Background characteristics of individual women:
A total of 33482 women (10,721 in 2005, 11,872 in 2011, and 10, 889 in 2016) were included for this study. Overall, 94.78%, 90.05%, and 73.44% of women gave birth at home in 2005, 2011, and 2016 EDHSs respectively. From the three consecutive surveys, more than 60% of the mothers were in the age group of 20-34 years and had the same mean ± SD age of 29 ± 6.6 years. Among the three surveys, a significant number (48%) of the female household head was observed in the 2011 EDHS survey. All most all (> 90%), of the women were married in five years preceding the survey in three consecutive surveys. The educational status of the women was 79%, 69%, and 66% were unable to read and write in each survey year respectively. As well, 31%, 47%, and 56% of women were had not any work in the consecutive surveys respectively. In EDHS 2016 survey, personal mobile and health insurance status was interviewed but not in EDHS 2005 and EDHS 2011. In five years, preceding the survey 16.4% of women were had personal mobile and 3.46 % of women were insured for health insurance ( Table 1). Despite home delivery prevalence in Ethiopia was high, all regions registered in the decreasing trend from 2005 to 2016 ( Figure 1).
Table 1: Sociodemographic characteristics of women who had a live birth in the five years preceding the survey from EDHS 2005, 2011, and 2016 in Ethiopia.
Characteristics
|
EDHS years
|
2005
|
2011
|
2016
|
Frequency (%)
|
Frequency (%)
|
Frequency (%)
|
Women age
|
<20 Years
|
1,315 (12.26)
|
1132 (9.53)
|
851 (7.81)
|
20 -34 years
|
6,655 (62.07)
|
7,861 (66.22)
|
7337 (67.38)
|
35-49 Years
|
2,752 (25.67)
|
2879 (24.25)
|
2701 (24.80)
|
Mean ± SD
|
29.01±6.94
|
29.04 ± 6.63
|
29.57± 6.60
|
Household head
|
Male
|
9558 (89.15)
|
6,168 (51.95)
|
9,371 (86.06)
|
Female
|
1163 (10.85)
|
5,704 (48.05)
|
1,518 (13.94)
|
Marital status
|
Not having partner
|
488 (4.50)
90()
|
721 (6.07) 6.07
|
681 (6.25)
|
Had partner
|
10,233 (95.50)
|
11,151 (93.93) 100.00
|
10,208 (93.75)
|
Total
|
10,721 (100)
90()
|
11, 872 (100)
|
10,889 (100)
|
Religion
|
Orthodox
|
4543(42.37)
|
4,519 (38.06) 38.06
|
3,718 (34.14)
|
Muslim
|
3752(35.00)
|
4,214 (35.49) 73.56
|
4,519 (41.50)
|
Protestant
|
2138(19.91)
|
2,758 (23.23) 96.78
|
2,297 (21.09)
|
Others
|
286(2.68)
|
382 (3.22) 100.00
|
355 (3.26)
|
Residence
|
Urban
|
760 (7.09)
|
1,528 (12.87) 12.87
|
1,213 (11.14)
|
Rural
|
9961 (92.91)
|
10,344 (87.13)
|
9676 (88.86)
|
Region
|
Tigray
|
684(6.39)
|
753 (6.34)
|
701 (6.44)
|
Afar
|
104(0.97)
|
121 (1.02)
|
114 (1.05)
|
Amhara
|
2571(23.99)
|
2,656 (22.37) 29.74
|
2,041 (18.74)
|
Oromia
|
4255(39.69)
|
5,014 (42.23) 71.97
|
4,813 (44.20)
|
Somalia
|
455(4.24)
|
365 (3.07)
|
508 (4.66)
|
Benishangul Gumuz
|
83(.78)
|
140 (1.18)
|
120 (1.11)
|
SNNP
|
2338(21.81)
|
2,494 (21.01)
|
2,250 (20.67)
|
Gambela
|
29(.27)
|
40 (0.34)
|
27 (0.24)
|
Harari
|
20(.19)
|
29 (0.24)
|
26 (0.24)
|
Addis Ababa
|
141(1.32)
|
222 (1.87)
|
243 (2.33)
|
Dire Dawa
|
36(.34)
|
39 (0.33)
|
46 (0.42)
|
Women education
|
Unable to read and write
|
8502(79.31)
|
8,227(69.30) 69.30
|
7,201 (66.13)
|
Primary education
|
1772(16.53)
|
3,211 (27.05)
|
2,904 (26.67)
|
Secondary education
|
409(3.79)
|
266 (2.24)
|
510 (4.68)
|
Higher education
|
40 (.38)
|
168 (1.42)
|
274 (2.52)
|
Husband education
|
Unable to read and write
|
6308 (59.03)
|
5,966 (50.60)
|
5,018 (48.57)
|
Primary education
|
3,245 (30.37)
|
4,867 (41.26)
|
4,051 (39.21)
|
Secondary education
|
1,023 (9.58)
|
584 (4.95) 6.81
|
790 (7.65)
|
Higher education
|
109 (1.01)
|
376 (3.19)
|
471 (4.56)
|
Women occupation
|
Not working
|
7,623 (30.71)
|
5,597 (47.14)
|
6,057 (55.62)
|
Working
|
3,098 (28.89)
|
6,275 (52.86)
|
4,832 (44.38)
|
Husband occupation
|
Not working
|
215 (2.00)
|
158 (1.33)
|
6,199 (56.93)
|
Working
|
10,506 (98.00)
|
11,714 (98.67)
|
4,690 (43.07)
|
Distance to health facility
|
Big problem
|
6,597(60.58)
|
8,956 (75.43)
|
6,597 (60.58)
|
Not a big problem
|
4,292(39.42)
|
2,916 (24.57)
|
4,292 (39.42)
|
Media exposure
|
No media exposure
|
7,347(68.41)
|
8,956 (75.43)
|
7,347 (68.41)
|
Has media exposure
|
3,392(31.59)
|
2,916 (24.57)
|
3,392 (31.59)
|
Had mobile
|
No
|
NA
|
NA
|
9,103 (83.60) 83.60
|
Yes
|
NA
|
NA
|
1,786 (16.40)
|
Insurance
|
Not insured
|
NA
|
NA
|
10,512 (96.54)
|
Insured
|
NA
|
NA
|
377 (3.46)
|
Had ANC
|
No
|
5041 (71.86)
|
4,543 (57.45)
|
2,792 (37.42)
|
Yes
|
1973 (28.14)
|
3,365 (42.55)
|
4,670 (62.58)
|
Total
|
7014 (100)
|
7908 (1000)
|
7462 (100)
|
Birth order
|
1
|
2030(18.65)
|
2,261 (19.05)
|
2,030 (18.65)
|
2-4
|
4660(42.80)
|
5,179 (43.62)
|
4,661 (42.80)
|
≥5
|
4197(35.55)
|
4,432 (37.33)
|
4,198 (38.55)
|
Parity
|
≤ 2
|
3135(28.80)
|
3,469 (29.22)
|
3,136 (28.80)
|
2-5
|
4319(39.67)
|
4,710 (39.67)
|
4,319 (39.67)
|
≥ 5+
|
3433(31.53)
|
3,693 (31.11)
|
3,434 (31.53)
|
Wealth index
|
Poor
|
5091(46.67)
|
5,368 (45.22)
|
5,091 (46.76)
|
|
Middle
|
2243(20.60)
|
2,437 (20.53)
|
2,243 (20.60)
|
|
Richer
|
3554(32.64)
|
4,067 (34.26)
|
3,555 (32.64)
|
Place of delivery
|
Health institution
|
560 (5.22)
|
1180 (9.95)
|
2891 (26.56)
|
Home
|
10,161 (94.78)
|
10.691 (90.05)
|
7997 (73.44)
|
Total
|
10721 (100)
|
11 872 (100)
|
10 889 (100)
|
NB: NA = Indicates data not available in that EDHS year.
Characteristics of the cluster:
The unit of analysis for the community factors on home delivery was a clusters. In EDHS 2016, 645 clusters were selected; from these 643 clusters were eligible in which the women give birth preceding five years the survey. Among the total number of clusters, 69% were rural in residence and almost half (49%) of the clusters were had a big problem accessing any health institution. Regarding the aggregate community ANC utilization rate half of the clusters were had low community utilization. From the total of clusters, half of them were had low community women educational attainment and high community poverty status. (Table 2).
Table 2: Characteristics of clusters in EDHS 2016, Ethiopia. 2019.
Community-level characteristics
|
Frequency
|
Percent (%)
|
Residence
|
Rural
|
441
|
68.68
|
Urban
|
202
|
31.32
|
Region
|
Tigray
|
63
|
9.80
|
Afar
|
53
|
8.24
|
Amhara
|
71
|
11.04
|
Oromia
|
74
|
11.51
|
Somali
|
67
|
10.42
|
Benishangul
|
50
|
7.78
|
SNNPR
|
71
|
11.04
|
Gambela
|
50
|
7.78
|
Harari
|
44
|
6.84
|
Addis Ababa
|
56
|
8.71
|
Dire Dawa
|
44
|
6.84
|
Community ANC utilization rate
|
Low
|
320
|
49.77
|
High
|
323
|
50.23
|
Community distance to the health facility
|
Big problem
|
314
|
48.83
|
Not a big problem
|
329
|
51.17
|
Community media exposure
|
Low
|
321
|
49.92
|
High
|
322
|
50.08
|
Community poverty status
|
High
|
319
|
49.61
|
Low
|
324
|
50.39
|
Community-women educational attainment
|
Low
|
318
|
49.46
|
High
|
325
|
50.54
|
Total
|
643
|
100
|
Spatio-temporal distribution of home delivery in Ethiopia.
The spatial distribution of home delivery in Ethiopia was non-random among the three consecutive surveys. The global Moran’s I value was 0.12 (P-value < 0.001) in 2005, 0.59 (P-value < 0.001) in 2011, and 0.44 (P-value < 0.001) in 2016 Ethiopian Demographic and health survey.
Hot spot analysis of the three surveys:
The spatial distribution of home delivery in Ethiopia was different in the three survey years. In EDHS 2005, a high proportion of home delivery was detected majorly at Amhara and SNNPR regional states of Ethiopia. In EDHS 2011, High clustering of home delivery detected in most parts of Tigray, Amhara, Afar, Benishangul and SNNPR, and western part of the Oromia region of Ethiopia. Furthermore, in EDHS 2016 high proportion of home delivery detected in the southern part of Afar, Southeastern part of Amhara, SNNPR, Benishangul Gumez and Somali region of Ethiopia (Figure 2).
Spatial scan statistics analysis of the three survey years:
As shown in figure 2 below, the red window indicates the identified significant clusters inside the window. In spatial scan statistics, a total of 158 most likely clusters were identified in EDHS 2005 survey. The most likely clusters of home delivery were detected in most parts of Amhara, southwestern part of Tigray, SNNPR, and Eastern part of the Benishangul Gumez region of Ethiopia. Among the most likely clusters, 128 of them were primary clusters which are located at 11.586460 N, 37.367962 E with 290.16 km radius (Relative risk (RR) = 1.04 and Log-Likelihood ratio (LLR) = 45.75, P-value < 0.001). In EDHS 2005, mothers live in the primary cluster were 4% more likely to give birth at home than outside the window.
In EDHS 2011, a total of 127 most likely clusters were identified in spatial scan statistics which is located at Southeastern Oromia and Sothern part of the Somali region of Ethiopia. Among the most likely cluster, 90 of them were primary clusters located at 5.842888 N, 42.105068 E with 396.38 km radius. Mothers live in the primary cluster were 11% more likely to deliver at home as compared to outside the window (RR= 1.11, LLR = 161.45, P-value <0.001).
Furthermore, in the 2016 EDHS survey, 70 most likely clusters were detected spatially. The most likely clusters of home deliver were detected at southeastern Oromia and western Somali region of Ethiopia. From the most likely clusters, 55 of them were primary clusters located at 5.330795 N, 41.837597 E with 400.35 km radius. Mothers who live in the primary clusters were 29% more likely to give birth at home as compared to outside the window (RR = 1.29, LLR = 210.89, P-value < 0.001) (Figure 3).
Prevalence of home delivery in Ethiopia in the three EDHS surveys:
Based on geostatistical Kriging analysis, in 2005 EDHS exclusively Dira Dawa, Hareri, and Addis Ababa had a prevalence of less than 73%. In the 2011 EDHS survey still, there is no significant difference in-home delivery reduction compared to 2005 EDHS across the region of Ethiopia. In EDHS 2016 survey, Tigray, Addis Ababa, some parts of Oromia, Dire Dawa, and Gambela region significantly decreased home delivery in Ethiopia as compared to the other region. (Figure 4)
Multilevel analysis (random effect analysis)
Home delivery prevalence rate was not similarly distributed across the communities. About 67.32% of the variance in the odds of home delivery in women could be attributed to community-level factors, as calculated by the ICC based on estimated intercept component variance and also the variation was statistically significant (p-value <0.001). After adjusting for individual-level and community-level factors, the variation in-home delivery across communities remained statistically significant. About 89 % of the odds of home delivery variation across communities was observed in the full model (model 4). Moreover, the MOR indicated that home delivery was attributed to community-level factors. The MOR for home delivery was 12.17 in the empty model (model 1); this showed that there was variation between communities (clustering) since MOR was 12.17 times higher than the reference (MOR=1). The unexplained community variation in-home delivery decreased to MOR of 2.54 when all factors were added to the null model (empty model). This indicates that when all factors are included, the effect of clustering is still statistically significant in the full model
Individual-level predictors for home delivery:
This study evidenced that on multivariable multilevel logistic regression analysis, individual-level factors such as religion, ANC visit, wealth index, birth order, parity, distance from the health facility, insurance, and mobile access were statistically significant factors for home delivery.
Keeping all individual and community level factors constant, the odds of giving birth at home for protestant in religion followers was 1.62 times higher than orthodox followers (AOR= 1.62, (95% CI, 1.15, 2.28)). Mothers who had personal mobile were 40% less likely gave birth at home than those who had no personal mobile (AOR = 0.60, (95%CI, 0.48, 0.76)). Women who had at least one ANC visit at the health facility were 82% less likely to deliver at home than those who had no ANC visit (AOR = 0.18, (95% CI, 0.14, 0.22)). The odds of giving birth at home among women who had 2 -5 and more than five birth order were 78% and 87% than birth order less than 2 respectively (AOR = 1.78, (95% CI, 1.36, 2.34), AOR = 1.87, (95%CI, 1.27, 2.76)). Multiparous women (para 2-5) had high odds of giving birth at home by 37% as compared to para II women (AOR = 1.37, 95% CI, 1.08, 1.74)). Furthermore, mothers rich in their wealth were 22% less likely to deliver at home as compared to poor wealth status (AOR = 0.78, 95%CI, 0.61, 0.99)).
Community-level predictors for home delivery.
In the multivariable multilevel logistic regression model residence, region, community ANC utilization rate, community women education, and distance to any health institution were significantly associated with community-level factors for place delivery.
Keeping all individual and community level factors constant, women in rural clusters were three times more likely giving birth at home than women in urban clusters (AOR= 2.92, (95%, 1.99, 4.29)). Regarding regions, women live in Afar, Amhara, and Somali had high odds of giving birth at home as compared to Addis Ababa (AOR= 11.03, (95% CI, 4.29, 28.41), AOR = 6.39, (95% CI, 2.60, 15.70), AOR = 6.02, (95% CI< 2.41, 15.06)). Women live in Tigray and Dire Dawa were no significant difference in the place of delivery as compared to Addis Ababa (AOR=1.46, (95% CI, 0.59, 3.60), AOR = 2.01, (95% CI, 0.77, 5.20)). Women who live in a high community ANC utilization were 50% less likely to give birth at home than in low community ANC utilization in five years preceding the survey (AOR =0.50, (95% CI, 0.39, 0.65)). Women in a cluster had no problem to access any health institution were less likely to deliver at home by 29% those women had a problem to access health facilities(AOR= 0.69, (95%CI,0.53, 0.90)). Furthermore, high women education status in the cluster (community) were 28% less likely deliver at home than low women education attainment at the cluster in five years preceding the survey (AOR = 0.78, (95% CI, 0.60, 0.99)) (Table 3).
Table 3: Multivariable multilevel logistic regression analysis of both individual and community-level factors associated with Home Delivery in Ethiopia, EDHS 2016:
Individual and community-level variables
|
Models
|
Null model
|
Model I
|
Model II
|
Model III
|
AOR (95%CI)
|
AOR (95%CI)
|
AOR (95%CI)
|
AOR (95%CI)
|
Mother’s age
<20 Years
20 -34 years
35-49 Years
|
|
|
|
|
<20 Years
|
|
1
|
|
1
|
20 -34 years
|
|
0.79 (0.59, 1.05)
|
|
0.90 (0.67,1.20)
|
35-49 Years
|
|
0.70 (0.49, 1.01)
|
|
0.91 (0.63, 1.31)
|
Household head
|
|
|
|
|
Male
|
|
1
|
|
1
|
Female
|
|
0.89 (0.72, 1.11)
|
|
0.89 (0.72, 1.11)
|
Marital status
|
|
|
|
|
Not having a partner
|
|
1
|
|
1
|
Had a partner
|
|
1.44 (0.71, 2.90)
|
|
1.10 (0.54, 2.25)
|
Religion
|
|
|
|
|
Orthodox
|
|
1
|
|
1
|
Muslim
|
|
1.52 (1.19, 1.95)
|
|
0.93 (0.68, 1.25)
|
Protestant
|
|
2.26 (1.68, 3.02)
|
|
1.62 (1.15, 2.28) *
|
Others
|
|
2.21 (1.15, 4.24)
|
|
1.48 (0.77, 2.84)
|
Women education
|
|
|
|
|
Unable to read and write
|
|
1
|
|
1
|
Primary education
|
|
0.69 (0.57, 0.83)
|
|
0.75 (0.62, 0.90) ***
|
Secondary education
|
|
0.39 (0.27, 0.56)
|
|
0.50 (0.34, 0.72) ***
|
Higher education
|
|
0.25 (0.13,0.47)
|
|
0.34 (0.18, 0.66) ***
|
Husband education
|
|
|
|
|
Unable to read and write
|
|
1
|
|
1
|
Primary education
|
|
0.74 (0.61, 0.89)
|
|
0.83 (0.69,0.99) *
|
Secondary education
|
|
0.44 (0.53, 0.59)
|
|
0.53 (0.39, 0.70) ***
|
Higher education
|
|
0.48 (0.33, 0.69)
|
|
0.56 (0.38, 0.81) ***
|
Women occupation
|
|
|
|
|
Not working
|
|
1
|
|
1
|
Working
|
|
1.31 (0.72, 2.40)
|
|
1.47 (0.80, 2.69)
|
Husband occupation
|
|
|
|
|
Not working
|
|
1
|
|
1
|
Working
|
|
0.78 (0.42, 1.42)
|
|
0.69 (0.38, 1.26)
|
Distance to any health facility
|
|
|
|
|
Big problem
|
|
1
|
|
1
|
Not a big problem
|
|
0.64 (0.54, 0.75)
|
|
0.79 (0.66, 0.94) **
|
Media exposure
|
|
|
|
|
No media exposure
|
|
1
|
|
1
|
Has media exposure
|
|
0.80 (0.67, 0.59)
|
|
0.91 (0.75, 1.11)
|
Had mobile
|
|
|
|
|
No
|
|
1
|
|
1
|
Yes
|
|
0.48 (0.38, 0.59)
|
|
0.60 (0.48, 0.76) ***
|
Insurance
|
|
|
|
|
Not insured
|
|
1
|
|
1
|
Insured
|
|
0.75 (0.32, 1.04)
|
|
0.9 (0.88, 1.08)
|
Had ANC
|
|
|
|
|
No
|
|
1
|
|
1
|
Yes
|
|
0.14 (0.12, 0.17)
|
|
0.18 (0.14, 0.22) ***
|
Birth order
|
|
|
|
|
1
|
|
1
|
|
1
|
2-4
|
|
1.76 (1.34, 2.30)
|
|
1.78 (1.36, 2.34) ***
|
>= 5
|
|
1.81 (1.24, 2.67)
|
|
1.87 (1.27, 2.76) **
|
Parity
|
|
|
|
|
<= 2
|
|
1
|
|
1
|
2-5
|
|
1.43 (1.14, 1.81)
|
|
1.37 (1.08, 1.74) **
|
> 5+
|
|
1.34 (0.93, 1.95)
|
|
1.14 (0.78, 1.66)
|
Wealth index
|
|
|
|
|
Poor
|
|
1
|
|
1
|
Middle
|
|
0.81 (0.65, 1.02)
|
|
0.91 (0.73, 1.14)
|
Richer
|
|
0.48 (0.38, 0.59)
|
|
0.78 (0.61, 0.99) *
|
Residence
|
|
|
|
|
Urban
|
|
|
1
|
1
|
Rural
|
|
|
4.55 (3.19, 6.49)
|
2.92 (1.99, 4.29) ***
|
Region
|
|
|
|
|
Addis Ababa
|
|
|
|
1
|
Tigray
|
|
|
2.54 (1.24, 5.23)
|
1.46 (0.59, 3.60)
|
Afar
|
|
|
17.24 (8.06, 6.89)
|
11.03 (4.29, 28.41) ***
|
Amhara
|
|
|
8.23 (3.98, 17.00)
|
6.39(2.60, 15.70) ***
|
Oromia
|
|
|
7.07 (3.42, 14.62)
|
3.69 (1.50, 9.08) **
|
Somali
|
|
|
10.66 (5.17, 1.98)
|
6.02(2.41, 15.06) ***
|
Benishangul
|
|
|
6.09 (2.90, 12.81)
|
3.94 (1.59, 9.82) **
|
SNNPR
|
|
|
5.75 (2.80, 11.77)
|
3.04 (1.23, 7.47) *
|
Gambla
|
|
|
6.22 (2.97, 13.02)
|
4.48 (1.77, 11.32) **
|
Harari
|
|
|
3.59 (1.69, 7.64)
|
2.68 (1.04, 6.89) *
|
Dire Dawa
|
|
|
2.73 (1.28, 5.84)
|
2.01 (0.77, 5.20)
|
Community ANC utilization
|
|
|
|
|
Low
|
|
|
1
|
1
|
High
|
|
|
0.32 (0.25, 0.41)
|
0.50 (0.39, 0.65) ***
|
Community Health facility distance
|
|
|
|
|
Big problem
|
|
|
1
|
1
|
Not a big problem
|
|
|
0.56 (0.43, 0.71)
|
0.69 (0.53, 0.90) **
|
Community media exposure
|
|
|
|
|
Low
|
|
|
1
|
1
|
High
|
|
|
0.85 (0.65, 1.09)
|
1.11 (0.85, 1.45)
|
Community poverty status
|
|
|
|
|
High
|
|
|
1
|
1
|
Low
|
|
|
0.58 (0.44, 0.75)
|
0.78 (0.58, 1.04)
|
Community-women education
|
|
|
|
|
Low
|
|
|
1
|
1
|
High
|
|
|
0.55 (0.43, 0.69)
|
0.78 (0.60, 0.99) *
|
Random effects
|
|
|
|
|
ICC%
|
67.32
|
|
|
|
PCV%
|
1
|
82
|
86
|
89
|
MOR
|
12.17
|
4.75
|
3.20
|
2.54
|
Model comparison
|
|
|
|
|
Log-likelihood ratio
|
-4749.96
|
-2606.26
|
-4303.85
|
-2450.68
|
NB: * = Significant atP-value < 0.05, ** = Significant at P-value , 0.01, *** = Significant at P-value 0.001 , CI = Confidence Interval, AOR = Adjusted Odds Ratio,