Abstract
Purpose:
In sub-Saharan Africa, sexually active adolescent girls and young women (AGYW) experience high rates of intimate partner violence (IPV) and low levels of contraceptive use, but the effect of IPV on contraceptive use is not well understood.
Methods:
In the Girl Power-Malawi study, AGYW age 15–24 were recruited from four health centers in Lilongwe, Malawi and followed for one year. At baseline, AGYW were assessed for IPV using the modified Conflict Tactics Scale. AGYW reported contraceptive method use at 6 and 12 months, characterized as barrier, non-barrier, or any modern method. Modified Poisson regression was implemented to estimate risk ratios (RRs) and 95% confidence intervals (CI) to examine the effect of IPV on contraceptive use.
Results:
One thousand AGYW were enrolled, and 954 non-pregnant participants were included. Baseline prevalence of IPV with the most recent partner was 35.5% (physical), 46.2% (sexual), and 76.9% (emotional). Baseline IPV did not affect contraceptive use at 6 months (aRR (95% CI): physical 0.98 (0.91–1.05); sexual 1.00 (0.94–1.07); emotional 1.03 (0.94–1.12)) or 12 months: physical 0.95 (0.89–1.02); sexual 0.96 (0.90–1.02); emotional 0.98 (0.91–1.05)). None of the three IPV categories affected contraceptive use when the outcome was restricted to either barrier or non-barrier methods.
Conclusions:
In this cohort, IPV was not a key driver of contraceptive use in longitudinal analyses. Interventions are needed to address the alarming rates of IPV in this population but addressing IPV alone may be insufficient to address low contraceptive use and multifaceted youth friendly health services (YFHS) may be necessary.
Keywords: Contraception, Intimate Partner Violence, Adolescent Health, Global Health, Reproductive Health Services
Introduction
The United Nations defines intimate partner violence (IPV) as behavior by an intimate partner or ex-partner that causes physical, sexual, or psychological harm. Prevalence of IPV is particularly high in sub-Saharan Africa (SSA), where intimate relationships are often characterized by severe gender inequality, with norms favoring male control of sexual intercourse and limited power among women for sexual decision making.1 In SSA, rates of IPV among AGYW range from 7.8% to 41.2%.2 Prevalence in Malawi specifically varies between studies with the UN reporting a lifetime prevalence of 38% and a separate study that focused on adolescents, which found a prevalence of 27%.3, 4
IPV is associated with multiple negative reproductive health outcomes including increased HIV and STI incidence, unintended pregnancy, induced abortions, and worse perinatal outcomes.5–8 Access to effective contraception can play a role in preventing these adverse health outcomes, yet women in violent relationships may face partner pressures limiting contraception use.9,10 A lower prevalence of contraception use among women experiencing IPV has been consistently observed in much of the global health literature11–13, but the findings in SSA are less consistent with some studies finding a higher prevalence of contraceptive use among women experiencing IPV.14–18
However, many studies on this topic in SSA have several methodologic limitations. These included a cross sectional study design which limits the ability to infer causality14–28, lack of disaggregation by violence type19,21,23,27,29, lack of disaggregation by type of modern contraception used9,14,15,17–22,25–28,30, and use of non-standardized violence surveys.17,19,24,28 Prior studies have concluded that these methodologic limitations may account for some of the variance in findings between Africa and the rest of the global literature.
Furthermore, research on IPV outcomes typically focus on married or adult women, with minimal data available on the experiences of adolescent girls and young women (AGYW) who face different relationship pressures, fertility expectations, and norms surrounding contraceptive use.31,32 Understanding the effect of IPV on contraception among young women is especially important in light of DREAMS (Determined, Resilient, Empowered, AIDS-free, Mentored and Safe), a multi-billion dollar initiative aimed at multi-layered HIV prevention. DREAMS targets gender norms with hypothesized downstream consequences of decreased IPV and increased contraceptive use.33 However, no prior studies have focused on the unique experiences of AGYW with respect to the impact of IPV on contraceptive use.
Girl Power-Malawi34, a study examining sexual and reproductive health service uptake among AGYW, provides an ideal longitudinal data source to examine this question. In this analysis we sought to examine the effect of IPV on barrier and non-barrier contraceptive use in a young cohort, aiming to understand this complex question.
Methods
Ethical Approvals
Girl Power-Malawi was approved by the Institutional Review Board (IRB) at the University of North Carolina at Chapel Hill and the Malawi National Health Sciences Research Committee. Additional approval for data sharing and to conduct this secondary analysis was obtained from the Duke University IRB. Written informed consent was provided by AGYW age 18–24 years and assent was provided by from AGYW age 15–17 years. Written informed consent was also obtained from a parent, guardian, or authorized representative for AGYW age 15–17.
Study Design, Setting, and Population
The Girl Power study was a multisite quasi experimental study conducted in Lilongwe, Malawi and Western Cape, South Africa, and was designed to assess the impact of four different care models that included combinations of youth friendly integrated services, behavioral sessions, and cash transfers on clinical and behavioral outcomes of AGYW. Methods of the study have been described in more detail in previous papers.34,35 We treated all participants as a single prospective cohort for this analysis. This analysis focuses on the four clinics in Malawi.
Participants were recruited on a rolling basis at each clinic from February-August 2016, with follow up visits occurring through August 2017. Two-hundred fifty AGYW were recruited at each clinic for a total sample size of 1000. Eligibility included female sex, age 15–24 years, from the clinic’s catchment area, and ability to participate for one year. While sexual activity was not a formal inclusion criterion, informal conversations with participants prior to being invited to participate inquired about their sexual activity to determine if they would be a good candidate for clinic participation. Therefore nearly all participants were currently or recently sexually active.
Data collection
Behavioral surveys was administered at baseline, 6, and 12 months. All surveys were conducted by trained young female research staff in the local language Chichewa. Data were recorded on tablets using Open Data Kit software.
Study Exposures, Outcomes, and Covariates
This analysis examined the effect of IPV at baseline on contraception use at 6 and 12 months.
Exposure
The primary exposures of interest were types of IPV, which were assessed at baseline using a 17-item survey, modified from the Conflict Tactics Scale.36,37 The scale contained six items for emotional violence, six for physical violence, two for sexual violence, and three for controlling behavior. Questions assessed experiences of these behaviors at any point in time with their most recent sexual partner. IPV was only assessed at baseline as it was not an outcome analyzed in the original Girl Power study. In this analysis, we focused on any experiences of physical, sexual, and emotional violence, analyzed separately by category.
Of the 14 questions, 13 provided binary response options. For the 14th question, which asked about frequency of feeling fearful of a partner, we dichotomized answers “yes all the time” and “sometimes” as positive responses and “not usually” and “no, never” as negative responses. For this analysis, we created dichotomous variables for each violence category (physical, sexual, emotional) where an affirmative response to one or more questions in the category was considered a positive exposure for that violence category.
Outcomes
The primary outcomes of interest in this analysis were self-reported use of contraceptive methods at the 6- and 12-month follow-up visits. At each visit, participants were asked about their current use of contraceptive methods, including male and female condoms, intrauterine device (IUD), implant, Depo Provera, and oral contraceptives (OCs). The primary outcome was use of any of these contraceptive methods. Additional dichotomous variables for barrier and non-barrier methods were also used. Those who reported using male or female condoms were considered users of barrier methods and those who reported IUD, implant, Depo Provera, or OCs were considered users of non-barrier methods.
Covariates
Frequency (n) and prevalence (%) were reported for baseline demographic and clinical characteristics for each category of violence (Table 1). These variables included age, education, marital status, cohabitation, employment and socioeconomic status, pregnancy and parity, HIV testing and status, sexual behaviors, and contraceptive use. Contraception use was defined as current use at baseline and analyzed as any use, barrier, or non-barrier methods. A chi squared test was conducted to examine the association between the potential covariate and each violence category at baseline. Covariates were selected based on associations seen at baseline and important confounders in the relationship between IPV and contraception.
Table 1:
Baseline Demographic and Clinical Characteristics by Violence Category (n=954)
| Physical Violence | Sexual Violence | Emotional Violence | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Yes (n=339) | No (n=615 ) | Chi 2 | Yes (n=441) | No (n=513 ) | Chi 2 | Yes (n=734) | No (n=220 ) | Chi 2 | |
| n (%) | n (%) | P | n (%) | n (%) | P | n (%) | n (%) | P | |
| Demographics | |||||||||
| Age, years | |||||||||
| 15–19 | 191 (56.3) | 363 (59.0) | 248 (56.2) | 306 (59.6) | 407 (55.4) | 147 (66.8) | |||
| 20–24 | 148 (43.7) | 252 (41.0) | 0.42 2 | 193 (43.8) | 207 (40.4) | 0.2 87 | 327 (44.6) | 73 (33.2) | 0.00 3 |
| Education | |||||||||
| Primary complete | 226 (66.7) | 446 (72.5) | 306 (69.4) | 366 (71.3) | 512 (69.8) | 160 (72.7) | |||
| Primary incomplete | 113 (33.3) | 160 (26.0) | 0.02 4 | 135 (3 0.6) | 138 (26.9) | 0.2 74 | 217 (29.6) | 56 (25.5) | 0.27 4 |
| Marital status | |||||||||
| Single | 228 (67.3) | 469 (76.3) | 302 (68.5) | 395 (77.0) | 519 (70.7) | 178 (80.9) | |||
| Married | 69 (20.4) | 119 (19.3, | 99 (22.4) | 89 (17.3) | 151 (20.6) | 37 (16.8) | |||
| Separated/Divorce d/Widowed | 41 (12.1) | 27 (4.4) | <0.0 01 | 40 (9.1) | 28 (5.5) | 0.0 07 | 63 (8.6) | 5 (2.3) | 0.00 1 |
| Employment | |||||||||
| Working for pay | 62 (18.3) | 103 (16.7) | 82 (18.6) | 83 (16.2) | 134 (18.3) | 31 (14.1) | |||
| Full time student | 71 (20.9) | 209 (34.0) | 121 (27.4) | 159 (31.0) | 189 (25.7) | 91 (41.4) | |||
| Unemployed | 171 (50.4) | 250 (40.7) | 199 (45.1) | 222 (43.3) | 349 (47.5) | 72 (32.7) | |||
| Other | 32 (9.4) | 44 (7.2) | <0.0 01 | 34 (7.7) | 42 (8.2) | 0.5 58 | 56 (7.6) | 20 (9.1) | <0.0 01 |
| Source of money | |||||||||
| Boyfriend, partner, or husband | 87 (25.7) | 112 (18.2) | 93 (21.1) | 106 (20.7) | 167 (22.8) | 32 (14.5) | |||
| Other | 187 (55.2) | 410 (66.7) | 0.00 1 | 276 (62.6) | 321 (62.6) | 0.9 02 | 455 (62.0) | 142 (64.5) | 0.02 3 |
| Asset index | |||||||||
| 0–3 | 192 (56.6) | 333 (54.1) | 264 (59.9) | 261 (50.9) | 399 (54.4) | 126 (57.3) | |||
| 4–6 | 90 (26.5) | 163 (26.5) | 106 (24.0) | 147 (28.7) | 202 (27.5) | 51 (23.2) | |||
| >6 | 57 (16.8) | 119 (19.3) | 0.60 4 | 71 (16.1) | 105 (20.5) | 0.0 20 | 133 (18.1) | 43 (19.5) | 0.43 8 |
| Clinical characteristics | |||||||||
| Contraceptive method at baseline | |||||||||
| Barrier method | |||||||||
| Yes | 223 (65.8) | 416 (67.6) | 280 (63.5) | 359 (70.0) | 485 (66.1) | 154 (70.0) | |||
| No | 116 (34.2) | 199 (32.4) | 0.55 9 | 161 (36.5) | 154 (30.0) | 0.0 34 | 249 (33.9) | 66 (30.0) | 0.27 8 |
| Non-barrier method | |||||||||
| Yes | 106 (31.3) | 158 (25.7) | 138 (31.3) | 126 (24.6) | 218 (29.7) | 46 (20.9) | |||
| No | 233 (68.7) | 457 (74.3) | 0.06 5 | 303 (68.7) | 387 (75.4) | 0.0 21 | 516 (70.3) | 174 (79.1) | 0.01 1 |
| Any method | |||||||||
| Yes | 273 (80.5) | 502 (81.6) | 352 (79.8) | 4 23 (82.5) | 595 (81.1) | 180 (81.8) | |||
| No | 66 (19.5) | 113 (18.4) | 0.67 8 | 89 (20.2) | 9 0 (17.5) | 0.2 98 | 139 (18.9) | 40 (18.2) | 0.80 1 |
| Prior pregnancy | |||||||||
| Yes | 160 (47.2) | 228 (37.1) | 201 (45.6) | 187 (36.5) | 325 (44.3) | 63 (28.6) | |||
| No | 177 (52.2) | 385 (62.6) | 0.00 2 | 237 (53.7) | 325 (63.4) | 0.0 03 | 405 (55.2) | 157 (71.4) | <0.0 01 |
| Number of living children | |||||||||
| 0 | 204 (60.2) | 402 (65.4) | 264 (59.9) | 342 (66.7) | 442 (60.2) | 164 (74.5) | |||
| 1 | 96 (28.3) | 184 (29.9) | 140 (31.7) | 140 (27.3) | 232 (31.6) | 48 (21.8) | |||
| >2 | 39 (11.5) | 29 (4.7) | <0.0 01 | 37 (8.4) | 31 (6.0) | 0.0 76 | 60 (8.2) | 8 (3.6) | <0.0 01 |
| HIV status | |||||||||
| Positive | 13 (3.8) | 13 (2.1) | 13 (2.9) | 13 (2.5) | 23 (3.1) | 3 (1.4) | |||
| Negative | 287 (84.7) | 506 (82.3) | 371 (84.1) | 422 (82.3) | 630 (85.9) | 163 (74.1) | |||
| Unknown/never tested | 39 (14.2) | 96 (15.6) | 0.07 7 | 57 (12.9) | 78 (15.2) | 0.5 71 | 82 (11.2) | 54 (24.5) | <0.0 01 |
| Sexual behaviors | |||||||||
| Age at sexual debut | |||||||||
| <15 | 82 (24.2) | 88 (14.3) | 82 (18.6) | 88 (17.2) | 137 (18.7) | 33 (15.0) | |||
| >15 | 256 (75.5) | 523 (85.0) | <0.0 01 | 356 (80.7) | 423 (82.5) | 0.5 48 | 592 (80.7) | 187 (85.0) | 0.19 9 |
| Lifetime sexual partners | |||||||||
| 1 | 117 (34.5) | 305 (49.6) | 180 (40.8) | 242 (47.2) | 286 (39.0) | 136 (61.8) | |||
| 2–3 | 164 (48.4) | 241 (39.2) | 196 (44.4) | 209 (40.7) | 333 (45.4) | 72 (32.7) | |||
| 3–10 | 48 (14.2) | 60 (9.8) | 53 (12.0) | 55 (10.7) | 98 (13.4) | 10 (4.5) | |||
| >10 | 10 (2.9) | 7 (1.1) | <0.001 | 10 (2.3) | 7 (1.4) | 0.1 53 | 15 (2.0) | 2 (0.9) | <0.001 |
Data Analysis
Primary analysis
All analyses were conducted using Stata SE 16.1. The primary analysis assessed the effect of each type of baseline intimate partner violence on any method, barrier method, and non-barrier method use at 6- and 12-month follow up. A modified poisson regression with robust error variance38 was used to estimate relative risks (RR) and 95% confidence intervals (CI) for the effect of IPV on contraception use. We controlled for marital status, parity, study arm, and contraception use at baseline, which were all considered potential confounders.
Secondary analyses
To further explore the effect of IPV on contraception, we conducted two additional analyses. We conducted stratified analyses by baseline marital status (single versus married) as these subpopulations commonly have differences in contraceptive access and fertility intention. Finally, we conducted additional analyses stratified by contraceptive use at baseline to assess differences in contraceptive uptake versus continuation. RRs, aRRs and 95% CIs were calculated in the same manner as the primary analysis with the exception of stratification by marital status and contraception use at baseline not additionally adjusted for these covariates.
Results
Study Population
Two hundred and fifty AGYW were recruited at each of the four clinics for a total of 1000 participants. Women who were pregnant at baseline (n=38), had not experienced sexual debut (n=5), or were missing data on IPV (n=3) were excluded from the analysis, for a total sample size of n=954.
Baseline characteristics
Baseline demographic characteristics are presented in Table 1, stratified by IPV exposure for each category of violence. Of the 954 participants, 339 (35.5%) had experienced physical violence, 441 (46.2%) had experienced sexual violence, and 734 (76.9%) had experienced emotional violence with their most recent partner.
Mean age was 19.1 years (standard deviation=2.5 years). One hundred eighty-eight (19.7%) were married at baseline, and 346 (36.3%) had at least one child. At baseline, married women were more likely to experience all three categories of violence (physical, n=953, 36.7% vs 32.7%, p<0.001; sexual, n=953, 52.7% vs 43.3%, p=0.007; emotional, n=953, 80.3% vs 74.4%, p=0.001). Having at least one child at baseline was associated with increased physical (n=954, 38.8% vs 33.7%, p<0.001) and emotional violence (n=954, 83.9% vs 72.9%, p<0.001).
At baseline, 81.2% of participants reported using at least one form of contraception; 67.0% reported using a barrier method, 27.7% reported using a non-barrier method, and 13.4% reported using both. Male condom use represented 95.7% of barrier method use, with an additional 3.5% reporting male and female condom use, and only 0.8% reporting only female condom use. For those using non-barrier methods, method choice was primarily Depo Provera (60.2%), OCs (52.3%), or an implant (35.0%), with only one person using an IUD. At baseline, sexual violence was associated with lower prevalence of barrier method use (n=954, 63.5% vs 70%, p=0.034) but a higher prevalence non-barrier method use (n=954, 31.3% vs 24.6%, p=0.021). Emotional violence was associated with higher prevalence of non-barrier method use (n=954, 29.7% vs 20.9%, p=0.011). Use of any contraceptive method was not associated with any violence category at baseline.
Relationship between IPV and Contraception use at 6 and 12 months
Study retention was 84.4% at 6 months and 86.7% at 12 months, with 92.6% completing at least one follow up visit. Loss to follow up did not differ by any violence category.
IPV did not affect 6- and 12-month contraceptive use for any violence category in all adjusted and unadjusted models (Table 2, Figure 1). For physical violence, the aRR was 0.98 (95% CI: 0.91–1.05) at 6 months and 0.95 (95% CI: 0.89–1.02) at 12 months. For sexual violence, the aRR was 0.98 (95% CI: 0.92–1.06) at 6 months and 0.96 (95% CI: 0.90–1.02) at 12 months. For emotional violence, the aRR was 1.03 (95% CI: 0.94–1.12) at 6 months and 0.98 (95% CI: 0.91–1.05) at 12 months.
Table 2:
Comparison of contraceptive use by violence category
| 6-month | 12-month | |||
|---|---|---|---|---|
| RR (95% CI) | aRR (95% CI) | RR (95% CI) | aRR (95% CI) | |
| Any contraception | ||||
| Physical | 0.98 (0.91–1.05) | 0.98 (0.91–1.05) | 0.96 (0.90–1.03) | 0.95 (0.89–1.02) |
| Sexual | 0.98 (0.92–1.06) | 1.00 (0.94–1.07) | 0.95 (0.90–1.01) | 0.96 (0.90–1.02) |
| Emotional | 1.08 (0.99–1.19) | 1.03 (0.94–1.12) | 1.02 (0.95–1.10) | 0.98 (0.91–1.05) |
| Married | ||||
| Physical | 1.06 (0.91–1.23) | 1.05 (0.93–1.19) | 0.97 (0.85–1.11) | 0.95 (0.83–1.09) |
| Sexual | 0.93 (0.80–1.08) | 1.01 (0.89–1.14) | 0.92 (0.81–1.05) | 0.97 (0.86–1.10) |
| Emotional | 1.08 (0.88–1.34) | 1.13 (0.95–1.34) | 1.03 (0.87–1.21) | 1.06 (0.92–1.22) |
| Unmarried | ||||
| Physical | 0.96 (0.88–1.04) | 0.96 (0.89–1.04) | 0.96 (0.89–1.04) | 0.95 (0.89–1.03) |
| Sexual | 1.00 (0.92–1.08) | 1.01 (0.94–1.09) | 0.96 (0.89–1.03) | 0.96 (0.90–1.03) |
| Emotional | 1.08 (0.98–1.20) | 0.95 (0.87–1.04) | 1.02 (0.94–1.11) | 0.96 (0.88–1.05) |
| Contraception continuation | ||||
| Physical | 1.00 (0.93–1.07) | 0.99 (0.92–1.06) | 0.97 (0.91–1.03) | 0.96 (0.90–1.02) |
| Sexual | 0.99 (0.92–1.05) | 0.99 (0.93–1.06) | 0.93 (0.87–0.99) | 0.92 (0.86–1.08) |
| Emotional | 1.08 (0.99–1.18) | 1.03 (0.94–1.13) | 1.02 (0.95–1.10) | 0.98 (0.91–1.06) |
| Contraception uptake | ||||
| Physical | 0.93 (0.70–1.23) | 0.91 (0.69–1.21) | 0.96 (0.78–1.19) | 0.92 (0.75–1.13) |
| Sexual | 1.05 (0.80–1.38) | 1.09 (0.84–1.42) | 1.12 (0.92–1.37) | 1.12 (0.92–1.36) |
| Emotional | 1.10 (0.78–1.56) | 1.02 (0.72–1.43) | 1.03 (0.80–1.32) | 0.94 (0.72–1.22) |
| Barrier | ||||
| Physical | 0.93 (0.84–1.04) | 0.97 (0.87–1.07) | 1.00 (0.92–1.09) | 1.01 (0.93–1.10) |
| Sexual | 0.94 (0.85–1.04) | 1.01 (0.92–1.11) | 0.93 (0.86–1.02) | 0.99 (0.91–1.07) |
| Emotional | 1.00 (0.89–1.12) | 0.97 (0.87–1.09) | 0.99 (0.90–1.09) | 0.95 (0.87–1.04) |
| Non-barrier | ||||
| Physical | 1.09 (0.92–1.30) | 0.98 (0.84–1.14) | 1.04 (0.86–1.24) | 0.91 (0.78–1.06) |
| Sexual | 1.02 (0.86–1.21) | 0.98 (0.85–1.13) | 1.04 (0.87–1.24) | 0.95 (0.82–1.10) |
| Emotional | 1.33 (1.05–1.68) | 0.96 (0.80–1.16) | 1.16 (0.93–1.46) | 0.88 (0.74–1.05) |
RRs adjusted for marital status, parity, study arm, contraceptive use at baseline
Fig. 1.
Forest plot prepresenting adjusted risk ratios for the impact of IPV on contraception at 6 and 12 months by type of contraception.
IPV did not affect barrier method use for any violence category at 6 or 12 months in all adjusted and unadjusted models. For physical violence, the aRR was 0.97 (95% CI: 0.87–1.07) at 6 months and 1.01 (95% CI: 0.93–1.10) at 12 months. For sexual violence, the aRR was 1.01 (95% CI: 0.92–1.11) at 6 months and 0.99 (95% CI: 0.91–1.07) at 12 months. For emotional violence, the aRR was 0.97 (95% CI: 0.87–1.09) at 6 months and 0.95 (95% CI: 0.87–1.04) at 12 months.
IPV did not affect non-barrier method use for any violence category at 6 or 12 months in all adjusted or unadjusted models. For physical violence, the aRR was 0.98 (95% CI: 0.84–1.14) at 6 months and 0.91 (95% CI: 0.78–1.06) at 12 months. For sexual violence, the aRR was 0.98 (95% CI: 0.85–1.13) at 6 months and 0.95 (95% CI: 0.82–1.10) at 12 months. For emotional violence, the aRR was 0.96 (95% CI: 0.80–1.16) at 6 months and 0.88 (95% CI: 0.74–1.05) at 12 months.
Secondary analyses
Similar findings were observed in analyses stratified by marital status and baseline contraception use. Married women were more likely to use non-barrier methods compared to unmarried women (n=953, 69.7% vs 17.4%, p<0.001). They were also less likely to use barrier methods (n=953, 26.6% vs 76.9%, p<0.001). However, IPV did not affect either outcome among either sub-population at either time point. Similarly, participants who used contraception at baseline were more likely to use contraception at six and twelve months, but IPV did not affect contraceptive use in either group (Table 2).
Discussion
This analysis investigated the effect of baseline IPV on 6- and 12- month contraceptive use among AGYW in Lilongwe, Malawi. We found at baseline in cross-sectional analyses that participants who experienced sexual violence were more likely to use non-barrier methods and less likely to use barrier methods than those who did not experience sexual violence. Participants who experienced emotional violence were more likely to use non-barrier methods than those who did not experience emotional violence. Yet, physical, sexual, and emotional IPV did not affect the use of any contraception, barrier method contraception, or non-barrier method contraception in longitudinal analyses. These findings were also observed in stratified analyses among married and unmarried AGYW, as well as those who were and were not using contraception at baseline.
Our study is the first to examine the effect of IPV on contraception longitudinally among AGYW in SSA, a large population with very high rates of IPV and low rates of contraceptive use.39 AGYW ages 15–24 years in SSA account for over 100 million people, making them the majority of women of reproductive age and a population at high risk for HIV acquisition and pregnancy.40,41 These populations face high rates of IPV and unmet need for contraception, and understanding the effect of IPV on contraceptive use is critically important.
To our knowledge, there is only one prior prospective cohort study investigating the effect of IPV on contraception in SSA. However, this study, which was conducted in Rakai, Uganda,42 included all women of reproductive age (15–49 years), with only a small fraction of the population being 15–24 years old. This study found that IPV had no effect on non-barrier contraception but did find that condom use at last sex was 20% less in those experiencing IPV.42 Our findings mirror the non-barrier results, but we did not observe decreased condom use among participants reporting IPV in longitudinal analyses. This suggests that this younger population may not experience the same relationship dynamics as older women.
Associations between IPV and contraception have been studied extensively in SSA, but has been almost entirely limited to cross-sectional study designs14–28 which limit ability to infer directionality or causality in the relationship. Findings from these studies have been mixed with many studies finding a higher prevalence of contraceptive use among women in violent relationships.14–18 Our results are not consistent with these paradoxical findings. These cross- sectional studies may, in fact, be capturing the effect of reverse causality or unmeasured confounding in this relationship.18 This possibility is supported by a qualitative study conducted in South Africa that found that women using the vaginal ring reported increased violence after initiating the method.43 We observe similar findings in our cross-sectional analysis at baseline with a significant associations that were not reproduced in our longitudinal findings. By using longitudinal methods, disaggregating both exposure and outcome measures, and using a validated and widely used survey to measure IPV, we addressed the common methodologic limitations of prior work in the region.
Despite the use of a validated exposure survey, questions about experiences of violence were asked referencing any point in time with the current partner. Participants were not asked to clarify when these behaviors had taken place or the length of time they had been with their current partner. Thus, we were not able to examine whether violence was ongoing or had occurred in the past. Additionally, we were also not able to measure IPV as a time varying exposure. Despite this limitation, IPV data is from the most recent partner rather than lifetime prevalence, and most women had a recent partner. Furthermore, previous violence from a partner still may affect contraceptive use or other reproductive choices.
Our methods for measurement of contraceptive use had several strengths and weaknesses. Contraception was measured through self-report at clinic visits and therefore is subject to social desirability bias. However, in a previous analysis of this cohort,35 comparisons between clinical and survey data showed consistency in self reporting on contraception suggesting that survey data was representative of actual use and study arm is unlikely to influence the extent of social desirability bias. Additionally, contraceptive use was measured in the context of an intervention study with the potential for the increased uptake of contraception associated with the youth friendly service model to overshadow the influence of IPV. To account for this, we adjusted for study arm in all analyses.
This study provides new insights into the relationship between IPV and contraceptive use among AGYW in sub-Saharan Africa. In this well-designed longitudinal analysis, IPV was not found to be a key driver of contraception use in AGYW. Interventions are needed to address the persistently high prevalence of IPV in SSA, but interventions targeting IPV alone may not ultimately impact contraception use in this population. Future research involving assessment of IPV interventions, such as portions of the DREAMS program, with contraception as an outcome would provide further evidence of the direct impact of IPV on contraception. Previous analysis of this cohort35 has shown multifaceted youth friendly health services (YFHS) significantly contribute to increased contraception uptake in this age group, another key component of the DREAMS program.24 This analysis suggests that rather than leveraging IPV alone to improve contraceptive use, these types of programs should focus attention on comprehensive YFHS programs that target multiple drivers to make services more accessible in the region.
Implications and Contributions.
Interventions to address the high rates of intimate partner violence (IPV) and unmet need for contraception are urgently needed. Using longitudinal methods, IPV was not a key driver of contraceptive use among adolescent girls and young women in Malawian clinical settings. Therefore, programs addressing IPV alone may not increase contraceptive use.
Acknowledgements:
The study was funded by Evidence for HIV Prevention in Southern Africa (EHPSA), a DFID program managed by Mott MacDonald. NER was funded by the National Institute of Mental Health (R00 MH104154) and the National Institute of Allergy and Infectious Diseases (P30 AI50410).
Abbreviations:
- SSA
sub-Saharan Africa
- AGYW
adolescent girls and young women
- IPV
intimate partner violence
- HIV
human immunodeficiency virus
- IUD
intrauterine device
- OC
oral contraceptive
- DREAMS
Determined, Resilient, Empowered
- AIDS-free
Mentored, and Safe
- PEPFAR
The United States President’s Emergency Plan for AIDS Relief
- YFHS
youth friendly health services
- RR
risk ratio
- aRR
adjusted risk ratio
- CI
confidence interval
Footnotes
Conflicts of Interest: All authors declare no conflicts of interest
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1).Ahinkorah BO, Dickson KS, Seidu AA. Women decision-making capacity and intimate parter violence among women in sub-Saharan Africa. Arch Public Health 2018;76:5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2).Ahinkorah BO. Intimate partner violence against adolescent girls and young women and its association with miscarriages, stillbirths and induced abortions in sub-Saharan Africa: Evidence from demographic and health surveys. SSM Popul Health 2021;13:100730. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3).Kidman R, Kohler HP. Emerging partner violence among young adolescents in a low-income country: Perpetration, victimization and adversity. PLoS One 2020;15(3): e0230085. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4).Global Database on Violence against Women: Malawi. United Nations Women. https://evaw-global-database.unwomen.org/fr/countries/africa/malawi?typeofmeasure=60305524975b40e6b0ede2ae54aa5c0e [Google Scholar]
- 5).World Health Organization. Global and regional estimates of violence against women: prevalence and health effects of intimate partner violence and non-partner sexual violence. 2013. https://apps.who.int/iris/bitstream/handle/10665/85239/9789241564625_eng.pdf;jsessionid=0A62DF20E0D5D252C3FBF55B47FAF247?sequence=1. Accessed July 28, 2021.
- 6).Jewkes RK, Dunkle K, Nduna M, et al. Intimate partner violence, relationship power inequity, and incidence of HIV infection in young women in South Africa: a cohort study. Lancet 2010;376(9734):41–8. [DOI] [PubMed] [Google Scholar]
- 7).Silverman JG, Gupta J, Decker MR, et al. Intimate partner violence and unwanted pregnancy, miscarriage, induced abortion, and stillbirth among a national sample of Bangladeshi women. BJOG 2007;114(10):1246–52. [DOI] [PubMed] [Google Scholar]
- 8).Chaves K, Eastwood J, Ogbo FA, et al. Intimate partner violence identified through routine antenatal screening and maternal and perinatal health outcomes. BMC Pregnancy Childbirth 2019;19(1):357. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9).Yaya S, Gunawardena N, Bishwajit G. Association between intimate partner violence and utilization of facility delivery services in Nigeria: a propensity score matching analysis. BMC Public Health 2019;19(1):1131. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10).Sabri B, Huerta J, Alexander KA, et al. Multiple Intimate Partner Violence Experiences: Knowledge, Access, Utilization and Barriers to Utilization of Resources by Women of the African Diaspora. J Health Care Poor Underserved 2015;26(4):1286–303. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11).Maxwell l, Devries K, Zionts D, et al. Estimating the Effect of Intimate Partner Violence on Women’s Use of Contraception: A Systematic Review and Meta-Analysis. PLoS One 2015;10(2):e0118234. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12).McCauley HL, Silverman J, Decker MR. Contraceptive Use, Pregnancy Testing and Intimate Partner Violence Among Adolescent and Young Adult Women Ages 16–24. Journal of Pediatric and Adolescent Gynecology 2013; 26(2):e57. [Google Scholar]
- 13).Northridge JL, Silver Ej, Talib HJ, et al. Reproductive Coercion in High School-Aged Girls: Associations with Reproductive Health Risk and Intimate Partner Violence. J Pediatr Adolesc Gynecol 2017; 30(6): 603–608. [DOI] [PubMed] [Google Scholar]
- 14).Alio A, Daley E, Nana P, et al. Intimate partner violence and contraception use among women in Sub-Saharan Africa. International Journal of Gynecology and Obstetrics 2009; 107(1):35–38. [DOI] [PubMed] [Google Scholar]
- 15).Okenwa L, Lawoko S, Jansson B. Contraception, reproductive health and pregnancy outcomes among women exposed to intimate partner violence in Nigeria. Eur J Contracept Reprod Health Care 2011;16(1):18–25. [DOI] [PubMed] [Google Scholar]
- 16).Emenike E, Lawoko S, Dalal K. Intimate partner violence and reproductive health of women in Kenya. Int Nurs Rev 2008;55(1):97–102. [DOI] [PubMed] [Google Scholar]
- 17).Wusu O. Predictors and Implications of Intimate Partner Violence Against Married Female Youths in Nigeria. J Fam Viol 2015;30:63–74. [Google Scholar]
- 18).Adjiwanou V, N’Bouke A. Exploring the Paradox of Intimate Partner Violence and Increased Contraceptive Use in sub-Saharan Africa. Stud Fam Plann 2015;46(2):127–42 [DOI] [PubMed] [Google Scholar]
- 19).Kabir R, Kordowicz M. Exploring the relationship between intimate partner violence and contraception use by Tanzanian women. Health Care Women Int 2020;1–14. [DOI] [PubMed] [Google Scholar]
- 20).Wandera SO, Kwagala B, Odimegwu C. Intimate partner violence and current modern contraceptive use among married women in Uganda: a cross-sectional study. Pan Afr Med J 2018;30:85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21).Yusuf RA, Dongarwar D, Yusuf ZI, et al. Association between Intimate Partner Violence, Knowledge and Use of Contraception in Africa: Comparative Analysis across Five African Regions. International Journal of Maternal and Child Health and AIDS 2020;9(1):42–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22).Antai D, Adaji S. Community-level influences on women’s experience of intimate partner violence and terminated pregnancy in Nigeria: a multilevel analysis. BMC Pregnancy Childbirth 2012;12:128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23).Chalachala J. Association of Intimate Partner Violence (IPV) and Current Contraceptive Use in the Democratic Republic of Congo (DRC): a secondary data analysis of the Demographic Health Survey (DHS) II 2013–2014. University of North Carolina at Chapel Hill Department of Maternal and Child Health; 2016. https://core.ac.uk/download/pdf/210606676.pdf Accessed July 28, 2021. [Google Scholar]
- 24).Diop-Sidibé N, Campbell JC, Becker S. Domestic violence against women in Egypt—wife beating and health outcomes. Soc Sci Med 2006;62(5)1260–77. [DOI] [PubMed] [Google Scholar]
- 25).Kidayi PL, Msuya S, Todd J, et al. Determinants of Modern Contraceptive Use among Women of Reproductive Age in Tanzania: Evidence from Tanzania Demographic and Health Survey Data. Advances in Sexual Medicine 2015;5:43–52. [Google Scholar]
- 26).Kidman R, Palermo T, Bertrand J. Intimate partner violence, modern contraceptive use and conflict in the Democratic Republic of the Congo. Soc Sci Med 2015;133:2–10. [DOI] [PubMed] [Google Scholar]
- 27).Oluwaseyi SD, Latifat I. Intimate Partner Violence and Contraceptive Behavior: Evidence from Malawi and Zambia. Southern African Journal of Demography 2015;16(1):123–50. [Google Scholar]
- 28).Oyedokun AO. Domestic violence and constrained contraceptive choices in selected areas of Osun State, Nigeria. Sexual and Relationship Therapy 2008;23(4):305–23. [Google Scholar]
- 29).Kacanek D, Bostrom A, Montgomery ET, et al. Intimate partner violence and condom and diaphragm nonadherence among women in an HIV prevention trial in southern Africa. J Acquir Immune Defic Syndr 2013;64(4):400–8. [DOI] [PubMed] [Google Scholar]
- 30).Kupoluyi JA. Intimate partner violence as a factor in contraceptive discontinuation among sexually active married women in Nigeria. BMC Women’s Health 2020;20(1):128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31).Kriel Y, Milford C, Cordero J, et al. Male partner influence on family planning and contraceptive use: perspectives from community members and healthcare providers in KwaZulu-Natal, South Africa. Reprod Health 2019;16(1):89. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32).OlaOlorun F, Seme A, Otupiri E, Ogunjuyigbe P, Tsui A. Women’s fertility desires and contraceptive behavior in three peri-urban communities in sub-Saharan Africa. Reprod Health 2016;13:12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33).Saul J, Bachman G, Allen S, Toiv NF, Cooney C, Beamon T. The DREAMS core package of interventions: A comprehensive approach to preventing HIV among adolescent girls and young women. PLoS One 2018;13(12):e0208167. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34).Rosenberg NE, Pettifor AE, Myers L, et al. Comparing four service delivery models for adolescent girls and young women through the ‘Girl Power’ study: protocol for a multisite quasi-experimental cohort study. BMJ Open 2017;7(12):e018480. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35).Rosenberg NE, Bhushan NL, Vansia D, et al. Comparing Youth-Friendly Health Services to the Standard of Care Through “Girl Power-Malawi”: A Quasi-Experimental Cohort Study. J Acquir Immune Defic Syndr 2018;79(4):458–466. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36).Kishor S. Domestic violence measurements in the demographic and health surveys: the history and the challenges. Retrieved from Expert paper for UN Division for the Advancement of Women, Geneva, 2005: http://www.un.org/womenwatch/daw/egm/vaw-stat-2005/docs/expert-papers/Kishor.pdf. Accessed July 28, 2021. [Google Scholar]
- 37).Bazargan-Hejazi S, Medeiros S, Mohammadi R, et al. Patterns of Intimate Partner Violence: a study of female victims in Malawi. Journal of Injury and Violence Research 2013; 5(1), 38–50. doi: 10.5249/jivr.v5i1.139 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38).Zou G. A modified poisson regression approach to prospective studies with binary data. Am J Epidemiol 2004;159(7):702–6. [DOI] [PubMed] [Google Scholar]
- 39).World Health Organization. Violence Against Women Fact Sheet. 2017. https://www.who.int/news-room/fact-sheets/detail/violence-against-women. Accessed July 28, 2021.
- 40).Population Pyramid. 2021. https://www.populationpyramid.net/sub-saharan-africa/2021/. Accessed July 28, 2021.
- 41).ICAP. Malawi Population-Based HIV Impact Assessment MPHIA 2015–2016. Columbia University Health MSoP; 2016. [Google Scholar]
- 42).Maxwell L, Brahmbhatt H, Ndyanabo A, et al. The impact of intimate partner violence on women’s contraceptive use: Evidence from the Rakai Community Cohort Study in Rakai, Uganda. Soc Sci Med 2018;209:25–32. [DOI] [PubMed] [Google Scholar]
- 43).Hartmann M, Palanee-Phillips T, O’Rourke S, et al. The relationship between vaginal ring use and intimate partner violence and social harms: formative research outcomes from CHARISMA study in Johannesburg, South Africa. AIDS Care 2019;31(6):660–666. [DOI] [PubMed] [Google Scholar]