Women’s primary contraceptive method type is impacted by the gender inequities of girl child marriage and intimate partner violence (IPV). Outside of South Asia, proximity to conflict zones has been found to impact contraceptive use, girl child marriage, and IPV, possibly moderating associations between these variables. We created multinomial regression models using the 2016 Demographic and Health Survey data from postconflict Sri Lanka to study associations between primary contraceptive method type (modern spacing methods, sterilization, and traditional methods compared to no method) and the gender inequities of girl child marriage and past year sexual, physical, and emotional IPV and to assess whether and how these associations were moderated by proximity to conflict. We found that proximity to conflict moderated the relationships between girl child marriage, past year physical and emotional IPV, and primary contraceptive method type. Girl child marriage was associated with increased relative risk (RR) of modern spacing methods (adjusted RR ratio/aRRR: 1.81–2.21) across all levels of proximity to conflict. In districts distal to conflict, past year physical IPV was associated with decreased RR of sterilization (aRRR: 0.67) and traditional methods (aRRR: 0.63), and past year emotional IPV was associated with decreased RR of traditional methods (aRRR: 0.71). In districts central to conflict, past year emotional IPV was associated with increased RR of modern spacing methods (aRRR: 1.50). Our findings suggest that policymakers and providers who seek to improve reproductive health in Sri Lanka must consider the moderating impact of proximity to conflict on the relationship between contraceptive use and the gender inequities of girl child marriage and IPV.

Background

The primary contraceptive method type used by an individual to prevent or delay pregnancy can have a major impact on their health and well-being. In Sri Lanka, women have free access to contraception [1], and sterilization is the most widely used contraceptive method, used by 14% of women aged 15–49 [2]. Couples who rely on sterilization as their only contraceptive option cannot space or delay births, so widespread use of sterilization over other options can drive early childbearing with lack of healthy birth spacing [3]. Contraceptive methods vary in accessibility, detectability by a partner, whether they can be reversed (allowing for birth spacing), effectiveness (with modern methods being more effective than traditional), and acceptability to women and their partners. Sri Lankan women are more likely to use less-effective traditional methods than modern methods of contraception if they are Muslim, older than 35 years old, have a youngest child older than 6 years, and live with extended family members [4]. Sri Lankan women are also more likely to use traditional methods than modern methods if they live farther from a health clinic, have not had a public health midwife visit them, and have poor knowledge of modern contraceptives or a negative opinion of services at the nearest clinic [4]. In addition to demographic factors and issues of contraceptive access or knowledge, gender inequities such as girl child marriage (marrying or cohabiting before age 18 [5]) and intimate partner violence (IPV) can also impact women and girls’ use of contraception.

The gender inequities of girl child marriage and IPV have been found in numerous settings to influence an individual’s primary method of contraception, possibly through the mechanism of reproductive coercion [6]. Girl child marriage has been found in multiple South Asian contexts to be associated with increased use of modern contraceptive methods including female sterilization. In a study of India, Bangladesh, Nepal, and Pakistan, girl child marriage was significantly associated with current modern contraceptive use and female sterilization (vs. not being sterilized) [7]. In India, girl child marriage has been found to be associated with increased likelihood of female sterilization, even when controlling for duration of marriage [8]. There is no scientific consensus on the impact of IPV on women’s primary contraceptive method type. A 2008 study found that in Bangladesh, women who experienced physical or sexual IPV were more likely to use current modern contraceptive methods than women who did not experience IPV [9]. In contrast, a 2015 multicountry study of Nepal, Bangladesh, and India found no associations between physical IPV and contraceptive methods as well as a positive association between sexual IPV and modern spacing methods and a negative association between sexual IPV and female sterilization [10]. Studies in the United States have found associations between physical and sexual IPV and increased likelihood of female sterilization [11, 12]. A 2015 systematic review and meta-analysis of longitudinal studies of IPV and contraception found that women’s experience of IPV was associated with a significant reduction in their odds of using contraception [13]. The authors of the review suggested that researchers of IPV and contraceptive methods consider separately methods that are modern versus traditional and methods that do or do not require ongoing negotiations between a woman and her male partner [13]. Researchers have identified local lifetime IPV prevalence rates in Sri Lanka of around 30–40% in various single-setting studies across the country [14, 15, 16, 17] and have found IPV to be more likely in Sri Lanka in the context of socioeconomic disadvantage [18]. While women’s primary contraceptive method type may be influenced by inequitable and/or violent experiences in their relationships, community-level violence can also have an impact on contraception.

One factor that may moderate associations between primary contraceptive method type and the gender inequities of girl child marriage and IPV is proximity to conflict zones, both during and postconflict. Conflict and postconflict settings often suffer from underdevelopment of health systems, which can lead to lowered health access and negative outcomes for inhabitants [19]. Although demand for contraception is not lower in postconflict compared to non-conflict-affected settings [20, 21], underdeveloped health facilities can decrease women’s access to and use of modern contraceptive methods [21]. Exposure to conflict has been linked to negative outcomes for adolescent girls including increased girl child marriage and underage pregnancies [22, 23, 24, 25]. Conflict has also been shown in many settings to result in increased IPV [26, 27, 28, 29], possibly through the everyday normalization of military violence leading to acceptance of violence as a problem-solving measure within the home [30, 31]. Because of conflict’s documented impact on girl child marriage, IPV, and contraception, it is possible that proximity to conflict could moderate any associations between these variables. Studies on the impact of conflict on women’s sexual and reproductive health have largely been conducted in Sub-Saharan Africa and have not explored how conflict might affect contraception in South Asia.

Within South Asia, Sri Lanka presents a unique context in which to study whether conflict moderates associations between contraceptive use, girl child marriage, and IPV. During the 26-year civil war, which ended in 2009, residents of the Northern and Eastern provinces where the majority of armed conflict occurred were exposed to tremendous amounts of military violence by both separatist and government forces [32, 33]. Research on the conflict’s impact on health outcomes has uncovered marked differences in health care utilization and outcomes between areas that were central to (within), proximal to (bordering), and distal to (furthest from) conflict zones, as well as an overall decrease in health systems development during the conflict [34], with rural and conflict-affected areas having far fewer doctors per inhabitants than urban areas far from conflict zones [35]. A 2016 Lancet report described how, 7 years after the end of the war, living standards in conflict zones were still poor, with limited access to toilets, water, and electricity [36]. During the war, families in the conflict zones practiced child marriage in order to remove their children from eligibility to be recruited by militant factions which were compelling youth to fight [37], and this practice may have persisted in the postconflict era [38]. IPV has also been found to be higher in the conflict zones many years after the end of the war [18, 39]. The enduring impact of the civil war on health systems, child marriage, and IPV in Sri Lanka suggests that the relationships between primary contraceptive method type, girl child marriage, and IPV may be moderated by individuals’ proximity to conflict.

About This Study

In postconflict Sri Lanka, the lasting impact of the conflict on IPV, girl child marriage, and contraception remains poorly understood. Using disaggregated data from the 2016 Sri Lankan Demographic and Health Survey, this study aims to understand how associations between girl child marriage, past year IPV, and primary contraceptive method type are moderated by proximity to conflict, a question that has not previously been considered in South Asia. Findings from this study will provide policymakers and health practitioners in Sri Lanka with much-needed information on how primary contraceptive method type is impacted by the gender inequities of girl child marriage and past year IPV and the ways in which these relationships are impacted by postconflict proximity to the conflict zone.

Data Source

This study used data from the 2016 Sri Lankan Demographic and Survey (DHS), which collected individual-level data on child and maternal health outcomes, domestic violence experience, reproductive health, and information on economic engagement and agency of women in Sri Lanka (N = 27,210 households, 18,510 women aged 15–49) [2]. The 2016 survey was the first DHS to collect data from a nationally representative sample of households in Sri Lanka—all previous DHS data collection occurred during the 30-year civil war and excluded portions of the Northern and Eastern provinces, which had been claimed as a separate state by the Liberation Tigers of Tamil Eelam organization [40]. Similar to other DHS data sets, the 2016 Sri Lankan DHS followed a multistage, clustered sample design that allowed for population-level estimates at both the national and district level [2, 41].

In addition to a general health survey administered to every eligible woman in each household, a domestic violence module was administered to one woman per household (n = 16,629). The 2016 DHS was the first to ask questions on experience of past year IPV in Sri Lanka and, to date, no peer-reviewed studies have been published using these IPV data. Following the World Health Organization’s (WHO) guidelines for the ethical collection of information on domestic violence, one eligible woman per household was randomly selected for this module, which was not implemented if privacy could not be obtained [42]. Each domestic violence module respondent was read an additional consent statement at the start of the module, informing her that the questions could be personal and reassuring her of the confidentiality of her responses. In the 2016 DHS, women taking part in the domestic violence module were asked nine questions about their experiences of IPV in the previous 12 months [2]. This study was restricted to women who were currently living with an intimate partner, answered the domestic violence module of the DHS, and were not pregnant nor trying to become pregnant, following a convention of counting women who wish to get pregnant in the 2 years post DHS as having no need for contraception [43]. This study was also restricted to women aged 18 and above, following a convention in girl child marriage research to censor participants under age 18 who might marry after the survey’s completion [5]. The complete sample of women included in this study comprised of 11,426 participants. Ethical approval for this research was obtained from the University of California, San Diego Institutional Review Board (Project number #191418XX).

Variables of Interest

Dependent Variable: Primary Contraceptive Method Type

As the dependent variable, we examined women’s primary contraceptive method type and divided their responses into four categories based on whether a method type was modern or traditional and among modern methods whether it could be reversed and used for birth spacing (separating out sterilization). The four categories are listed below:

  1. No method (this was used as the reference category in multinomial analyses).

  2. Modern spacing methods: pill, intrauterine device, implant, condom, injection, female condom, or emergency contraception.

  3. Sterilization: male or female sterilization. Over 99% of this category comprised of female sterilization, but the respondents who listed male sterilization were included in order to retain a population-representative sample.

  4. Traditional methods: lactational amenorrhea, rhythm method, or withdrawal (the same grouping used by the Sri Lankan Department of Census and Statistics in their descriptive report of DHS data) [2].

Independent Variables: Girl Child Marriage and IPV

We examined four variables as independent variables of interest to account for the separate and overlapping effects of different experiences. The four binary independent variables we considered are listed below:

  1. Girl child marriage: having married or cohabited with a male partner before age 18, yes or no (conventional definition used by girl child marriage researchers [5]).

  2. Past year sexual IPV: having been forced to have sex by a partner in the last 12 months, yes or no.

  3. Past year physical IPV: having experienced at least one of six types of physical IPV: (1) slapping or beating with a hand, (2) pushing or shoving, (3) strangulation, (4) dragging or pulling, (5) beating with an object, or (6) burned, in the last 12 months, yes, or no.

  4. Past year emotional IPV: having experienced at least one of two incidents, either being belittled/seriously offended or prevented from leaving home by a partner, in the last 12 months, yes or no.

We considered past year physical, sexual, and emotional IPV as separate independent variables to allow estimation of the independent effects of each form of IPV. We calculated Spearman’s ρ correlation estimates across all three IPV variables. All IPV correlation estimates were below 0.5, justifying retaining them as independent variables.

Moderating Variable: Proximity to Conflict

We considered proximity to conflict as a potential moderator of the associations between girl child marriage, past year IPV, and primary contraceptive method type. This variable was defined as having three ordinal levels: central, proximal, and distal. Participants assigned to the “central” category reported residing in either the Northern or Eastern provinces of Sri Lanka, where the majority of the armed conflict occurred during the civil war. Central districts were Ampara, Batticaloa, Jaffna, Kilinochchi, Mannar, Mullaitivu, Trincomalee, and Vavuniya. Participants assigned to the “proximal” category reported residing in one of the seven districts of Sri Lanka outside of the Northern and Eastern regions that shared a border with one or both of these regions. Proximal districts were Anuradhapura, Badulla, Hambantota, Matale, Monaragala, Polonnaruwa, and Puttalam. Finally, participants assigned to the “distal” category resided in districts that were neither in nor bordering the Northern and Eastern provinces of the country. Distal districts were Colombo, Gampaha, Galle, Kalutara, Kandy, Kegalle, Kurunegala, Matara, Nuwara Eliya, and Ratnapura.

Covariates

We included as covariates variables known or hypothesized to be associated with girl child marriage, IPV, and contraception. These variables included each respondent’s age, education, household wealth, parity, the age difference between her and her partner, her participation in decisions about her health care, years married, household size, religion, and ethnicity. Beyond individual- and family-level characteristics, we also included as covariates whether the respondent lived in an urban setting or not and which district in Sri Lanka she lived in to control for unmeasured variation at the community level.

Statistical Analysis

All statistical analyses were conducted using R software version 3.6.3 [44], and estimates were adjusted for complex survey design and participant-level weights using the “survey” package [45] in order to calculate population-representative measures. We first assessed the distributions of all considered covariates, girl child marriage, all three forms of past year IPV, proximity to conflict, and primary contraceptive method type in the sample. We used χ2 tests for all comparisons across primary contraceptive method types after reducing all variables to categories (ordinal or nominal). In preparation for our multivariable model, we assessed all variables for multicollinearity by calculating their variance inflation factor (VIF). One of each pair of similarly distributed variables was excluded from the following multivariable models, resulting in a set of covariates in which all VIF values were less than 5. The following variables were excluded from multivariable models due to multicollinearity: age and ethnicity.

Next, we examined the role of proximity to conflict as a potential moderator of the relationships between child marriage, past year IPV, and primary contraceptive method type by first creating four separate interaction terms that combined proximity to conflict with child marriage and with each form of past year IPV [46]. We assessed the significance of these interaction terms in a multinomial logistic regression which estimated the adjusted change in relative risk of using modern spacing contraceptive methods, sterilization, or traditional methods (each compared to no method). The multinomial model was necessary in order to compare the multiple types of contraceptive methods within one parsimonious model. We included in this multinomial model girl child marriage, all three past year IPV variables, proximity to conflict, and all noncollinear covariates that were significantly associated (p < .05) with primary contraceptive method type. If an interaction term between proximity to conflict and an independent variable was statistically significant (p < .05), it indicated that proximity to conflict moderated that variable’s impact on primary contraceptive method type (it partitioned that independent variable into domains of varying association with the dependent variable). To understand how relationships between the independent variables and primary contraceptive method type were impacted by differing levels of proximity to conflict, we examined how relationships between all independent variables and primary contraceptive method type in the multinomial models changed when stratified by the different levels of proximity to conflict. Finally, a sensitivity analysis was performed on the data set to assess whether removing respondents using male sterilization resulted in any changes in the direction or significance of any of the observed associations.

Descriptive characteristics of the sample and their distributions across primary contraceptive method types are summarized with unweighted frequencies and weighted percentage values in table 1. Almost half (49%) of women reported using modern spacing methods of contraception. Nearly one in six (15%) of women had married before the age of 18. Women’s experiences of IPV in the past year varied, with 2% reporting having experienced past year sexual IPV and 14% having experienced past year emotional IPV. The majority of women (64%) lived in districts that were distal to conflict. Most women were older than 29, had attended secondary education, and had given birth two or more times. Over one third of women (34%) had partners who were 6 or more years older than them. Eighty-six percent of women reported that they made decisions about their health care individually or as an equal with their partner. Most women had been married 10 or more years and lived in households with four or more people. Finally, most women lived outside of urban settings, with Buddhism as the most practiced religion and Sinhala as the largest ethnic group represented.

Table 1.

Demographic Details of Currently Partnered Women Aged 18–49 Who Are Not Pregnant or Trying to Become Pregnant and Participated in the 2016 Sri Lanka DHS Domestic Violence Module, by Type of Contraceptive Method. DOI: https://doi.org/10.1525/agh.2022.1539582.t1

Primary Contraceptive Method Type
TotalNoneModern SpacingSterilizationTraditionalχ2
Characteristicn%n%*n%*n%*n%*p Value
Total 11,426 100 2,551 19 5,383 49 2,026 17 1,466 14 — 
Girl child marriage 
 No 9,622 85 2,199 20 4,499 49 1,577 16 1,347 15 <.01 
 Yes 1,804 15 352 17 884 50 449 25 119  
Past year sexual IPV 
 No 11,104 98 2,455 19 5,254 49 1,956 17 1,439 14 <.01 
 Yes 322 96 26 129 41 70 23 27 10  
Past year physical IPV 
 No 10,288 91 2,250 19 4,870 49 1,780 17 1,388 15 <.01 
 Yes 1,138 301 24 513 47 246 21 78  
Past year emotional IPV 
 No 9,735 87 2,098 19 4,633 50 1,682 17 1,322 15 <.01 
 Yes 1,691 13 453 24 750 46 344 20 144 10  
Proximity to conflict 
 Distal 6,373 64 1,078 17 3,120 50 1,100 17 1,075 17 <.01 
 Proximal 2,654 23 408 15 1,402 54 557 21 287 11  
 Central 2,399 13 1,065 42 861 38 369 15 104  
Age** 
 18–29 2,111 18 374 16 1,500 72 32 205 11 <.01 
 30–39 4,834 42 858 15 2,648 57 745 15 583 13  
 40–49 4,481 39 1,319 26 1,235 30 1,249 28 678 17  
Education 
 Primary (01–05) or less 1,111 334 26 318 30 378 35 81 <.01 
 Secondary (6–12) 7,682 67 1,651 18 3,799 51 1,366 18 866 13  
 Higher than Secondary 2,633 24 566 19 1,266 49 282 11 519 21  
Household wealth quintile 
 Lowest 2,642 18 746 22 1,156 48 593 22 147 <.01 
 Second 2,349 20 514 19 1,130 50 423 18 282 13  
 Middle 2,217 21 431 18 1,154 53 346 15 286 14  
 Fourth 2,180 21 438 18 1,054 50 344 16 344 16  
 Highest 2,038 20 422 20 889 44 320 15 407 21  
Parity 
 0–1 2,390 22 590 22 1,316 56 29 455 21 <.01 
 2 4,832 44 1,036 19 2,764 58 321 711 16  
 3 or more 4,204 34 925 18 1,303 33 1,676 41 300  
Age difference between woman and partner** 
 Same or woman is older 1,996 17 481 21 941 50 317 15 257 14 <.01 
 1–5 years 5,550 49 1,218 19 2,657 50 940 16 735 15  
 6–10 years 3,075 27 643 18 1,449 49 602 20 381 14  
 Over 10 years 805 209 23 336 43 167 20 93 13  
Person who makes decisions about health care 
 Woman or woman and partner 9,725 86 2,075 18 4,643 50 1,716 17 1,291 15 <.01 
 Partner or someone else 1,701 14 476 26 740 45 310 17 175 12  
Years married** 
 0–9 3,197 28 618 17 2,086 67 76 417 14 <.01 
 10–19 5,033 44 1,010 17 2,459 51 902 17 662 15  
 20 or more 3,196 28 923 26 838 28 1,048 33 387 14  
Household size 
 1–3 1,713 15 434 23 778 47 226 13 275 18 <.01 
 4 3,748 33 799 19 1,977 54 384 10 588 17  
 5 3,129 27 643 17 1,404 47 738 24 344 12  
 6 or more 2,836 24 675 20 1,224 46 678 23 259 10  
Religion 
 Buddhism 7,480 73 1,113 15 3,925 52 1,281 17 1,161 16 <.01 
 Hinduism 1,917 11 735 38 664 34 432 23 86  
 Islam 992 379 34 406 44 119 13 88  
 All other religions 1,037 324 22 388 43 194 19 131 16  
Ethnicity 
 Sinhala 7,996 79 1,205 15 4,158 52 1,371 17 1,262 16 <.01 
 Sri Lankan Tamil 2,210 12 927 39 744 35 443 21 96  
 Other ethnicities*** 1,220 10 419 31 481 42 212 17 108  
Urban setting 
 No 9,634 85 2,035 18 4,626 50 1,765 18 1,208 14 <.01 
 Yes 1,792 15 516 26 757 44 261 14 258 15  
District 
 Distal to conflict zone 
  Colombo 881 10 193 22 426 48 116 13 146 17 <.01 
  Galle 541 78 15 254 47 96 18 113 20  
  Gampaha 916 10 146 15 442 49 145 16 183 20  
  Kalutara 511 60 12 251 49 81 16 119 23  
  Kandy 645 130 20 316 50 128 18 71 12  
  Kegalle 412 77 19 227 55 66 16 42 10  
  Kurunegala 881 124 14 478 54 127 15 152 17  
  Matara 465 88 19 242 53 61 12 74 16  
  Nuwara Eliya 422 103 22 144 37 156 37 19  
  Ratnapura 699 79 11 340 49 124 17 156 23  
 Proximal to conflict zone 
  Anuradhapura 494 68 13 303 63 92 18 31  
  Badulla 431 73 15 199 48 124 29 35  
  Hambantota 375 79 22 176 47 67 18 53 14  
  Matale 257 28 10 144 56 55 21 30 13  
  Monaragala 350 45 12 186 54 79 22 40 12  
  Polonnaruwa 294 40 14 178 61 57 19 19  
  Puttalam 453 75 16 216 49 83 17 79 17  
Central to conflict zone 
  Ampara 450 160 34 197 45 58 13 35  
  Batticaloa 337 181 55 104 30 34 10 18  
  Jaffna 297 120 39 88 30 73 25 16  
  Kilinochchi 236 64 27 105 45 61 25  
  Mannar 285 219 76 34 13 32 11  
  Mullaitivu 232 49 20 129 58 42 18 12  
  Trincomalee 274 102 36 132 49 29 11 11  
  Vavuniya 288 170 58 72 25 40 14  
Primary Contraceptive Method Type
TotalNoneModern SpacingSterilizationTraditionalχ2
Characteristicn%n%*n%*n%*n%*p Value
Total 11,426 100 2,551 19 5,383 49 2,026 17 1,466 14 — 
Girl child marriage 
 No 9,622 85 2,199 20 4,499 49 1,577 16 1,347 15 <.01 
 Yes 1,804 15 352 17 884 50 449 25 119  
Past year sexual IPV 
 No 11,104 98 2,455 19 5,254 49 1,956 17 1,439 14 <.01 
 Yes 322 96 26 129 41 70 23 27 10  
Past year physical IPV 
 No 10,288 91 2,250 19 4,870 49 1,780 17 1,388 15 <.01 
 Yes 1,138 301 24 513 47 246 21 78  
Past year emotional IPV 
 No 9,735 87 2,098 19 4,633 50 1,682 17 1,322 15 <.01 
 Yes 1,691 13 453 24 750 46 344 20 144 10  
Proximity to conflict 
 Distal 6,373 64 1,078 17 3,120 50 1,100 17 1,075 17 <.01 
 Proximal 2,654 23 408 15 1,402 54 557 21 287 11  
 Central 2,399 13 1,065 42 861 38 369 15 104  
Age** 
 18–29 2,111 18 374 16 1,500 72 32 205 11 <.01 
 30–39 4,834 42 858 15 2,648 57 745 15 583 13  
 40–49 4,481 39 1,319 26 1,235 30 1,249 28 678 17  
Education 
 Primary (01–05) or less 1,111 334 26 318 30 378 35 81 <.01 
 Secondary (6–12) 7,682 67 1,651 18 3,799 51 1,366 18 866 13  
 Higher than Secondary 2,633 24 566 19 1,266 49 282 11 519 21  
Household wealth quintile 
 Lowest 2,642 18 746 22 1,156 48 593 22 147 <.01 
 Second 2,349 20 514 19 1,130 50 423 18 282 13  
 Middle 2,217 21 431 18 1,154 53 346 15 286 14  
 Fourth 2,180 21 438 18 1,054 50 344 16 344 16  
 Highest 2,038 20 422 20 889 44 320 15 407 21  
Parity 
 0–1 2,390 22 590 22 1,316 56 29 455 21 <.01 
 2 4,832 44 1,036 19 2,764 58 321 711 16  
 3 or more 4,204 34 925 18 1,303 33 1,676 41 300  
Age difference between woman and partner** 
 Same or woman is older 1,996 17 481 21 941 50 317 15 257 14 <.01 
 1–5 years 5,550 49 1,218 19 2,657 50 940 16 735 15  
 6–10 years 3,075 27 643 18 1,449 49 602 20 381 14  
 Over 10 years 805 209 23 336 43 167 20 93 13  
Person who makes decisions about health care 
 Woman or woman and partner 9,725 86 2,075 18 4,643 50 1,716 17 1,291 15 <.01 
 Partner or someone else 1,701 14 476 26 740 45 310 17 175 12  
Years married** 
 0–9 3,197 28 618 17 2,086 67 76 417 14 <.01 
 10–19 5,033 44 1,010 17 2,459 51 902 17 662 15  
 20 or more 3,196 28 923 26 838 28 1,048 33 387 14  
Household size 
 1–3 1,713 15 434 23 778 47 226 13 275 18 <.01 
 4 3,748 33 799 19 1,977 54 384 10 588 17  
 5 3,129 27 643 17 1,404 47 738 24 344 12  
 6 or more 2,836 24 675 20 1,224 46 678 23 259 10  
Religion 
 Buddhism 7,480 73 1,113 15 3,925 52 1,281 17 1,161 16 <.01 
 Hinduism 1,917 11 735 38 664 34 432 23 86  
 Islam 992 379 34 406 44 119 13 88  
 All other religions 1,037 324 22 388 43 194 19 131 16  
Ethnicity 
 Sinhala 7,996 79 1,205 15 4,158 52 1,371 17 1,262 16 <.01 
 Sri Lankan Tamil 2,210 12 927 39 744 35 443 21 96  
 Other ethnicities*** 1,220 10 419 31 481 42 212 17 108  
Urban setting 
 No 9,634 85 2,035 18 4,626 50 1,765 18 1,208 14 <.01 
 Yes 1,792 15 516 26 757 44 261 14 258 15  
District 
 Distal to conflict zone 
  Colombo 881 10 193 22 426 48 116 13 146 17 <.01 
  Galle 541 78 15 254 47 96 18 113 20  
  Gampaha 916 10 146 15 442 49 145 16 183 20  
  Kalutara 511 60 12 251 49 81 16 119 23  
  Kandy 645 130 20 316 50 128 18 71 12  
  Kegalle 412 77 19 227 55 66 16 42 10  
  Kurunegala 881 124 14 478 54 127 15 152 17  
  Matara 465 88 19 242 53 61 12 74 16  
  Nuwara Eliya 422 103 22 144 37 156 37 19  
  Ratnapura 699 79 11 340 49 124 17 156 23  
 Proximal to conflict zone 
  Anuradhapura 494 68 13 303 63 92 18 31  
  Badulla 431 73 15 199 48 124 29 35  
  Hambantota 375 79 22 176 47 67 18 53 14  
  Matale 257 28 10 144 56 55 21 30 13  
  Monaragala 350 45 12 186 54 79 22 40 12  
  Polonnaruwa 294 40 14 178 61 57 19 19  
  Puttalam 453 75 16 216 49 83 17 79 17  
Central to conflict zone 
  Ampara 450 160 34 197 45 58 13 35  
  Batticaloa 337 181 55 104 30 34 10 18  
  Jaffna 297 120 39 88 30 73 25 16  
  Kilinochchi 236 64 27 105 45 61 25  
  Mannar 285 219 76 34 13 32 11  
  Mullaitivu 232 49 20 129 58 42 18 12  
  Trincomalee 274 102 36 132 49 29 11 11  
  Vavuniya 288 170 58 72 25 40 14  

N = 11,426. n values are unweighted, while percent and p values are weighted according to the survey’s complex sampling design.

*Cross-tabulated weighted percent values are calculated within rows. **Age, age difference between woman and partner, and years married are presented in tables 1 and 2 as categorical but were included in regression models as continuous variables. ***Other ethnicities include Muslim, Malay, Indian Tamil, Burgher, and other.

Women’s primary contraceptive method type varied significantly by their demographic characteristics (χ2 test of distribution, p < .05). These distributions are shown in the nine rightmost columns in table 1. Women who were married as children were more likely to use sterilization (25%) and less likely to use traditional methods (8%) than women married as adults. Women who experienced each form of IPV were similarly more likely to use sterilization (20–23%) and less likely to use traditional methods (8–10%) than women who did not experience each IPV form, and they were also more likely to not use any method (24–26%). Over two in five (42%) women living in districts central to conflict were not using any method of contraception, which was much higher than the proportion of women not using any method in the distal (17%) and proximal (15%) districts. A majority of women under 40 were using modern spacing methods, while women aged 40–49 were most likely to be using no method of contraception (26%) or sterilization (28%). One in three women who had not attended secondary school reported using sterilization, while about half of women who had attended secondary school or higher were using modern spacing methods. Over one in five women (22%) from the lowest wealth quintile were using sterilization, which was a larger proportion than in wealthier groups. Women who had given birth three or more times were much more likely to use sterilization (41%) than women who had given birth two or fewer times (less than 10%). Over one in four women (26%) who did not participate in their own health care decisions were not using any form of contraception, compared to fewer than one in five women (18%) who participated in decisions about their health care. Women who had been married for 20 years or longer were most likely to use no method (26%) or sterilization (33%) than women married for shorter amounts of time, while a majority of women married for fewer than 20 years were using modern spacing methods.

We observed statistically significant (p < .05) interaction effects between proximity to conflict and three of the four independent variables: girl child marriage, past year physical IPV, and past year emotional IPV (see Table S1 for all interaction results). Additionally, there was an interaction effect trending toward significance (p = .10) between proximity to conflict and past year sexual IPV. Results of the multivariable regression models stratified by proximity to conflict are shown in table 2. We found that women married as children had increased relative risk of modern spacing contraceptives compared to women who married as adults across all levels of proximity to conflict (distal adjusted relative risk ratio/aRRR: 1.81, 95% confidence interval [CI] = [1.43, 2.30]; proximal aRRR: 2.05, CI = [1.44, 2.91]; central aRRR: 2.21, CI = [2.54, 3.18]). Women who experienced past year physical IPV had decreased relative risk of both sterilization (aRRR: 0.67, CI = [0.46, 0.96]) and traditional methods (aRRR: 0.63, CI = [0.43, 0.91]) compared to women who did not experience past year physical IPV in districts distal to conflict. Women who experienced past year emotional IPV had decreased relative risk of traditional methods (aRRR: 0.71, CI = [0.51, 0.95]) in districts distal to conflict and increased relative risk of modern spacing methods (aRRR: 1.50, CI = [1.08, 2.10]) in districts central to conflict compared to women who had not experienced past year emotional IPV. Women who experienced past year sexual IPV did not have statistically significant changes in the relative risk of any primary contraceptive method type compared to women who did not experience past year sexual IPV, and no form of past year IPV was associated with variation in primary contraceptive method type compared to women who did not experience that form of past year IPV in districts proximal to conflict. Finally, a sensitivity analysis excluding all respondents who named male sterilization as their primary method (n = 8) did not find any changes in the direction or significance of any of the measured associations calculated throughout our analyses.

Table 2.

Stratified Multinomial Logistic Regression Models Across Proximity to Conflict Comparing Relative Risk of Different Primary Contraceptive Method Types to No Method Among Currently Partnered Women Aged 18–49 Who Are Not Pregnant or Trying to Become Pregnant and Participated in the 2016 Sri Lanka DHS Domestic Violence Module. DOI: https://doi.org/10.1525/agh.2022.1539582.t2

Primary Contraceptive Method Type (Compared to No Method)
Modern SpacingSterilizationTraditional
Proximity to Conflict VariableRRRCIp ValueRRRCIp ValueRRRCIp Value
Distal (n = 6,373) 
 Girl child marriage 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 1.81 [1.43, 2.30] <.01* 1.31 [0.99, 1.72] .06 1.00 [0.75, 1.35] >.9 
 Past year sexual IPV 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 1.28 [0.70, 2.34] .42 1.41 [0.69, 2.88] .34 1.76 [0.85, 3.63] .12 
 Past year physical IPV 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 0.81 [0.61, 1.07] .14 0.67 [0.46, 0.96] .03* 0.63 [0.43, 0.91] .01* 
 Past year emotional IPV 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 0.94 [0.75, 1.19] .63 1.03 [0.77, 1.39] .84 0.71 [0.52, 0.95] .02* 
Proximal (n = 2,654) 
 Girl child marriage 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 2.05 [1.44, 2.91] <.01* 1.41 [0.95, 2.10] .09 0.69 [0.40, 1.19] .18 
 Past year sexual IPV 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 1.87 [0.63, 5.50] .26 1.53 [0.47, 4.99] .48 1.56 [0.31, 7.73] .59 
 Past year physical IPV 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 1.05 [0.59, 1.87] .86 1.43 [0.75, 2.75] .28 0.78 [0.34, 1.80] .55 
 Past year emotional IPV 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 0.98 [0.58, 1.71] .94 0.79 [0.42, 1.50] .48 0.61 [0.26, 1.39] .24 
Central (n = 2,399) 
 Girl child marriage 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 2.21 [1.54, 3.18] <.01* 1.11 [0.70, 1.74] .66 0.31 [0.10, 1.01] .05 
 Past year sexual IPV 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 0.67 [0.38, 1.16] .15 0.86 [0.45, 1.65] .66 1.38 [0.50, 3.83] .53 
 Past year physical IPV 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 1.19 [0.80, 1.76] .38 1.37 [0.83, 2.26] .22 0.85 [0.36, 2.01] .72 
 Past year emotional IPV 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 1.50 [1.08, 2.10] .02* 1.09 [0.69, 1.71] .71 1.69 [0.90, 3.15] .10 
Primary Contraceptive Method Type (Compared to No Method)
Modern SpacingSterilizationTraditional
Proximity to Conflict VariableRRRCIp ValueRRRCIp ValueRRRCIp Value
Distal (n = 6,373) 
 Girl child marriage 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 1.81 [1.43, 2.30] <.01* 1.31 [0.99, 1.72] .06 1.00 [0.75, 1.35] >.9 
 Past year sexual IPV 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 1.28 [0.70, 2.34] .42 1.41 [0.69, 2.88] .34 1.76 [0.85, 3.63] .12 
 Past year physical IPV 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 0.81 [0.61, 1.07] .14 0.67 [0.46, 0.96] .03* 0.63 [0.43, 0.91] .01* 
 Past year emotional IPV 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 0.94 [0.75, 1.19] .63 1.03 [0.77, 1.39] .84 0.71 [0.52, 0.95] .02* 
Proximal (n = 2,654) 
 Girl child marriage 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 2.05 [1.44, 2.91] <.01* 1.41 [0.95, 2.10] .09 0.69 [0.40, 1.19] .18 
 Past year sexual IPV 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 1.87 [0.63, 5.50] .26 1.53 [0.47, 4.99] .48 1.56 [0.31, 7.73] .59 
 Past year physical IPV 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 1.05 [0.59, 1.87] .86 1.43 [0.75, 2.75] .28 0.78 [0.34, 1.80] .55 
 Past year emotional IPV 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 0.98 [0.58, 1.71] .94 0.79 [0.42, 1.50] .48 0.61 [0.26, 1.39] .24 
Central (n = 2,399) 
 Girl child marriage 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 2.21 [1.54, 3.18] <.01* 1.11 [0.70, 1.74] .66 0.31 [0.10, 1.01] .05 
 Past year sexual IPV 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 0.67 [0.38, 1.16] .15 0.86 [0.45, 1.65] .66 1.38 [0.50, 3.83] .53 
 Past year physical IPV 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 1.19 [0.80, 1.76] .38 1.37 [0.83, 2.26] .22 0.85 [0.36, 2.01] .72 
 Past year emotional IPV 
  No Ref Ref Ref Ref Ref Ref Ref Ref Ref 
  Yes 1.50 [1.08, 2.10] .02* 1.09 [0.69, 1.71] .71 1.69 [0.90, 3.15] .10 

N = 11,426. RRR = relative risk ratio; CI = 95% confidence interval.

*Significant at α = .05.

Multivariate models included as covariates education, household wealth quintile, parity, age difference between woman and partner, person that makes decisions about health care, years married, household size, religion, urban setting, and district.

Our study adds depth to the research literature on contraceptive use in South Asia and in postconflict settings. We found that proximity to conflict moderates the associations of girl child marriage, past year physical IPV, and past year emotional IPV with primary contraceptive method type for women in Sri Lanka. Our findings imply that any interventions to influence Sri Lankan women’s contraceptive method use need to not only consider the impact of girl child marriage and IPV on primary contraceptive method type but also take into account the varying ways that proximity to conflict may impact these relationships.

We found that girl child marriage was associated with increased likelihood of modern contraceptive methods across all levels of proximity to conflict and was not significantly associated with changes in any other type of contraceptive method. Our findings only partially aligned with the findings of other studies in South Asia, which have found girl child marriage to be associated with both increased modern spacing methods [7] and increased female sterilization [7, 8, 24]. The lack of significant associations between girl child marriage and sterilization may reflect an acceptance of birth spacing and greater use of reversible contraceptive methods in Sri Lanka among women married as children, different from women married as children in other South Asian countries. The finding could also be explained by the widespread use of sterilization among all women in Sri Lanka regardless of their age at marriage. Our study is the first to examine associations between girl child marriage and contraceptive use in Sri Lanka and lays the groundwork for more in-depth explorations of the role of girl child marriage on the reproductive and sexual health of Sri Lankan women and girls.

Our study’s findings on associations between past year IPV and primary contraceptive method type add complexity to the literature on this topic from other countries across South Asia. We did not find any significant associations between past year sexual IPV and primary contraceptive method type, contradicting results from other South Asian research on IPV and contraception which found sexual IPV to be associated with increased use of modern spacing methods and decreased use of sterilization [10]. One possible explanation of the lack of influence of sexual IPV on contraceptive use in Sri Lanka might be increased societal acceptance of birth spacing and small family sizes, thus separating sexual violence from the reproductive coercion that it is associated with in other South Asian countries. This possible difference in family size preference might be influenced by Sri Lanka’s high rates of education for women compared to other South Asian countries, which may also explain Sri Lanka’s differing rates and consequences of IPV [47]. Past year physical IPV was only associated with changes in primary contraceptive method type in districts that were distal to conflict, where it was associated with decreased relative risk of both sterilization and traditional methods. This finding diverged from other research in South Asia, which found physical IPV to be associated with increased likelihood of using contraceptive methods [9, 48, 49]. One possible explanation for the decrease in sterilization and traditional methods but not modern spacing methods with physical IPV may be that women who experience physical IPV are less able to use methods that are nonreversible (such as sterilization) or which require partner cooperation (such as traditional methods) [13]. The idea that use of traditional methods may be less possible for women experiencing physical IPV is supported by qualitative research on Sri Lankan women who prefer traditional contraceptive methods and list the support and cooperation of their partners as a crucial factor for the success of traditional methods [50]. Finally, we found diverging associations between emotional IPV and primary contraceptive method type based on women’s proximity to conflict: In distal districts, emotional IPV was associated with a decreased likelihood of using traditional methods, while in central districts, it was associated with an increased likelihood of using modern spacing methods. The differing associations between each form of past year IPV and primary contraceptive method type in Sri Lanka reveal the importance of considering each form as distinct, with unique causes and impacts. This idea is supported by research on men’s perpetration of IPV in Sri Lanka, which found varying associations between adverse childhood experiences and the perpetration of sexual, physical, and emotional IPV [16].

The differing associations of physical and emotional IPV with primary contraceptive method type across proximity to conflict reveal the importance of considering these relationships across conflict exposure and postconflict place of residence. Past year physical IPV was not associated with any contraceptive method type in districts that were proximal or central to conflict despite being significantly associated with decreases in sterilization and traditional methods in districts distal to conflict. It is possible that prolonged exposure to armed conflict in districts that were central and proximal to conflict has in some way altered the perpetration and impact of physical IPV in these communities, where physical IPV may be more acceptable as a conflict resolution tactic [30, 31]. Past year emotional IPV was associated with decreased use of traditional methods in distal districts and increased use of modern spacing methods in central districts. Traditional methods such as withdrawal and the rhythm method require male partner communication and cooperation, which are less likely in association with emotional IPV [13, 50]. The increase in modern spacing methods among women experiencing emotional IPV in districts central to the conflict zone was the only IPV and contraceptive method association that was aligned with the literature from other contexts. It’s possible that in the conflict zone in Sri Lanka, women who experience emotional IPV feel less stigma seeking help, including modern spacing methods to delay pregnancy, compared to women who have experienced physical or sexual IPV. It is also possible that in the conflict zone, perpetrators of emotional IPV are more likely to attempt to use reproductive coercion on their partners, triggering covert use of modern spacing contraceptives by women experiencing their abuse. Finally, the positive association between past year emotional IPV and modern spacing contraceptive methods could be temporally reversed, and women who used modern spacing contraceptives may face more emotional abuse by partners who don’t approve of these methods in the conflict zone, possibly because of mismatched fertility desires and poor couple communication around contraception.

Limitations

This study had multiple limitations. Because all variables were collected during a cross-sectional survey with a lack of clear temporal difference between our independent and dependent variables, we cannot infer causality from the findings. This is a challenge shared across many DHS analyses [9]. It is also possible that women responding to the survey may have underreported IPV or girl child marriage due to social desirability bias or that women from districts with differing proximity to conflict recalled IPV experiences differently due to their varied contexts. The 2016 DHS only collected information on the most recent location that women had moved from; therefore, we did not have the complete data needed to analyze women’s lived experience of the conflict and cannot make conclusions about the direct impact of the conflict on survey participants. For this reason, our proximity to conflict variable must be interpreted in a postconflict and cross-sectional context, 7 years after the end of the Sri Lankan civil war, and as a geographic rather than experiential variable.

Because the DHS was designed to be nationally representative, we were not able to focus our analyses on specific cultural groups in the country with known higher rates of IPV and child marriage. One example of this is the small proportion of participants from the group labeled “Indian Tamils,” and defined as Tamils living in the up-country region of Sri Lanka and often working on tea “estates” (plantations) [2]. Although multiple studies focused on this population have highlighted high rates of IPV and sterilization compared to national averages [17, 39, 51, 52, 53, 54], the proportions in this study sample were too small to separate upcountry Tamils from other minority ethnic groups or to separate women living in estate areas from other nonurban participants in our analyses. Muslim communities in Sri Lanka also differentially experience child marriage as the only group for which it is legal [55]; however, in the context of this study, we were only able to include religion as a covariate rather than study it as an independent variable. More research should be conducted on the experiences of minority groups in Sri Lanka to better understand differing patterns of girl child marriage, IPV, and contraception that exist within them. Finally, we were limited in our analyses by the lack of certain individual-level data, including access to reproductive health care, distance from the nearest health center, location over one’s lifetime, and lifetime experience of IPV (rather than past year IPV). We suggest that future DHS implementations in Sri Lanka collect anonymized geolocation data and ask all participants about their locations over time, access to reproductive health care, and their lifetime experiences of IPV so that the impact of these variables can be examined in future analyses of contraceptive use.

Policy and Practice Implications

Our findings can inform policy and practice in the field of public health and gender equity in postconflict Sri Lanka. We found that child marriage and IPV can influence women’s primary contraceptive method type and that these relationships are moderated by proximity to conflict. These findings suggest that any interventions to increase women’s access to effective contraception in Sri Lanka will need to take the postconflict context into account [19]. The WHO suggests that policymakers developing reproductive health programs in conflict-affected areas use regional policies to reflect the needs of regions that have been differentially affected by the conflict, rather than a centralized approach, and programs need to be developed by or in collaboration with local community members, with special attention paid to ensure that the voices of marginalized groups are heard [29]. The WHO has also highlighted the additional need for reproductive health staff training and support in conflict-affected areas and to address gender-based violence [29]—this need has been shown to exist in Sri Lanka, where a large majority of nursing training programs include no instruction on IPV [56]. Finally, Sri Lankan researchers as well as the WHO recommend that contraceptive services be made available and accessible to all potential users, including those using traditional methods [4, 29]. Creating a welcoming environment for all who might need contraception will increase the likelihood that women across Sri Lanka will be able to access and use effective contraception.

In order for all women across Sri Lanka to have access to effective contraception, health care policymakers need to consider the impact of the gender inequities of girl child marriage and IPV on contraceptive use as well as the moderating influence of proximity to conflict. Programs need to be developed that take into account the regional differences in contraceptive use based on differential conflict exposure and address the needs of minority groups with unique experiences of contraception, girl child marriage, and IPV. Contraceptive service providers in Sri Lanka need to be adequately trained to mitigate conflict-related factors affecting contraception, as well as to address IPV and provide care to women married as children or using traditional methods of contraception. Sri Lanka can move closer to eliminating its postconflict health disparities by making contraceptive services accessible and welcoming to all.

RWF conceptualized the study, obtained data, performed analyses, and interpreted data for this work. JGS contributed to the study’s conceptualization, acquisition of data, and interpretation of the work, and provided significant critical revisions. LM contributed to conceptualization, acquisition of data, and interpretation. AR, ER, LU, and RL contributed to the design of the work, interpretation of data, and critical interpretations of the findings.

The authors would like to thank the Bill & Melinda Gates Foundation (grant number OPP1179208) for providing funding support to this study and the Department of Census and Statistics in Sri Lanka for collecting and sharing the data used in this study. Thank you to Drs. Vagisha Gunasekara and Ravindra Rannan-Eliya, Professors M. W. Amarasiri de Silva and Deborah DeGraff, and Emeritus Professor K. A. P. Siddhisena for providing assistance in requesting the data. Thank you to Professor Pia Axemo, Professor Kumudu Wijewardena, and Vathsala Illesinghe for providing feedback in the early stages of this project during discussions in Colombo, Sri Lanka. We are grateful to Nabamallika Dehingia and Arnab Dey for sharing their statistical expertise throughout this study. Finally, we would like to express our gratitude to the thousands of women who shared their time and personal information as participants in the 2016 Sri Lankan Demographic and Health Survey.

Data from the 2016 Sri Lankan Demographic and Health Survey can be requested from the Sri Lankan Office of the Census. Information on requesting this data can be found at the Department of Census and Statistics website: http://www.statistics.gov.lk.

The authors have no competing interests to declare.

LM, JGS and AR received funding support from the Bill & Melinda Gates Foundation (grant number OPP1179208) to complete this study. The funding agency played no role in the design, implementation, or interpretation of the results of this study.

Table S1.

1
Degraff
DS
,
Siddhisena
KAP
.
Unmet need for family planning in Sri Lanka: low enough or still an issue?
Int Perspect Sex Reprod Health
.
2015
;
41
(
4
):
200
209
.
Available
: http://www.ncbi.nlm.nih.gov/pubmed/26871728.
Accessed 14 February 2019
.
2
Department of Census and Statistics (DCS), Ministry of Health–Nutrition and Indigenous Medicine
.
Sri Lanka Demographic and Health Survey
2016
.
Colombo, Sri Lanka
:
Ministry of Health–Nutrition and Indigenous Medicine
;
2017
.
3
Matthews
Z
,
Padmadas
SS
,
Hutter
I
et al.
Does early childbearing and a sterilization-focused family planning programme in India fuel population growth?
Demogr Res
.
2009
;
20
:
693
720
.
4
Hettiarachchi
J
,
Gunawardena
NS
.
Factors related to choice of modern vs traditional contraceptives among women in rural Sri Lanka
.
Sri Lanka J Obstet Gynaecol
.
2012
;
33
(
1
):
20
.
5
Loaiza
Sr
E
,
Wong
S
.
Marrying Too Young: End Child Marriage
. Vol.
11
.
New York, NY
:
United Nations Population Fund
.
Available
: https://www.unfpa.org/sites/default/files/pub-pdf/MarryingTooYoung.pdf.
Accessed 31 January 2021
.
6
Grace
KT
,
Fleming
C
.
A systematic review of reproductive coercion in international settings
.
World Med Heal Policy
.
2016
;
8
(
4
):
382
408
.
7
Godha
D
,
Hotchkiss
DR
,
Gage
AJ
.
Association between child marriage and reproductive health outcomes and service utilization: a multi-country study from south Asia
.
J Adolesc Heal
.
2013
;
52
(
5
):
552
558
.
8
Raj
A
,
Saggurti
N
,
Balaiah
D
et al.
Prevalence of child marriage and its effect on fertility and fertility-control outcomes of young women in India: a cross-sectional, observational study
.
Lancet
.
2009
;
373
(
9678
):
1883
1889
.
9
Hindin
MJ
,
Kishor
S
,
Ansara
DL
.
Intimate Partner Violence Among Couples in 10 DHS Countries: Predictors and Health Outcomes
.
Macro International Incorporated
.
Available
https://www.popline.org/node/204198.
Accessed 4 November 2018
.
10
Raj
A
,
McDougal
L
,
Reed
E
et al.
Associations of marital violence with different forms of contraception: cross-sectional findings from South Asia
.
Int J Gynecol Obstet
.
2015
;
130
:
E56
E61
.
Available
: http://doi.wiley.com/10.1016/j.ijgo.2015.03.013.
Accessed 10 April 2020
.
11
Stockman
J
,
Campbell
J
,
Campbell
D
et al.
Sexual intimate partner violence, sexual risk behaviors, and contraceptive practices among women of African descent
.
Contraception
.
2010
;
82
(
2
):
212
.
12
McCloskey
LA
,
Doran
KA
,
Gerber
MR
.
Intimate partner violence is associated with voluntary sterilization in women
.
J Women’s Heal
.
2017
;
26
(
1
):
64
70
.
Available
: www.liebertpub.com.
Accessed 11 April 2020
.
13
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
.
14
Jayatilleke
AC
,
Poudel
KC
,
Yasuoka
J
et al.
Intimate partner violence in Sri Lanka
.
Biosci Trends
.
2010
;
4
(
3
):
90
95
.
Available
: http://www.ncbi.nlm.nih.gov/pubmed/20592458.
Accessed 14 February 2019
.
15
Jayasuriya
V
,
Wijewardena
K
,
Axemo
P
.
Intimate partner violence against women in the capital province of Sri Lanka: prevalence, risk factors, and help seeking
.
Violence Against Women
.
2011
;
17
(
8
):
1086
1102
.
Available
: http://journals.sagepub.com/doi/10.1177/1077801211417151.
Accessed 30 May 2019
.
16
Fonseka
RW
,
Minnis
AM
,
Gomez
AM
.
Impact of adverse childhood experiences on intimate partner violence perpetration among Sri Lankan men
.
PLoS One
.
2015
;
10
(
8
):
e0136321
.
Available
: http://www.ncbi.nlm.nih.gov/pubmed/26295577.
Accessed 11 June 2018
.
17
de Mel
N
,
Peiris
P
,
Gomez
S
.
Broadening Gender: Why Masculinities Matter
.
Colombo, Sri Lanka
:
Care International
;
2013
.
Available
: http://www.care.org/sites/default/files/documents/Broadening-Gender_Why-Masculinities-Matter.pdf.
18
Bandara
P
,
Knipe
D
,
Munasinghe
S
et al.
Socioeconomic and geographic correlates of intimate partner violence in Sri Lanka: analysis of the 2016 demographic and health survey
.
medRxiv
.
2021
.
Available
: https://doi.org/10.1101/2021.03.21.21254059.
Accessed 14 April 2021
.
19
Cometto
G
,
Fritsche
G
,
Sondorp
E
.
Health sector recovery in early post-conflict environments: experience from southern Sudan
.
Disasters
.
2010
;
34
(
4
):
885
909
.
Available
: http://doi.wiley.com/10.1111/j.1467-7717.2010.01174.x.
Accessed 12 May 2020
.
20
Warren
N
,
Alvarez
C
,
Makambo
MT
et al.
“Before the war we had it all”: family planning among couples in a post-conflict setting
.
Health Care Women Int
.
2017
;
38
(
8
):
796
812
.
Available
: https://www.tandfonline.com/doi/full/10.1080/07399332.2017.1329307.
Accessed 21 February 2019
.
21
McGinn
T
,
Austin
J
,
Anfinson
K
et al.
Family planning in conflict: results of cross-sectional baseline surveys in three African countries
.
Confl Health
.
2011
;
5
(
1
):
11
.
22
Neal
S
,
Stone
N
,
Ingham
R
.
The impact of armed conflict on adolescent transitions: a systematic review of quantitative research on age of sexual debut, first marriage and first birth in young women under the age of 20 years
.
BMC Public Health
.
2016
;
16
(
1
):
225
.
Available
: http://www.pcr.uu.se/research/UCDP.
Accessed 19 January 2020
.
23
Boerma
T
,
Tappis
H
,
Saad-Haddad
G
et al.
Armed conflicts and national trends in reproductive, maternal, newborn and child health in sub-Saharan Africa: what can national health surveys tell us?
BMJ Glob Heal
.
2019
;
4
(
Suppl 4
):
e001300
.
24
Raj
A
.
When the mother is a child: the impact of child marriage on the health and human rights of girls
.
Arch Dis Child
.
2010
;
95
:
931
935
.
Available
: http://www.ncbi.nlm.nih.gov/pubmed/20930011.
Accessed 25 February 2019
.
25
Mootz
JJ
,
Stabb
SD
,
Mollen
D
.
Gender-based violence and armed conflict: a community-informed socioecological conceptual model from Northeastern Uganda
.
Psychol Women Q
.
2017
;
41
(
3
):
368
388
.
Available
: http://journals.sagepub.com/doi/10.1177/0361684317705086.
Accessed 10 February 2021
.
26
Kinyanda
E
,
Weiss
HA
,
Mungherera
M
et al.
Intimate partner violence as seen in post-conflict eastern Uganda: prevalence, risk factors and mental health consequences
.
BMC Int Health Hum Rights
.
2016
;
16
(
1
):
5
.
Available
: http://bmcinthealthhumrights.biomedcentral.com/articles/10.1186/s12914-016-0079-x.
Accessed 21 February 2019
.
27
Kelly
JTD
,
Colantuoni
E
,
Robinson
C
et al.
From the battlefield to the bedroom: a multilevel analysis of the links between political conflict and intimate partner violence in Liberia
.
BMJ Glob Heal
.
2018
;
3
(
2
):
e000668
.
Available
: http://gh.bmj.com.
Accessed 21 February 2019
.
28
Ekhator-Mobayode
UE
,
Hanmer
LC
,
Matulevich
ECR
et al.
Effect of Armed Conflict on Intimate Partner Violence: Evidence From the Boko Haram Insurgency in Northeastern Nigeria
.
Washington, DC
:
World Bank Group Gender Global Theme
;
2020
. Vol.
9168
.
29
World Health Organization
.
Reproductive Health During Conflict and Displacement: A Guide for Programme Managers
.
Geneva, Switzerland
:
World Health Organization;
2000
.
30
Manjoo
R
,
McRaith
C
.
Gender-based violence and justice in conflict and post-conflict areas
.
Cornell Int’l LJ
.
2011
;
44
:
11
.
31
Saile
R
,
Ertl
V
,
Neuner
F
et al.
Does war contribute to family violence against children? Findings from a two-generational multi-informant study in Northern Uganda
.
Child Abuse Negl
.
2014
;
38
(
1
):
135
146
.
Available
: https://www.sciencedirect.com/science/article/pii/S0145213413003128?via%3Dihub.
Accessed 15 January 2020
.
32
Elbert
T
,
Schauer
M
,
Schauer
E
et al.
Trauma-related impairment in children—a survey in Sri Lankan provinces affected by armed conflict
.
Child Abuse Negl
.
2009
;
33
(
4
):
238
246
.
Available
: https://www.sciencedirect.com/science/article/pii/S0145213409000349.
Accessed 7 March 2019
.
33
Somasundaram
D
,
Sivayokan
S
.
Rebuilding community resilience in a post-war context: developing insight and recommendations—a qualitative study in Northern Sri Lanka
.
Int J Ment Health Syst
.
2013
;
7
(
1
):
3
.
Available
: https://ijmhs.biomedcentral.com/articles/10.1186/1752-4458-7-3
34
Johnson
SA
.
The cost of war on public health: an exploratory method for understanding the impact of conflict on public health in Sri Lanka
.
PLoS One
.
2017
;
12
(
1
):
166674
.
Available
: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0166674.
Accessed 8 May 2020
.
35
Jayasekara
RS
,
Schultz
T
.
Health status, trends, and issues in Sri Lanka
.
Nurs Heal Sci
.
2007
;
9
(
3
):
228
233
.
Available
: http://doi.wiley.com/10.1111/j.1442-2018.2007.00328.x.
Accessed 18 March 2021
.
36
McCall
C
.
Sri Lanka’s war wounds run deep
.
Lancet
.
2016
;
387
(
10032
):
1986
.
Available
: http://dx.doi.org/10.1016/S0140-6736(16)30523-2
37
Kottegoda
S
,
Samuel
K
,
Emmanuel
S
.
Reproductive health concerns in six conflict-affected areas of Sri Lanka
.
Reprod Health Matters
.
2008
;
16
(
31
):
75
82
.
Available
: https://www.tandfonline.com/action/journalInformation?journalCode=zrhm21.
Accessed 13 April 2020
.
38
Fokus Women
.
Post War Trends in Child Marriage: Sri Lanka
;
2015
.
Available
: https://docplayer.net/24861265-Post-war-trends-in-child-marriage-sri-lanka.html.
39
Senanayake
L
.
Sexual and Gender Based Violence. In: Monitoring & Evaluation Unit of the Family Planning Association of Sri Lanka, editor
.
Sexual and Reproductive Health Research in Sri Lanka: Current Status, Challenges, and Directions (2010-2019) [Internet]
.
2019
.
Available
: www.fpasrilanka.org.
Accessed 18 February 2021
.
40
Sengupta
S
.
War’s End in Sri Lanka: Bloody Family Triumph
.
The New York Times
.
19
May
2009
.
Available
: https://www.nytimes.com/2009/05/20/world/asia/20lanka.html
41
ICF International
.
Sampling and Household Listing Manual. Demographic and Health Surveys Methodology
.
Calverton, MD
:
ICF International
;
2012
.
42
Ellsberg
M
,
Heise
L
.
Researching Violence Against Women. A Practical Guide for Researchers and Activists
.
Washington, DC
:
PATH
;
2005
.
Available
: https://www.who.int/reproductivehealth/publications/violence/9241546476/en
43
Sinai
I
,
Igras
S
,
Lundgren
R
.
A practical alternative to calculating unmet need for family planning
.
Open Access J Contracept
.
2017
;
8
:
53
.
44
Chambers
J
.
Software for Data Analysis: Programming With R
.
Berlin, Germany
:
Springer Science & Business Media
;
2008
.
45
Lumley
T
.
Package “survey” Title Analysis of Complex Survey Samples [Internet]
.
Version 3.30-3. 2020
.
Available
: https://cran.r-project.org/web/packages/survey/survey.pdf
46
Baron
RM
,
Kenny
DA.
The moderator-mediator variable distinction in social psychological research. Conceptual, strategic, and statistical considerations
.
J Pers Soc Psychol
.
1986
;
51
(
6
):
1173
1182
.
47
Bhardwaj
N
,
Miller
J
.
Comparative cross-national analyses of domestic violence: insights from South Asia
.
Fem Criminol
.
2021
;
16
(
3
):
351
365
. https://doi.org/101177/1557085120987635.
Available
: https://journals.sagepub.com/doi/full/10.1177/1557085120987635.
Accessed 6 October 2021
.
48
Dalal
K
,
Andrews
J
,
Dawad
S
.
Contraception use and associations with intimate partner violence among women in Bangladesh
.
J Biosoc Sci
.
2012
;
44
(
1
):
83
94
.
49
Acharya
A
,
Weissman
A
,
Thapa
B
et al.
Intimate partner violence and contraceptive use among married women in Nepal
.
Int J Gynecol Obstet
.
2019
;
146
(
3
):
344
349
.
Available
: https://onlinelibrary.wiley.com/doi/abs/10.1002/ijgo.12883.
Accessed 11 April 2020
.
50
Perera
BN
.
Preference for traditional contraceptive use and women’s education: the case of Sri Lanka
.
Sri Lanka J Popul Stud
.
2014
;
14
:
115
129
.
51
Jegathesan
M
.
State-industrial entanglements in women’s reproductive capacity and labor in Sri Lanka
.
South Asia Multidiscip Acad J
.
2019
;
20
.
Available
: http://journals.openedition.org/samaj/5095.
Accessed 23 May 2019
.
52
Balasundaram
S
.
Stealing wombs: sterilization abuses and women’s reproductive health in Sri Lanka’s Tea Plantations
.
Indian Anthropol
.
2011
;
41
(
2
):
57
78
.
Available
: http://www.jstor.org/stable/41921991.
Accessed 18 May 2018
.
53
Fonseka
RW
.
Empowering youth in rural, up-country Sri Lanka through gender-equitable education and employment
.
Gend Dev
.
2018
;
26
(
3
):
569
585
.
Available
: https://www.tandfonline.com/doi/full/10.1080/13552074.2018.1523288
54
Department of Census and Statistics Sri Lanka
.
Women’s Wellbeing Survey—2019: Findings From Sri Lanka’s First Dedicated National Survey on Violence Against Women and Girls (Final Report) [Internet]
;
2020
.
Available
: https://asiapacific.unfpa.org/en/publications/sri-lanka-womens-wellbeing-survey-2019.
Accessed 30 November 2020
.
55
Muslim Marriage and Divorce Act [Internet]
.
1975
.
Available
: https://www.lawnet.gov.lk/marriage-and-divorce-muslim-4
56
Seneviratne
S
,
Guruge
S
,
Sivayogan
S
et al.
The status of intimate partner violence-related education for nurses in Sri Lanka: a cross-sectional survey of the nursing curricula
.
Open Univ Sri Lanka J
.
2020
;
15
(
2
):
19
44
.

How to cite this article: Fonseka RW, McDougal L, Raj A et al. Does proximity to conflict zones moderate associations between girl child marriage, intimate partner violence, and contraception in postconflict Sri Lanka? Adv Glob Health. 2022;1(1). DOI: https://doi.org/10.1525/agh.2022.1539582

Editor-in-Chief: Craig R. Cohen, University of California San Francisco, CA, USA

Senior Editor: Sarah Ssali, Makerere University, Uganda

Section: Achieving Gender Equality

This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International License (CC-BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. See http://creativecommons.org/licenses/by/4.0/.

Supplementary data