) |
| Pulse Co-Ox (unfunded) | 17 | 687 | 2.2 | 1.14 | 0.18 | (−22.1,26.6) |
| Pulse Co-Ox (blood loss > 1 litre) | 5 | 305 | −2.5 | 0.90 | 0.80 | (−27.9,22.9) |
| HemoCue | 10 | 525 | −0.3 | 0.48 | 0.06 | (−11.1,10.5) |
| HemoCue (arterial) | 5 | 314 | 0.1 | 0.41 | 0.07 | (−9.6,9.87) |
| HemoCue (capillary) | 5 | 228 | 1.4 | 0.93 | 0.12 | (−18.5,21.2) |
| HemoCue (venous) | 2 | 90 | −1.2 | 0.51 | 0.19 | (−14.5,12.1) |
| iSTAT | 6 | 146 | −0.3 | 0.32 | 0.06 | (−8.4,7.8) |
| Blood gas analyser | 10 | 821 | −2.6 | 0.51 | 0.32 | (−17.8,12.7) |
| Low risk of bias (Pulse Co-Ox) | 13 | 675 | 1.9 | 1.20 | 0.97 | (−29.2,32.9) |
POCT, point-of-care tests; LOA, limits of agreement.