3-1 DESCRIPTIVE FINDINGS
The mean age of the control group is 28.60 years and the mean age of the case group is 27.30 years. 36.59 percent of the case group and 40.16 percent of the control group are 15-24.9 years. 69.92 percent of cases and 54.33 of controls live in the village(Table3).
Table 3 Demographic characteristics of severe maternal morbidity cases and controls
variables
|
case
(number=123)
|
controls
(number=127)
|
frequency
|
Percent(%)
|
frequency
|
Percent(%)
|
Age(year)
|
15-24.9
|
45
|
36.59
|
51
|
40.16
|
25-29.9
|
27
|
21.95
|
29
|
22.83
|
30-34.9
|
22
|
17.89
|
24
|
18.90
|
46-35
|
29
|
23.58
|
12
|
18.11
|
address
|
urban
|
37
|
30.08
|
58
|
45.67
|
village
|
86
|
69.92
|
69
|
54.33
|
Employment status
|
employee
|
3
|
2.44
|
6
|
4.72
|
housewife
|
120
|
97.56
|
121
|
95.28
|
children
|
Less than 2 children
|
71
|
57.75
|
83
|
65.35
|
More than 2 children
|
52
|
42.28
|
44
|
34.65
|
Family relationship with spouse
|
yes
|
13
|
10.57
|
7
|
5.51
|
no
|
110
|
89.43
|
120
|
94.49
|
abortion
|
Less than 2 times
|
122
|
99.19
|
126
|
99.21
|
More than 2 times
|
1
|
0.81
|
1
|
0.79
|
stillbirth
|
yes
|
4
|
3.25
|
4
|
3.15
|
no
|
119
|
96.75
|
123
|
96.85
|
According to Table 4, the distribution of mediators (pregnancy, delivery, and neonatal risk factors) in the case group is significantly higher than in the control group. Neonatal risk factors have the most frequency, 75.61 percent of cases and 62.99 of controls have neonatal risk factors.
Table 4 Distribution of cases and controls based on pregnancy, delivery and neonatal risk factors(mediators)
variables
|
Cases(123)
|
Controls(127)
|
Number
|
Percent(%)
|
Number
|
Percent(%)
|
Pregnancy risk factors
|
yes
|
66
|
53.66
|
18
|
14.29
|
no
|
57
|
46.34
|
108
|
85.71
|
Delivery risk factors
|
yes
|
49
|
39.84
|
19
|
14.96
|
no
|
74
|
60.16
|
108
|
14.96
|
Neonatal risk factors
|
yes
|
93
|
75.61
|
80
|
62.99
|
no
|
30
|
24.39
|
47
|
37.01
|
The most prevalent violence in table5 for cases is social violence(68.29%) and the most prevalence violence for control group is total violence(48.78%) and mental violence(44.88%)(Table5).
Table 5 Distribution of case and controls based on violence and it's subgroup
Violence subgroup
|
Cases(123)
|
Controls(127)
|
frequency
|
Percent(%)
|
frequency
|
Percent(%)
|
Total violence
|
yes
|
63
|
51.22
|
60
|
48.78
|
no
|
60
|
48.78
|
66
|
51.97
|
Mental violence
|
yes
|
53
|
43.09
|
57
|
44.88
|
no
|
70
|
56.91
|
70
|
55.12
|
Physical violence
|
yes
|
37
|
30.08
|
24
|
18.90
|
no
|
86
|
69.92
|
103
|
81.10
|
Social violence
|
yes
|
84
|
68.29
|
21
|
16.54
|
no
|
39
|
31.71
|
106
|
83.46
|
Economicalal violence
|
yes
|
56
|
45.53
|
52
|
40.94
|
no
|
67
|
54.47
|
75
|
59.06
|
Sexual violence
|
yes
|
21
|
17.7
|
21
|
16.54
|
no
|
102
|
82.93
|
106
|
83.46
|
3-2 ANALYTICAL FINDINGS
In table 6,7,8,9 there is no significant effect for total and mental violence and severe maternal morbidity.
Table 6 Mediation analysis between total violence and severe maternal morbidity by PARAMED
adjusted |
crude |
effect
|
mediator
|
P value
|
CI
|
OR
|
P value
|
CI2
|
OR1
|
0.193
|
(0.15- 1.46)
|
0.47
|
0.117
|
(0.13-1.24)
|
0.42
|
CDE3
|
Pregnancy risk factors
|
0.766
|
(0.54-2.27)
|
1.11
|
0.861
|
(0.45-1.92)
|
0.94
|
PDE4
|
0.148
|
(0.92-1.71)
|
1.26
|
0.150
|
(0.92-1.63)
|
1.23
|
TIE5
|
0.406
|
(0.63-3.10)
|
1.40
|
0.731
|
(0.50-2.26)
|
1.15
|
MTE6
|
0.921
|
(0.31-3.59)
|
1.06
|
0.923
|
(0.32-2.77)
|
0.95
|
CDE
|
Delivery risk factors
|
0.562
|
(0.64-2.22)
|
1.20
|
0.529
|
(0.67-2.13)
|
1.20
|
PDE
|
0.403
|
(0.74-1.120
|
0.91
|
0.589
|
(0.77-1.15)
|
0.94
|
TIE
|
0.763
|
(0.59-2.02)
|
1.10
|
0.680
|
(0.61-2.08)
|
1.14
|
MTE
|
0.203
|
(0.78-3.29)
|
1.58
|
0.482
|
(0.68-2.25)
|
1.24
|
CDE
|
Neonatal risk factors
|
0.323
|
(0.74-2.48)
|
1.36
|
0.547
|
(0.70-1.94)
|
1.17
|
PDE
|
0.541
|
(0.83-1.09)
|
0.96
|
0.610
|
(0.80-1.08)
|
0.97
|
TIE
|
0.395
|
(0.71-2.37)
|
1.30
|
0.626
|
(0.60-1.89)
|
1.14
|
MTE
|
1OR=odds ratio, 2CI=confidence interval, 3CDE= controlled direct effect, 4PDE=pure direct effect, 5TIE=total indirect effect, 6MTE=marginal total effect
Table 7 Mediation analysis between total violence and severe maternal morbidity by MED4WAY
adjusted
|
crude
|
effect
|
mediator
|
P value
|
CI
|
ERR
|
P value
|
CI
|
ERR1
|
0.478
|
(-0.71-1.51)
|
0.40
|
0.749
|
(-0.79-1.10)
|
0.15
|
TE2
|
Pregnancy risk factors
|
0.180
|
(-7.59-1.41)
|
-3.09
|
0.095
|
(-8.36-0.66)
|
-3.80
|
CDE3
|
0.134
|
(-0.98-7.39)
|
3.20
|
0.078
|
(-0.42-7.99)
|
3.79
|
INT REF4
|
0.250
|
(-1.53-0.40)
|
-0.57
|
0.256
|
(-1.23-0.32)
|
-0.45
|
INT MED5
|
0.144
|
(-0.29-1.99)
|
0.85
|
0.187
|
(-0.32-1.67)
|
0.67
|
PIE6
|
0.774
|
(-0.57-0.77)
|
0.10
|
0.699
|
(-0.55-0.82)
|
0.14
|
TE
|
Delivery risk factors
|
0.922
|
(-2.42-2.68)
|
0.12
|
0.922
|
(-3.13-2.83)
|
-0.15
|
CDE
|
0.945
|
(-1.97-2.12)
|
0.07
|
0.781
|
(-2.12-2.83_
|
0.35
|
INT REF
|
0.945
|
(-0.21-0.23)
|
0.01
|
0.801
|
(0.00-0.13)
|
0.01
|
INT MED
|
0.314
|
(-0.32-0.10)
|
-0.11
|
0.564
|
(-0.35-0.19)
|
-0.08
|
PIE
|
0.426
|
(-0.46-1.10)
|
0.32
|
0.648
|
(-0.44-0.71)
|
0.14
|
TE
|
Neonatal risk factors
|
0.270
|
(-0.48-1.74)
|
0.63
|
0.518
|
(-0.56-1.11)
|
0.27
|
CDE
|
0.268
|
(-0.67-0.18)
|
-0.24
|
0.566
|
(-0.47-0.25)
|
-0.11
|
INT REF
|
0.525
|
(-0.21-0.11)
|
-0.05
|
0.697
|
(-0.08-0.05)
|
-0.01
|
INT MED
|
0.732
|
(-0.07-0.05)
|
-0.01
|
0.623
|
(-0.09-0.05)
|
-0.02
|
PIE
|
1ERR=excess relative risk,2CI= confidence interval, 3TE=total effect,4CDE=controlled direct effect, 5INT REF=reference interaction, 6INT MED=mediated interaction, 7PIE=pure indirect effect
Table8 Mediation analysis between mental violence and severe maternal morbidity by PARAMED
adjusted
|
crude
|
Effect
|
mediator
|
P value
|
CI
|
OR
|
P value
|
CI2
|
OR1
|
0.293
|
(0.18-1.67)
|
0.55
|
0.345
|
(0.20-1.73)
|
0.60
|
CDE3
|
pregnancy risk factors
|
0.147
|
(0.33-1.17)
|
0.63
|
0.276
|
(0.38-1.31)
|
0.71
|
PDE4
|
0.215
|
(0.86-1.87)
|
1.27
|
0.159
|
(0.89-1.98)
|
1.33
|
TIE5
|
0.564
|
(0.37-1.70)
|
0.80
|
0.887
|
(0.44-2.02)
|
0.95
|
MTE6
|
0.330
|
(0.16-1.82)
|
0.55
|
0.510
|
(0.23-2.03)
|
70/0
|
CDE
|
delivery risk factors
|
0.148
|
(0.35-1.17)
|
0.64
|
0.441
|
(0.46-1.39)
|
0.81
|
PDE
|
0.495
|
(0.89-1.25)
|
1.06
|
0.278
|
(0.89-1.48)
|
1.15
|
TIE
|
0.229
|
(0.36-1.27)
|
0.68
|
0.815
|
(0.50-1.71)
|
0.93
|
MTE
|
0.514
|
(0.38-1.60)
|
0.79
|
0.526
|
(0.66-2.21)
|
1.21
|
CDE
|
neonatal risk factors
|
0.221
|
(0.36-1.26)
|
0.68
|
0.954
|
(0.58-1.66)
|
0.98
|
PDE
|
0.451
|
(0.80-1.10)
|
0.94
|
0.492
|
(0.80-1.11)
|
0.94
|
TIE
|
0.158
|
(0.34-1.19)
|
0.64
|
0.788
|
(0.54-1.58)
|
0.93
|
MTE
|
1OR=odds ratio, 2CI=confidence interval, 3CDE= controlled direct effect, 4PDE=pure direct effect, 5TIE=total indirect effect, 6MTE=marginal total effect
Table9 Mediation analysis between mental violence and severe maternal morbidity by MED4WAY
adjusted
|
crude
|
effect
|
mediator
|
P value
|
CI
|
ERR
|
P value
|
CI
|
ERR1
|
0.517
|
(-0.80-0.40)
|
-0.20
|
0.884
|
(-0.77-0.66)
|
-0.05
|
TE2
|
pregnancy risk factors
|
0.272
|
(-4.87-1.37)
|
-1.75
|
0.318
|
(-5.76-1.87)
|
-1.95
|
CDE3
|
0.348
|
(-1.49-4.25)
|
1.38
|
0.356
|
(-1.86-5.17)
|
1.66
|
INT REF4
|
0.455
|
(-0.56-0.25)
|
-0.15
|
0.445
|
(-0.63-0.27)
|
-0.18
|
INT MED5
|
0.270
|
(-0.25-0.91)
|
0.33
|
0.222
|
(-0.24-1.07)
|
0.41
|
PIE6
|
0.143
|
(-0.74-0.10)
|
-0.32
|
0.808
|
(-0.63-0.49)
|
-0.07
|
TE
|
delivery risk factors
|
0.327
|
(-3.00-1.00)
|
-1.00
|
0.498
|
(-66.3-1.78)
|
-0.94
|
CDE
|
0.496
|
(-1.20-2.49)
|
0.64
|
0.553
|
(-1.73-3.23)
|
0.75
|
INT REF
|
0.612
|
(-0.20-0.12)
|
-0.04
|
0.598
|
(-0.31-0.18)
|
-0.06
|
INT MED
|
0.494
|
(-0.15-0.31)
|
0.08
|
0.306
|
(-0.17-0.55)
|
0.18
|
PIE
|
0.114
|
(-0.74-0.07)
|
-0.33
|
0.780
|
(-0.56-0.42)
|
-0.70
|
TE
|
neonatal risk factors
|
0.563
|
(-0.82-0.44)
|
-0.19
|
0.556
|
(-0.53-0.99)
|
0.23
|
CDE
|
0.503
|
(-0.39-0.19)
|
-0.10
|
0.169
|
(-0.59-0.10)
|
-0.24
|
INT REF
|
0.587
|
(-0.12-0.06)
|
-0.03
|
0.527
|
(-0.17-0.09)
|
-0.04
|
INT MED
|
0.623
|
(-0.08-0.05)
|
-0.02
|
0.660
|
(-0.06-0.03)
|
-0.01
|
PIE
|
1ERR=excess relative risk,2CI= confidence interval, 3TE=total effect,4CDE=controlled direct effect, 5INT REF=reference interaction, 6INT MED=mediated interaction, 7PIE=pure indirect effect
According to Table 10, the odds ratio (OR) of pure direct effect and marginal total effect of physical violence is 2.26(p=0.048) and 2.30(p=0.029) through pregnancy risk factors(mediator), respectively. The odds ratio of pure and controlled direct effect and the marginal total effect of physical violence through neonatal risk factors are 2.42(p=0.043), 2.29(p=0.025), and 2.15(p=0.035) respectively. These significant effects show a direct effect of physical violence on severe maternal morbidity.
The excess relative risk of reference interaction between physical violence and pregnancy risk factor is 3.94(P=0.022) which shows that this interaction increased the risk of SMM by 394%(table 11).
Table10 Mediation analysis between physical violence and severe maternal morbidity by PARAMED
adjusted
|
crude
|
effect
|
mediator
|
P value
|
CI
|
OR
|
P value
|
CI2
|
OR1
|
0.189
|
(0.15-1.45)
|
0.47
|
0.098
|
(0.13-1.18)
|
0.40
|
CDE3
|
pregnancy risk factors
|
0.048*
|
(1.00-5.08)
|
2.26
|
0.162
|
(0.79-3.81)
|
1.75
|
PDE4
|
0.933
|
(0.72-1.42)
|
1.01
|
0.561
|
(0.79-1.53)
|
1.10
|
TIE5
|
0.029*
|
(1.08-4.84)
|
2.30
|
0.087
|
(0.90-4.07)
|
1.93
|
MTE6
|
0.839
|
(0.24-5.58_
|
1.17
|
0.801
|
(0.22-3.15)
|
0.84
|
CDE
|
delivery risk factors
|
0.058
|
(0.97-3.85)
|
1.94
|
0.062
|
(0.97-3.43)
|
1.82
|
PDE
|
0.711
|
(0.84-1.11)
|
0.97
|
0.815
|
(0.91-1.11)
|
1.01
|
TIE
|
0.068
|
(0.95-3.75)
|
1.89
|
0.064
|
(0.96-3.53)
|
1.85
|
MTE
|
0.043*
|
(1.02-5.71)
|
2.42
|
0.143
|
(0.82-3.62)
|
1.73
|
CDE
|
neonatal risk factors
|
0.025*
|
(1.10-4.73)
|
2.29
|
0.039
|
(1.03-3.45)
|
1.89
|
PDE
|
0.486
|
(0.78-1.11)
|
0.94
|
0.668
|
(0.88-1.02)
|
0.98
|
TIE
|
0.035*
|
(1.05-4.39)
|
2.15
|
0.048
|
(1.00-3.37)
|
1.85
|
MTE
|
1OR=odds ratio, 2CI=confidence interval,3 CDE= controlled direct effect,4 PDE=pure direct effect, 5TIE=total indirect effect, 6MTE=marginal total effect
Table11 Mediation analysis between physical violence and severe maternal morbidity by MED4WAY
adjusted
|
crude
|
effect
|
mediator
|
P value
|
CI
|
ERR
|
P value
|
CI
|
ERR1
|
0.138
|
(-0.41-3.02)
|
1.30
|
0.209
|
(0.51-2.37)
|
0.93
|
TE2
|
pregnancy risk factors
|
0.128
|
(-6.12-0.77)
|
-2.68
|
0.049*
|
(-7.16- -0.02)
|
-3.59
|
CDE3
|
0.022*
|
(0.57-7.31)
|
3.94
|
0.013*
|
(0.92-7.75)
|
4.34
|
INT REF4
|
0.101
|
(-2.83-0.25)
|
-0.29
|
0.094
|
(-2.61-0.20)
|
-1.20
|
INT MED5
|
0.058
|
(-0.04-2.96)
|
1.32
|
0.056
|
(-0.03-2.80)
|
1.38
|
PIE6
|
0.178
|
(-0.40-2.19)
|
0.89
|
0.166
|
(-0.35-2.04)
|
0.84
|
TE
|
delivery risk factors
|
0.849
|
(-3.67-4.46)
|
0.40
|
0.788
|
(-4.12-3.13)
|
-0.50
|
CDE
|
0.755
|
(-2.87-3.97)
|
0.54
|
0.410
|
(-1.87-4.47)
|
1.32
|
INT REF
|
0.776
|
(-0.22-0.29)
|
0.04
|
0.811
|
(-0.30-0.23)
|
-0.03
|
INT MED
|
0.505
|
(-0.34-0.16)
|
-0.09
|
0.804
|
(-0.37-0.47)
|
0.05
|
PIE
|
0.142
|
(-0.38-2.68)
|
1.15
|
0.137
|
(-0.26-1.96)
|
0.85
|
TE
|
neonatal risk factors
|
0.169
|
(-0.68-3.90)
|
1.61
|
0.253
|
(-0.64-2.45)
|
0.90
|
CDE
|
0.448
|
(-1.14-0.50)
|
-0.32
|
0.959
|
(-0.70-0.67)
|
-0.02
|
INT REF
|
0.574
|
(-0.43-0.24)
|
-0.10
|
0.959
|
(-0.11-0.10)
|
-0.00
|
INT MED
|
0.485
|
(-0.15-0.07)
|
-0.04
|
0.610
|
(-0.18-0.10)
|
-0.04
|
PIE
|
1ERR=excess relative risk,2CI= confidence interval, 3TE=total effect,4CDE=controlled direct effect, 5INT REF=reference interaction, 6INT MED=mediated interaction, 7PIE=pure indirect effect
Pure direct effect and the marginal total effect of social violence through pregnancy risk factors increased the odds of severe maternal morbidity by 2.54(p=0.024) and 2.65(p=0.013)(table12).
Pure direct effect and the marginal total effect of social violence through delivery risk factors increased the odds of severe maternal morbidity by 2.67(p=0.023) and 2.24(0.033)(table12).
direct effects (controlled and pure) and the marginal total effect of social violence through neonatal risk factors increased the risk of severe maternal morbidity by 2.45(p=0.033), 2.58(p=0.008), and 2.67(p=0.007) respectively. These significant effects show the direct association between social violence and SMM(table12).
Table12 Mediation analysis between social violence and severe maternal morbidity by PARAMED
adjusted
|
crude
|
effect
|
mediator
|
P value
|
CI
|
OR
|
P value
|
CI2
|
OR1
|
0.396
|
(0.18-1.93)
|
0.60
|
0.278
|
(0.18-1.63)
|
0.54
|
CDE3
|
pregnancy risk factors
|
0.024*
|
(1.13-5.70)
|
2.54
|
0.061
|
(0.96-4.45)
|
2.08
|
PDE4
|
0.824
|
(0.72-1.50)
|
1.04
|
0.505
|
(0.80-1.55)
|
1.12
|
TIE5
|
0.013*
|
(1.23-5.70)
|
2.65
|
0.027*
|
(1.09-4.90)
|
2.32
|
MTE6
|
0.224
|
(0.51-16.97)
|
2.96
|
0.215
|
(0.55-13.70)
|
2.75
|
CDE
|
delivery risk factors
|
0.023*
|
(1.14-6.22)
|
2.67
|
0.015*
|
(1.21-6.11)
|
2.72
|
PDE
|
0.431
|
(0.53-1.13)
|
0.84
|
0.476
|
(0.57-1.29)
|
0.86
|
TIE
|
0.033*
|
(1.06-4.70)
|
2.24
|
0.031*
|
(1.08-5.07)
|
2.34
|
MTE
|
0.033*
|
(1.07-5.60)
|
2.45
|
0.046*
|
(1.01-4.19)
|
2.06
|
CDE
|
neonatal risk factors
|
0.008*
|
(1.28-5.16)
|
2.58
|
0.008*
|
(1.23-4.19)
|
2.28
|
PDE
|
0.601
|
(0.90-1.18)
|
1.04
|
0.661
|
(0.90-1.16)
|
1.03
|
TIE
|
0.007*
|
(1.33-5.44)
|
2.67
|
0.006*
|
(1.27-4.32)
|
2.34
|
MTE
|
1OR=odds ratio, 2CI=confidence interval, 3CDE= controlled direct effect, 4PDE=pure direct effect, 5TIE=total indirect effect, 6MTE=marginal total effect
Table13 Mediation analysis between social violence and severe maternal morbidity by MED4WAY
adjusted
|
crude
|
effect
|
mediator
|
P value
|
CI
|
ERR
|
P value
|
CI
|
ERR1
|
0.110
|
(-0.37-3.96)
|
1.66
|
0.136
|
(-0.41-3.05)
|
1.32
|
TE2
|
pregnancy risk factors
|
0.331
|
(-5.68-1.91)
|
-1.88
|
0.201
|
(-6.41-1.34)
|
-2.53
|
CDE3
|
0.064
|
(-0.19-7.06)
|
0.43
|
0.049*
|
(0.01-7.20)
|
3.61
|
INT REF4
|
0.162
|
(-2.72-0.45)
|
-1.13
|
0.160
|
(-2.20-0.36)
|
-0.92
|
INT MED5
|
0.074
|
(-0.12-2.60)
|
1.24
|
0.078
|
(-0.13-2.46)
|
1.17
|
PIE6
|
0.14
|
(-0.42-2.90)
|
1.24
|
0.146
|
(-0.43-3.15)
|
1.34
|
TE
|
delivery risk factors
|
0.438
|
(-5.79-13.78)
|
3.91
|
0.420
|
(-6.90-16.50)
|
4.82
|
CDE
|
0.585
|
(-10.24-5.77)
|
-2.23
|
0.536
|
(-12.92-6.71)
|
-3.10
|
INT REF
|
0.607
|
(-1.44-0.84)
|
-0.30
|
0.607
|
(-1.15-0.67)
|
-0.24
|
INT MED
|
0.230
|
(-0.35-0.08)
|
-0.13
|
0.377
|
(-0.43-0.16)
|
-0.13
|
PIE
|
0.085
|
(-0.22-3.48)
|
1.63
|
0.066
|
(-0.09-2.77)
|
1.34
|
TE
|
neonatal risk factors
|
0.154
|
(-0.62-3.95)
|
1.67
|
0.415
|
(-0.44-3.05)
|
1.30
|
CDE
|
0.839
|
(-1.25-1.01)
|
-0.18
|
0.967
|
(-1.01-0.97)
|
-0.02
|
INT REF
|
0.844
|
(-0.26-0.32)
|
0.03
|
0.967
|
(-0.25-0.26)
|
0.00
|
INT MED
|
0.467
|
(-0.08-0.19)
|
0.05
|
0.396
|
(-0.07-0.19)
|
0.06
|
PIE
|
1ERR=excess relative risk,2CI= confidence interval, 3TE=total effect,4CDE=controlled direct effect, 5INT REF=reference interaction, 6INT MED=mediated interaction, 7PIE=pure indirect effect
The OR of pure direct effect and marginal total effect are 1.99(P=0.030) and 1.89(P=0.043) for economicalal violence as exposure, neonatal risk factors as mediator, and severe maternal morbidity as outcome(table 14).
Table14 Mediation analysis between economicalal violence and severe maternal morbidity by PARAMED
adjusted
|
crude
|
effect
|
mediator
|
P value
|
CI
|
OR
|
P value
|
CI2
|
OR1
|
0.649
|
(0.25-2.35)
|
0.77
|
0.385
|
(0.21-1.80)
|
0.62
|
CDE3
|
pregnancy risk factors
|
0.207
|
(0.78-3.01)
|
1.54
|
0.628
|
(0.60-2.28)
|
18/1
|
PDE4
|
0.412
|
(0.86-1.40)
|
1.11
|
0.717
|
(0.82-1.31)
|
1.04
|
TIE5
|
0.167
|
(0.80-3.62)
|
1.70
|
0.589
|
(0.58-2.59)
|
1.23
|
MTE6
|
0.375
|
(0.49-6.28)
|
1.78
|
0.993
|
(0.33-2.98)
|
0.99
|
CDE
|
delivery risk factors
|
0.051
|
(0.99-3.51)
|
1.87
|
0.435
|
(0.70-2.21)
|
1.25
|
PDE
|
0.510
|
(0.77-1.13)
|
0.94
|
0.697
|
(0.78-1.17)
|
0.96
|
TIE
|
0.080
|
(0.93-3.29)
|
1.75
|
0.544
|
(0.65-2.20)
|
1.20
|
MTE
|
0.052
|
(0.99-4.22)
|
2.05
|
0.570
|
(0.64-2.19)
|
1.19
|
CDE
|
neonatal risk factors
|
0.030*
|
(1.06-3.73)
|
1.99
|
0.414
|
(0.74-2.07)
|
1.24
|
PDE
|
0.425
|
90.83-1.07)
|
0.95
|
0.553
|
(0.88-1.06)
|
0.97
|
TIE
|
0.043*
|
(1.01-3.52)
|
1.89
|
0.479
|
(0.71-2.02)
|
1.20
|
MTE
|
1OR=odds ratio, 2CI=confidence interval, 3CDE= controlled direct effect, 4PDE=pure direct effect, 5TIE=total indirect effect, 6MTE=marginal total effect
Table 15 Mediation analysis between economicalal violence and severe maternal morbidity by MED4WAY
adjusted
|
crude
|
effect
|
mediator
|
P value
|
CI
|
ERR
|
P value
|
CI
|
ERR1
|
0.281
|
(-0.58-1.99)
|
0.71
|
0.625
|
(-0.68-1.14)
|
0.23
|
TE2
|
pregnancy risk factors
|
0.624
|
(-5.04-3.02)
|
-1.01
|
0.319
|
(-5.48-1.78)
|
-1.85
|
CDE3
|
0.376
|
(-1.89-5.00)
|
1.55
|
0.207
|
(-1.11-5.16)
|
2.02
|
INT REF4
|
0.536
|
(-0.56-0.29)
|
-0.13
|
0.734
|
(-0.37-0.26)
|
-0.05
|
INT MED5
|
0.405
|
(-0.40-0.01)
|
0.30
|
0.726
|
(-0.48-0.70)
|
0.11
|
PIE6
|
0.180
|
(-0.35-1.86)
|
0.75
|
0.580
|
(-0.53-0.93)
|
0.20
|
TE
|
delivery risk factors
|
0.461
|
(-2.60-5.75)
|
1.57
|
0.993
|
(-3.17-3.10)
|
-0.01
|
CDE
|
0.683
|
(-4.06-2.66)
|
0.70
|
0.843
|
(-2.38-2.91)
|
0.27
|
INT REF
|
0.716
|
(-0.30-0.20)
|
-0.05
|
-0.858
|
(-0.08-0.09)
|
0.01
|
INT MED
|
0.465
|
(-0.25-0.11)
|
-0.07
|
0.648
|
(-0.33-0.21)
|
-0.06
|
PIE
|
0.148
|
(-0.29-1.96)
|
0.83
|
0.519
|
(-0.41-0.82)
|
0.20
|
TE
|
neonatal risk factors
|
0.160
|
(-0.45-2.75)
|
1.15
|
0.598
|
(-0.63-1.10)
|
0.24
|
CDE
|
0.441
|
(-0.77-0.33)
|
-0.22
|
0.976
|
(-0.36-0.37)
|
0.00
|
INT REF
|
0.552
|
(-0.24-0.13)
|
-0.06
|
0.976
|
(-0.06-0.06)
|
0.00
|
INT MED
|
0.437
|
(-0.14-0.06)
|
-0.04
|
0.526
|
(-0.14-0.07)
|
-0.03
|
PIE
|
1ERR=excess relative risk,2CI= confidence interval, 3TE=total effect,4CDE=controlled direct effect, 5INT REF=reference interaction, 6INT MED=mediated interaction, 7PIE=pure indirect effect
There is no significant effect for sexual violence as exposure and SMM as outcome in table 16 and 17.
Table16 Mediation analysis between sexual violence and severe maternal morbidity by MED4WAY
adjusted
|
crude
|
effect
|
mediator
|
P value
|
CI
|
OR
|
P value
|
CI2
|
OR1
|
0.510
|
(0.15-2.50)
|
0.63
|
0.377
|
(0.14-2.05)
|
0.55
|
CDE3
|
pregnancy risk factors
|
0.715
|
(0.53-2.46)
|
1.15
|
0.866
|
(0.44-1.96)
|
0.95
|
PDE4
|
0.458
|
(0.78-1.70)
|
1.16
|
0.564
|
(0.80-1.50)
|
1.10
|
TIE5
|
0.513
|
(0.56-3.15)
|
1.33
|
0.945
|
(0.44-2.36)
|
1.03
|
MTE6
|
0.492
|
(0.30-11.76)
|
1.89
|
0.683
|
(0.26-7.52)
|
1.42
|
CDE
|
delivery risk factors
|
0.462
|
(0.56-3.53)
|
1.41
|
0.651
|
(0.06-2.96)
|
1.23
|
PDE
|
0.515
|
(0.57-1.32)
|
0.87
|
0.476
|
(0.53-1.33)
|
0.85
|
TIE
|
0.639
|
(0.52-2.89)
|
1.23
|
0.929
|
(0.44-2.41)
|
1.04
|
MTE
|
0.300
|
(0.63-4.38)
|
1.69
|
0.500
|
(0.56-3.19)
|
1.35
|
CDE
|
neonatal risk factors
|
0.346
|
(0.64-3.51)
|
1.50
|
0.598
|
(0.57-2.57)
|
1.22
|
PDE
|
0.237
|
(0.60-1.13)
|
0.83
|
0.242
|
(0.64-1.11)
|
0.85
|
TIE
|
0.598
|
(0.54-2.83)
|
1.25
|
0.917
|
(0.50-2.14)
|
1.04
|
MTE
|
1OR=odds ratio, 2CI=confidence interval, 3CDE= controlled direct effect, 4PDE=pure direct effect, 5TIE=total indirect effect, 6MTE=marginal total effect
Table17 Mediation analysis between sexual violence and severe maternal morbidity by MED4WAY
adjusted
|
crude
|
effect
|
mediator
|
P value
|
CI
|
ERR
|
P value
|
CI2
|
ERR1
|
0.566
|
(-0.81-1.48)
|
0.33
|
0.946
|
(-0.82-0.88)
|
0.03
|
TE3
|
pregnancy risk factors
|
0.434
|
(-4.70-2.02)
|
-1.34
|
0.270
|
(-5.20-1.45)
|
-1.87
|
CDE4
|
0.325
|
(-1.48-4.48)
|
1.50
|
0.227
|
(-1.13-4.75)
|
1.81
|
INT REF5
|
0.469
|
(-0.90-0.41)
|
-0.24
|
0.588
|
(-0.55-0.31)
|
-0.12
|
INT MED6
|
0.308
|
(0.39-1.23)
|
0.42
|
0.549
|
(-0.47-0.89)
|
0.21
|
PIE7
|
0.671
|
(-0.82-1.28)
|
0.23
|
0.930
|
(-0.83-0.91)
|
0.04
|
TE
|
delivery risk factors
|
0.600
|
(-4.55-7.88)
|
1.66
|
0.725
|
(4.87-7.03)
|
1.07
|
CDE
|
0.635
|
(-6.40-3.90)
|
-1.25
|
0.742
|
(-5.86-4.16)
|
-0.84
|
INT REF
|
0.675
|
(-0.61-0.39)
|
-0.11
|
0.757
|
(-0.47-0.34)
|
-0.06
|
INT MED
|
0.425
|
(-0.25-0.10)
|
-0.07
|
0.381
|
(-0.39-0.15)
|
-0.12
|
PIE
|
0.717
|
(-0.77-1.12)
|
0.17
|
0.919
|
(-0.71-0.79)
|
0.04
|
TE
|
neonatal risk factors
|
0.474
|
(-1.05-2.26)
|
0.60
|
0.556
|
(0.93-1.73)
|
0.40
|
CDE
|
0.502
|
(-0.77-0.38)
|
-0.20
|
0.490
|
(-0.68-0.32)
|
-0.18
|
TDE
|
0.529
|
(-0.54-0.28)
|
-0.13
|
0.525
|
(-0.39-0.20)
|
-0.10
|
TIE
|
0.233
|
(-0.25-0.06)
|
-0.10
|
0.228
|
(-0.22-0.05)
|
-0.08
|
PIE
|
1ERR=excess relative risk,2CI= confidence interval, 3TE=total effect,4CDE=controlled direct effect, 5INT REF=reference interaction, 6INT MED=mediated interaction, 7PIE=pure indirect effect
3-3 INTERPRETATION OF RESULT
Controlled direct effect of physical and social violence on SMM mediated by neonatal risk factors shows that: if neonatal risk factors are fixed at a specific value in participants (adjust or control neonatal risk factors), physical and social violence increase the odds of SMM by 2.42(p=0.043) and 2.45(p=0.033) respectively(table 10,12).
Pure direct effect of physical violence on SMM mediated by pregnancy and neonatal risk factor show that in women who expose to physical violence but they do not have any of the pregnancy and neonatal risk factors(=0), odds ratio of SMM are 2.263 and 2.289 for them respectively(table10).
Pure direct effect of social violence on SMM mediated by pregnancy, delivery, and neonatal risk factors show that women who are exposed to physical violence but do not have any of pregnancy, delivery, and neonatal risk factors (=0), odds ratio of SMM are 2.542, 2.673 and 2.577 respectively(table12).
Pure direct effect of economicalal violence on SMM mediated by pregnancy, delivery, and neonatal risk factors show that women who are exposed to economical violence but do not have neonatal risk factors (=0), odds ratio of SMM are 1.993(table14).
Marginal total effect of physical violence on SMM mediated by pregnancy and neonatal risk factors shows that the direct effect and indirect effect of physical violence through pregnancy and neonatal risk factors increase the odds of SMM by 2.296 and 2.151, respectively(table10).
Marginal total effect of social violence on SMM mediated by pregnancy, delivery, and neonatal risk factors show that the direct effect and indirect effect of social violence through pregnancy, delivery, and neonatal risk factors increase the odds of SMM by 2.651, 2.239, and 2.673(table12).
Marginal total effect of economical violence on SMM mediated by neonatal risk factors shows that the direct effect and indirect effect of economical violence through neonatal risk factors increase the odds of SMM by 1.894(table14).
Reference interaction between physical and social violence with pregnancy risk factors increases the risk of SMM by 394.6% and 360% (even if after making interaction we prevent physical and social violence as exposure)(table11,13).
Note that the difference between pure and controlled direct effect is: in the first case, we assume that there are no mediators that can be caused by exposure, but in the second case, the mediator is assumed to be fixed at a certain value in population, this value can be the reference or other value of mediator.