Literature identification and Descriptions of studies
Two hundred duplicates were discovered after examining the titles and abstracts of the 440 papers found. We also excluded 80 studies because they failed to mention LARC in Ethiopia. A total of 160 studies were evaluated based on their themes and abstracts; with 60 being removed since they were presented in mixed method (includes all modern contraception). Finally, the full text of the remaining 75 publications was downloaded and extensively checked to confirm that they complied with all of the criteria for inclusion. We also eliminated 35 articles because they were reported prevalence of early discontinuation of LARC, ten publications because they didn’t reflect the prevalence of LARC and three publications because they reported the rationale for implanon early termination. Finally, twenty-seven studies met the review's inclusion criteria (11–37) (Fig. 1).
Fourteen studies were conducted at health facilities (11, 12, 14, 15, 20, 22, 23, 26, 29–31, 34, 35, 37),while thirteen studies were conducted at the community level(13, 16–19, 21, 24, 25, 27, 28, 32, 33, 36)and all of the studies were cross sectional in design (11–37). The studies included our review were published between 2017 and 2021(11–37).
The study subjects in twenty-four investigations were women of reproductive age ((11, 13–16, 18–33, 35–37), three were about postpartum women (12, 28, 34), one study was about HIV/AIDS-positive women of reproductive age(20), and one study was about female health professionals of reproductive age (15). The average age of study participants was ranged from 25(33) to 36(37) years old,with a range of 15–49 years (17, 19). Six studies were done in the Oromia region(12, 19, 23, 24, 29, 31), nine studies in the Amhara region(13–15, 17, 20, 21, 26, 35, 37)and nine studies in SNNP region (16, 18, 19, 22, 27, 28, 33, 34, 36), and one study each in the Afar region (11), Harari region(30) ,and Addis Ababa city administration(32). Sample size of included studies ranges from 262(27) to 731(36) and prevalence of LARC users among Ethiopian women over the last five years ranges from 12.6%(21) to 53.2%(12). Twenty-four studies (13–15, 17, 19–37)report IUCD and implanon prevalence, with just one study reporting Jadelle prevalence in addition to IUCD and implanon(12). Twenty-six studies were carried out in urban (11–34, 36, 37),while only one study was carried out in rural areas(35) (Table 1).
Table 1
Characteristics of studies included in systematic review and meta-analysis, 2022
Author | Year | Region | Age | Sample size | Number LARC user | Prevalence (%) | Newcastle Ottawa scale |
Tebeje(31) | 2017 | Oromia | CD | 422 | 97 | 22.9 | 6 |
Mesfin(28) | 2021 | SNNP | 28.2 | 416 | 94 | 22.6 | 7 |
Shiferaw(30) | 2017 | Harar | 26.7 | 402 | 153 | 38 | 6 |
Wariyo(12) | 2019 | Oromia | 27 | 393 | 209 | 53.2 | 4 |
Woldu(34) | 2020 | SNNP | 27 | 381 | 139 | 36.5 | 6 |
Biza(17) | 2017 | Amhara | CD | 381 | 111 | 29.1 | 8 |
Gujo(22) | 2021 | SNNP | 27.1 | 376 | 142 | 37.8 | 6 |
Halil(25) | 2020 | SNNP | 28.8 | 528 | 97 | 18.3 | 5 |
Mohamed(29) | 2017 | Oromia | 25.91 | 644 | 180 | 27.9 | 8 |
Fekadu(19) | 2017 | Oromia | CD | 693 | 210 | 30.3 | 8 |
Asteray(14) | 2019 | Amhara | 26 | 359 | 66 | 18.4 | 6 |
Abdu(11) | 2019 | Afar | 26.71 | 524 | 175 | 33.4 | 8 |
Asirat(27) | 2017 | SNNP | 30.4 | 262 | 66 | 25.2 | 7 |
Mariam(26) | 2018 | Amhara | 25.9 | 614 | 207 | 33.7 | 7 |
Hibstu(23) | 2020 | Oromia | 27.4 | 361 | 121 | 33.5 | 8 |
Zenebe(37) | 2017 | Amhara | 25.91 | 644 | 180 | 27.9 | 8 |
Dereje(16) | 2019 | SNNP | 30 | 561 | 164 | 29.2 | 8 |
Samson(18) | 2020 | SNNP | NA | 346 | 134 | 38.7 | 6 |
Aklilu(33) | 2020 | SNNP | 25 | 660 | 145 | 22 | 5 |
Getahun(21) | 2018 | Amhara | 27.8 | 730 | 92 | 12.6 | 4 |
Gelagay(20) | 2018 | Amhara | 31.59 | 505 | 138 | 27.3 | 8 |
Teshome(32) | 2018 | Addis Ababa | 31.2 | 469 | 81 | 17.3 | 5 |
Alemu(35) | 2021 | Amhara | 30.1 | 356 | 112 | 32.3 | 8 |
Yemaneh(36) | 2017 | SNNP | 26.6 | 731 | 145 | 19.8 | 5 |
Asigedew(13) | 2017 | Amhara | 31.69 | 499 | 174 | 34.9 | 7 |
Liknaw(15) | 2019 | Amhara | 27.4 | 392 | 89 | 22.7 | 6 |
Abebe(24) | 2017 | Oromia | 30.5 | 419 | 139 | 33.2 | 7 |
CD...age in categorical variables, NA...not available |
Pooled prevalence estimate and heterogeneity
The prevalence of LARC among Ethiopian women was reported in twenty -seven (27) studies including 12,741 women. As shown in Fig. 2, among the 3,588 women used LARC, the pooled prevalence was estimated to be 29% (95% CI 26–32%). The prevalence estimate differed significantly between studies (X2 = 505.29, P = 0.001, I2 = 94.85%) (Fig. 2)
Subgroup analysis based on the type of LARC
1. Pooled estimates of prevalence of Implanon among Ethiopian women
The prevalence of Implanon among women in Ethiopia was reported in twenty-six (26) studies including 11,871 women. As shown in Fig. 3, 3,588 women used LARC and the pooled prevalence of Implanon was estimated to be 23% (95% CI 20–26%). The prevalence estimate differed significantly between studies (X2 = 337.79, P = 0.001, I2 = 92.6%) (Fig. 3)
2. Pooled estimates of prevalence of IUCD among Ethiopian women
In Ethiopia, the prevalence of IUCD among women in Ethiopia was reported in twenty -six (26) studies including 11,871 women. As shown in Fig. 4, among the 3,588 women used LARC, the pooled prevalence of IUCD was estimated to be 5% (95% CI 4–7%). The prevalence estimate differed significantly between studies (X2 = 506.36, P = 0.001, I2 = 95.26%) (Fig. 4)
Subgroup analysis
LARC users were detected in 33% (95% CI 26–41%) of Oromia, 28% (95% 23–33) of SNNP region, and 27% (95% CI 21–33%) of Amhara region. We looked at sample sizes of ≤ 400 and > 400 ,since we assumed sample size may be a source of the considerable heterogeneity we identified in our study, and we found that sample size was not a major driver of heterogeneity (p = 0.001). Between 2017 and 2018, the estimated pooled prevalence of LARC users was 28% (95% CI 23–32%), and between 2019 and 2021, the estimated pooled prevalence of LARC users was 30% (95% CI 25–36%), with no major source of heterogeneity. We also ran an analysis to see if the source of heterogeneity was related to the study setting, and we found that the study setting had no bearing on the cause of heterogeneity (P = 0.001)(Table 2).
Table 2
Sub group analysis of the prevalence of LARC users among Ethiopian, 2022
Variables | Variable categories | Included studies | Pre valence (95% CI) | I2 | P Values |
Study setting | Community | 13 | 25(21–30) | 94.48 | 0.001 |
Facility | 14 | 32(28–36) | 92.75 | 0.001 |
Region | Oromia | 6 | 33(26–41) | 94.84 | 0.001 |
Amhara | 9 | 27(21–33) | 95.48 | 0.001 |
SNNP | 9 | 28(23–33) | 92.89 | 0.001 |
Afar | 1 | 33(29–38) | - | - |
Harari | 1 | 38(33–43) | - | - |
Addis Ababa | 1 | 17(14–21) | - | - |
Sample size | > 400 | 16 | 26(23–30 | 94.49 | 0.00 |
≤ 400 | 11 | 33(27–38) | 93.69 | 0.00 |
Year of publication | 2017–2018 | 14 | 28(23–32) | 94.73 | 0.00 |
2019–2021 | 13 | 30(25–36) | 94.99 | 0.00 |
Year of publication (p = 0.83), regional state (p = 0.58), sample size (p = 0.85), and study setting (P = 0.1) did not exhibit any statistically significant association with prevalence of LARC usage in a univariate meta-regression (Table 3).
Table 3
Univariate meta-regression of factors related to heterogeneity of prevalence of LARC users among Ethiopian women, 2022
Variables | β coefficient | 95% CI | P values |
Regional states | 2.8418 | -0.7-13.2 | 0.576 |
Study setting | -25.49 | -56.5-5.5 | 0.103 |
Sample size | -4.39 | -50.89-42.11 | 0.847 |
Year of publication | -1.21 | -12-9.99 | 0.826 |
Leave out one sensitivity analysis
By systematically removing each study from the pooled prevalence of LARC users among Ethiopian women, a leave-one analysis was performed to assess the influence of each study on the pooled prevalence of LARC users among Ethiopian women. The omitted studies had no significant impact on the projected prevalence of LARC users among Ethiopian women, according to the findings (Table 4).
Table 4
leave out one analysisof the prevalence of LARC users among Ethiopian women for the last five years, 2022
Study omitted | Estimated pool prevalence with 95% CI | Study omitted | Estimated pool prevalence with 95% CI |
Wariyo (2019) | 28(25–30) | Zenebe (2017) | 29 (26–31) |
Tebeje (2017) | 29(26–32) | Dereje (2019) | 29 (26–32) |
Shiferaw (2017) | 29 (26–31) | Samson (2020) | 29 (26–32) |
Woldu (2020) | 29 (26–31) | Aklilu (2020) | 29 (26–32) |
Biza (2017) | 29 (26–32) | Getahun (2018) | 29 (27–32) |
Gujo (2021) | 29 (26–31) | Gelagay (2018) | 29 (26–32) |
Halil(2020) | 29 (27–31) | Teshome (2018) | 29 (27–32) |
Mohammad (2017) | 29 (26–32) | Alemu (2021) | 29 (27–32) |
Fekadu (2017) | 29 (26–32) | Yemaneh (2017) | 29 (26–32) |
Asteray (2019) | 29 (26–32) | Asigedew (2017) | 29 (26–31) |
Abdu (2019) | 29 (26–31) | Liknaw (2019) | 29 (26–32) |
Asirat (2017) | 29 (26–32) | Abebe (2017) | 29 (26–31) |
Mariam (2018) | 29 (26–31) | Hibstu (2020) | 29 (26–31) |
Publication bias assessment
The visual assessment of the asymmetry of the funnel plot revealed the presence of publication bias, which was statistically confirmed by Egger's regression test (β = 16.45, 95% CI 10.86–22.04) p = 0.001) (Fig. 5).
Factors associated with LARC use
As shown in the Table 5, a summary of the data collected to determine the associated variables among LARC users among Ethiopian women. When compared to poor knowledge and a negative attitude, having good knowledge(p = 0.001,95% CI 1.16–1.98) and a positive attitude(p = 0.001,95% CI 1.5–2.3) were approximately two times more likely to use LARC. When compared to those who had never used LARC before, former users were approximately four times (P = 0.001, 3.56, 95% CI 2.55–4.57) more likely to utilize it again. Age of women 25–34 less likely to use LARC than those aged below 25 or above 35 years (P = 0.001, 95% CI 0.24–0.65). Occupation, discussion with husband, health provider counselling, parity, and educational level, while being included in many articles, but had no association with LARC usage in our review (Table 5).
Table5. Analysis of factors associated with use of LARC among Ethiopian women for the last five years, 2022
Factors
|
Comparison group
|
Number
of
studies
|
OR (95%)
|
P values
|
Heterogeneity
|
Publication bias
(Egg’s test)
|
|
Q value
|
Df
|
I2
|
P values
|
|
P values
|
|
Parity
|
≥ 5 vs.<5
|
5
|
1.09
(0.4-1.8)
|
0.669
|
2.36
|
4
|
0
|
0.002
|
0.137
|
|
Knowled-ge
|
Good knowledge vs. poor knowledge
|
11
|
1.6
(1.16-1.98)
|
0.001
|
16.54
|
10
|
39.6
|
0.085
|
0.001
|
|
Attitude
|
Positive vs. negative
|
10
|
1.9(1.5-2.3)
|
0.001
|
12.36
|
9
|
35.3
|
0.136
|
0.001
|
|
Previous use of LARC
|
Yes vs. No
|
5
|
3.56(2.55-4.57)
|
0.001
|
5.25
|
4
|
23.7
|
0.263
|
0.161
|
|
Health provider counselling
|
Yes vs. No
|
8
|
1.18(0.85-1.5)
|
|
21.52
|
7
|
67.5
|
0.003
|
0.002
|
|
Educational level
|
Diploma and above vs. others*
|
9
|
1.94(0.98-2.8)
|
|
4.85
|
8
|
0
|
0.773
|
0.039
|
|
Age
|
25-34 years vs. others**
|
8
|
0.44
(0.26-0.65)
|
0.005
|
20.6
|
7
|
65.1
|
0.005
|
0.005
|
|
Discussion with husband
|
Yes vs. No
|
8
|
0.93
(0.58-1.28)
|
|
25.41
|
7
|
72.4
|
0.001
|
0.009
|
|
Occupation
|
Governmental employee vs. others***
|
9
|
1.22
(0.8-1.65)
|
0.204
|
10.96
|
8
|
27
|
0.201
|
0.033
|
|
Others* -cannot read and write, primary education, secondary education
Others**-age below 24 or above 35
Others***- housewives, students, merchants, private employee, daily labourer