Table 4.
Multivariable logistic and multinomial regression models for past-year IPV and changes in frequency relative to COVID-19 (referent = decreased IPV), per country
| Burkina Faso | Kenya | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Multivariable Logistic | Multivariable Multinomial | Multivariable Logistic | Multivariable Multinomial | |||||||||||
| Any IPV (n = 904; 25.7%) | Decreased (ref) (n = 542; 64.8%) | Increased (n = 165; 16.0%) | Sustained (n = 175; 19.2%) | Any IPV (n = 966; 23.5%) | Decreased (ref) (n = 367; 36.4%) | Increased (n = 313; 33.6%) | Sustained (n = 286; 30.0%) | |||||||
| %* | OR (95% CI) |
%* | %* | RRR (95% CI) |
%* | RRR (95% CI) |
%* | OR (95% CI) |
%* | %* | RRR (95% CI) |
%* | RRR (95% CI) |
|
| Residence | ||||||||||||||
| Urban | 29.0 | 53.8 | 22.3 | ref | 24.0 | ref | 22.6 | 27.9 | 37.7 | ref | 34.5 | ref | ||
| Rural | 24.9 | 67.9 | 14.3 |
0.53 (0.27, 1.06) |
17.8 |
0.34 (0.13, 0.88)* |
23.9 | 39.9 | 31.9 |
0.70 (0.41, 1.19) |
28.2 |
0.55 (0.30, 1.00)* |
||
| Partner dyad | ||||||||||||||
| Husband’s nights away from home in the past 12 months | ||||||||||||||
| 0 night away | 24.3 | ref | 62.0 | 18.2 | 19.8 | 20.7 | ref | 33.1 | 32.0 | 35.0 | ||||
| Less than 30 | 31.6 |
1.47 (1.09, 1.97)* |
64.3 | 14.3 | 21.4 | 30.2 |
1.65 (1.31, 2.09)*** |
40.1 | 34.6 | 25.4 | ||||
| 30 or more | 23.6 |
0.96 (0.70, 1.33) |
70.0 | 14.4 | 15.6 | 23.4 |
1.17 (0.90, 1.54) |
38.8 | 36.4 | 24.8 | ||||
| Partner education | ||||||||||||||
| None | 25.5 | 63.8 | 16.0 | ref | 20.2 | ref | 39.6 | 61.8 | 25.5 | 12.7 | ||||
| Primary | 27.2 | 61.6 | 17.6 |
0.85 (0.48, 1.53) |
20.9 |
1.00 (0.62, 1.60) |
27.2 | 35.6 | 33.6 | 30.8 | ||||
| Secondary or Higher | 24.6 | 73.7 | 14.1 |
0.37 (0.19, 0.74)** |
12.2 |
0.47 (0.25, 0.90)* |
19.1 | 33.4 | 34.9 | 31.7 | ||||
| Husband has other wives | ||||||||||||||
| No | 25.6 | 64.0 | 18.5 | 20.2 | 22.1 | ref | 35.0 | 35.0 | 30.0 | |||||
| Yes | 25.8 | 66.8 | 16.4 | 16.8 | 33.8 |
1.48 (1.13, 1.94)** |
42.8 | 27.0 | 30.0 | |||||
| Decision-making score | ||||||||||||||
| ≤ 3 | 23.4 | 66.3 | 17.1 | 16.6 | 27.3 | ref | 40.6 | 32.7 | 26.7 | |||||
| > 3 | 31.6 | 61.9 | 14.0 | 24.2 | 21.9 |
0.79 (0.64, 0.98)* |
34.1 | 34.1 | 31.8 | |||||
| Economically reliant on husband/partner for basic needs | ||||||||||||||
| No | 30.2 | ref | 64.4 | 15.9 | 19.7 | 25.1 | 35.2 | 30.6 | 34.3 | |||||
| Yes | 21.5 |
0.65 (0.49, 0.87)** |
65.4 | 16.1 | 18.5 | 22.6 | 37.2 | 35.7 | 27.1 | |||||
| Socioeconomic | ||||||||||||||
| Education | ||||||||||||||
| None | 24.1 | 66.5 | 14.1 | ref | 19.4 | ref | 31.8 | 59.7 | 26.8 |
0.81 (0.30, 2.15) |
13.6 |
0.58 (0.22, 1.56) |
||
| Primary | 28.1 | 54.6 | 18.8 |
1.39 (0.73, 2.63) |
26.7 |
1.62 (0.94, 2.79) |
27.1 | 33.1 | 34.3 | ref | 32.6 | ref | ||
| Secondary or Higher | 30.0 | 70.7 | 19.8 |
1.25 (0.51, 3.11) |
9.5 |
0.44 (0.22, 0.89)* |
19.0 | 37.6 | 34.0 |
0.67 (0.46, 0.97)* |
28.4 |
0.67 (0.42, 1.06) |
||
| Household Wealth | ||||||||||||||
| Lowest | 22.0 | 67.4 | 6.6 | ref | 26.1 | ref | 29.0 | 38.2 | 31.5 | 30.3 | ||||
| Middle | 27.3 | 66.4 | 20.0 |
3.36 (1.66, 3.78)** |
13.6 |
0.46 (0.23, 0.92)* |
22.3 | 38.5 | 30.5 | 31.0 | ||||
| Highest | 28.1 | 60.9 | 20.4 |
3.33 (1.26, 8.82)* |
18.7 |
0.56 (0.20, 1.55) |
18.1 | 29.8 | 42.2 | 28.1 | ||||
| Income loss, past year | ||||||||||||||
| None | 22.5 | 66.5 | 13.3 | 20.3 | 17.4 | ref | 43.6 | 20.9 | ref | 35.5 | ref | |||
| Partial | 28.1 | 63.3 | 18.6 | 18.1 | 22.3 |
1.68 (1.19, 2.38)** |
36.4 | 32.7 |
1.55 (0.88, 2.72) |
30.9 |
0.84 (0.42, 1.67) |
|||
| Complete | 33.4 | 63.7 | 17.5 | 18.8 | 30.4 |
2.38 (1.66, 3.42)*** |
33.4 | 40.1 |
2.03 (1.05, 3.92)* |
26.5 |
0.70 (0.35, 1.42) |
|||
Multivariable logistic regression compares any IPV experience to no IPV experience (referent). Multivariable multinomial regression compares IPV experience trajectory since COVID-19 restrictions (increased, sustained) to decreased IPV experience (referent). Model adjusted for all variables with p < 0.1 from bivariate models (not presented). Associations that remained significant (p < 0.05) after adjustment are presented. Burkina Faso logistic model adjusted for age, education, parity, household wealth tertile, worked in the last year, decision spending score, and income loss in the last 12 months. Burkina Faso multinomial model adjusted for age, education, residence, parity, household wealth tertile, level of financial knowledge, knows where to go for financial advice, and partner education. Kenya logistic model adjusted for education, household wealth tertile, savings, mobile money account, knows where to go for financial advice, financial goal, husband's nights away from home, partner education, husband has other wives, decision spending score, and income loss in last 12 months. Kenya multinomial model adjusted for education, residence, household wealth tertile, work last year, mobile money account, financial goal, husband's nights away from home, partner education, husband has other wives, decision spending score, and income loss in last 12 months. * p<0.05*