ABCDEFGHIJKLMNOPQRSTUVWXYZ
1
Application of 4% chlorhexidine to umbilical stump in the community - meta analysis inputs and outputs
2
3
Table 1: Meta-analysis inputs
4
CHX group total deathsCHX group NControl group total deathsControl group NRegion Control group NNM Risk ratio95%Weight in meta analysis
5
Arifeen et al. 2012 (ES estimates from Table 4)
4871975228310008
South Asia
28.30.8719
0.75-1.01
34%
6
Mullany et al. 2006 (ES estimates from Table 5)
724924985082
South Asia
19.30.7583
0.56-1.02
8%
7
Soofi et al., 2012 (ES estimates from Table 2)11148671764874
South Asia
36.10.6316
0.499-0.80
13%
8
Sazawal et al. 2016 (ES estimates from Table 3)
1891801522118896Africa11.70.8970
0.73-1.08
19%
9
Semrau et al. 2016 (ES estimates from Table 4)
2811851026319346Africa13.61.1167
0.94-1.31
26%
10
NOTES
11
Death rates calculated from deaths per 1,000 as reported in cited tables.
12
Arifeen results from single-cleansing and multiple-cleansing trial arms were pooled for this analysis, analyzed separately for meta-regression of number of days on outcome (see Table 5 in this workbook)
13
Meta analysis conducted using Comprehensive Meta Analysis version 3.3.070
14
15
Table 2: Pooled results
16
All studiesSouth Asia onlyAfrica onlyHigh NNM only**
17
Odds ratioRisk ratiop-valueRisk ratiop-valueRisk ratiop-value
Risk ratio
p-value
18
Fixed effects model0.8840.8860.0050.7920.0001.0170.7960.7980.000
19
Random effects model*0.8510.8540.0840.7600.0071.0060.9600.7560.040
20
21
Q-valuedf (Q)P-valueI-squared
22
Heterogeneity in RE model (all studies) 16.490 4.000 0.002 75.743
23
NOTES
24
*Random effects model preferred
25
**NNM>25 per 1,000. Includes Arifeen and Soofi studies only.
26
Pooled results were calculated using Comprehensive Meta Analysis Version 3.3.070
27
28
Table 3: Pooled results: Forest plot, random effects model only
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
Table 4: Meta-regression results (Random effects Method of Moments, Z-Distribution, Log risk ratio)
49
CoefficientStandard95%95%Z-value2-sided
50
ErrorLowerUpperP-value
51
Intercept0.18330.1848-0.1790.54550.990.3214
52
Control group NNM-0.0150.0079-0.03080-1.960.0505
53
54
Statistics for Model 1
55
Test of the model: Simultaneous test that all coefficients (excluding intercept) are zero
56
Q = 3.83, df = 1, p = 0.0505
57
Goodness of fit: Test that unexplained variance is zero
58
Tau² = 0.0153, Tau = 0.1238, I² = 59.88%, Q = 7.48, df = 3, p = 0.0581
59
60
Comparison of Model 1 with the null model
61
Total between-study variance (intercept only)
62
Tau² = 0.0307, Tau = 0.1753, I² = 75.74%, Q = 16.49, df = 4, p = 0.0024
63
Proportion of total between-study variance explained by Model 1
64
R² analog = 0.50
65
66
67
68
Table 5: Meta-regression results (Random effects Method of Moments, Z-Distribution, Log risk ratio)
69
 CoefficientStandard95%95%Z-value2-sided
70
ErrorLowerUpperP-value
71
Intercept-0.0730.2391-0.54160.3956-0.310.7601
72
Number of days of application-0.0120.0262-0.06350.039-0.470.6389
73
74
Statistics for Model 1
75
Test of the model: Simultaneous test that all coefficients (excluding intercept) are zero
76
Q = 0.22, df = 1, p = 0.6389
77
Goodness of fit: Test that unexplained variance is zero
78
Tau² = 0.0518, Tau = 0.2276, I² = 83.58%, Q = 18.27, df = 3, p = 0.0004
79
80
Comparison of Model 1 with the null model
81
Total between-study variance (intercept only)
82
Tau² = 0.0361, Tau = 0.1899, I² = 78.23%, Q = 18.38, df = 4, p = 0.0010
83
Proportion of total between-study variance explained by Model 1
84
R² analog = 0.00 (computed value is -0.44)
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100