Skip to main content

Table 2 Result summaries of meta-analysis for levels of skin pigmentation and ethnic groups

From: The accuracy of pulse oximetry in measuring oxygen saturation by levels of skin pigmentation: a systematic review and meta-analysis

Subgroup categories

No. of studies (evaluations)

Sample size (data pairs)

Calculated Arms, %

Pooled mean bias (95% CI), %

Pooled SD (95% CI), %

Calculated 95% limits of agreement, %

Overall I2 (between-studies and within-study heterogeneity) in mean bias data pooling

High (dark) skin pigmentation

 Primary analysis

8 (24)

221 (3270)

1.88

1.11 (0.29 to 1.93)

1.52 (1.30 to 1.79)

 − 1.87 to 4.09

98.03% (0% and 98.03%)

 Sensitivity analysis, excluding studies with high overall risk of bias

6 (15)

177 (2691)

1.75

0.87 (− 0.46 to 2.19)

1.52 (1.20 to 1.93)

 − 2.11 to 3.84

–

 Sensitivity analysis, excluding studies with no use of standardised scales for measuring skin pigmentation

5 (9)

160 (474)

1.79

0.89 (− 1.37 to 3.14)

1.55 (1.12 to 2.15)

 − 2.16 to 3.93

–

 Sensitivity analysis, excluding studies with participants of high skin pigmentation alone

6 (14)

88 (2738)

1.95

1.12 (− 0.16 to 2.39)

1.60 (1.26 to 2.03)

 − 2.02 to 4.26

–

Medium pigmentation

 Primary analysis

4 (10)

406 (1323)

1.58

 − 0.58 (− 2.25 to 1.09)

1.47 (1.08 to 2.00)

 − 3.46 to 2.30

92.65% (82.39% and 10.25%)

 Sensitivity analysis, excluding studies with no use of standardised scales for measuring skin pigmentation

3 (4)

399 (399)

1.66

 − 0.79 (− 3.03 to 1.45)

1.46 (1.00 to 2.14)

 − 3.66 to 2.08

–

Low (light) skin pigmentation

 Primary analysis

6 (15)

670 (2865)

1.53

 − 0.35 (− 1.36 to 0.67)

1.49 (1.23 to 1.81)

 − 3.27 to 2.58

92.73% (22.42% and 70.31%)

 Sensitivity analysis, excluding studies with high overall risk of bias

5 (12)

660 (2245)

1.57

 − 0.47 (− 1.77 to 0.83)

1.50 (1.21 to 1.87)

 − 3.42 to 2.47

–

 Sensitivity analysis (scale)

4 (6)

648 (667)

1.62

 − 0.54 (− 2.52 to 1.43)

1.53 (1.18 to 1.98)

 − 3.53 to 2.45

–

Participants described as Black/African American

 Primary analysis

9 (22)

459 (5753)

2.27

1.52 (0.95 to 2.09)

1.68 (1.32 to 2.14)

 − 1.78 to 4.82

96.39% (0% and 96.39%)

 Sensitivity analysis, excluding studies with high overall risk of bias

4 (10)

67 (2892)

1.99

1.47 (− 0.21 to 3.16)

1.35 (1.08 to 1.69)

 − 1.17 to 4.11

–

 Sensitivity analysis, excluding studies with data transformation

7 (20)

316 (3110)

2.26

1.55 (0.85 to 2.25)

1.64 (1.28 to 2.11)

 − 1.67 to 4.77

–

 Sensitivity analysis, excluding studies with Black/African American populations alone

8 (16)

426 (5621)

2.35

1.63 (0.77 to 2.49)

1.69 (1.25 to 2.28)

 − 1.68 to 4.94

–

Participants from other ethnic groups, described as Asian, Hispanic or mixed ethnicity

 Primary analysis

3 (9)

522 (2646)

1.58

0.31 (0.09 to 0.54)

1.55 (0.53 to 4.53)

 − 2.72 to 3.35

47.95% (0% and 47.95%)

 Sensitivity analysis, excluding studies with high overall risk of bias

2 (7)

41 (2165)

1.23

0.36 (− 0.24 to 0.95)

1.18 (0.14 to 9.78)

 − 1.95 to 2.66

–

 Sensitivity analysis, excluding studies with data transformation

2 (8)

488 (1405)

2.07

0.30 (− 0.80 to 1.40)

2.04 (0.32 to 12.81)

 − 3.70 to 4.31

–

White/Caucasian

 Primary analysis

13 (48)

2195 (12,870)

1.64

0.55 (− 0.21 to 1.31)

1.55 (1.31 to 1.82)

 − 2.48 to 3.58

94.39% (69.92% and 24.47%)

 Sensitivity analysis, excluding studies with high overall risk of bias

8 (26)

1293 (7770)

1.42

0.36 (− 0.88 to 1.61)

1.38 (1.19 to 1.59)

 − 2.33 to 3.06

–

 Sensitivity analysis, excluding studies with data transformation

10 (45)

1044 (5485)

1.67

0.63 (− 0.38 to 1.63)

1.54 (1.32 to 1.80)

 − 2.40 to 3.65

–

 Sensitivity analysis, excluding studies with White/Caucasian populations alone

8 (16)

1246 (9791)

1.84

0.98 (− 0.08 to 2.05)

1.56 (1.16 to 2.10)

 − 2.07 to 4.04

–