556108 research-article2014
JIVXXX10.1177/0886260514556108Journal of Interpersonal ViolenceThananowan and Vongsirimas
Article
Factors Mediating the Relationship Between Intimate Partner Violence and Cervical Cancer Among Thai Women
Journal of Interpersonal Violence 1–17 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/0886260514556108 jiv.sagepub.com
Nanthana Thananowan, RN, PhD,1 and Nopporn Vongsirimas, RN, PhD1
Abstract Previous research suggests that intimate partner violence (IPV), particularly physical or sexual violence, was associated with cervical cancer. However, there is less work examining the mechanism of the relationship between IPV and cervical cancer. The purpose of this cross-sectional study was to examine psychosocial factors (e.g., stress, social support, self-esteem, and depressive symptoms) as mediators of the relationship between IPV and cervical cancer among 532 Thai women with gynecological problems. About 21.1% of participants reported any type of IPV (e.g., physical, sexual, or emotional violence) in the past year and 22.2% had cervical cancer. IPV was significantly positively associated with stress, depressive symptoms, and cervical cancer but negatively correlated with social support and self-esteem. Results from structural equation modeling indicated that not only did IPV exhibit significantly direct effects on social support, stress, and depressive symptoms, and indirect effects on self-esteem, but it also had a significant, positive, total effect on cervical cancer. IPV exhibited the significant indirect effect on cervical cancer through social support, self-esteem, stress, and 1Mahidol
University, Bangkok, Thailand
Corresponding Author: Nanthana Thananowan, Department of Obstetrics and Gynecological Nursing, Faculty of Nursing, Mahidol University, 2 Siriraj Hospital, Prannok Rd, Bangkok 10700, Thailand. Email: [email protected]
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
2
Journal of Interpersonal Violence
depressive symptoms. The model fitted very well to the empirical data and explained 9% of variance. The findings affirmed that those psychosocial factors were mediators of the relationship between IPV and cervical cancer. Health care protocols for abused women should include screening for and treatment of IPV-related psychosocial factors. Interventions that provide social support and protect self-esteem should reduce stress and depressive symptoms among abused women, thereby reducing the risk of cervical cancer. Keywords abuse, cervical cancer, intimate partner violence, mediating, Thai women
Introduction Intimate partner violence (IPV), a pattern of physical, sexual, and/or emotional violence by an intimate partner in the context of coercive control (Tjaden & Thoennes, 2000), is a worldwide public health concern that predominantly affects women of reproductive age (Moracco, Runyan, Bowling, & Earp, 2007). Results from a growing body of research have consistently shown a significant association between IPV and women’s health. For example, physical and sexual assaults may result in fatal and nonfatal injuries (Wu, Huff, & Bhandari, 2010), chronic pain (Tiwari, Fong, Chan, & Ho, 2013), unwanted pregnancies (Bailey, 2010), sexually transmitted infections (STIs; Humphreys, 2011), vaginal irritation and discharge, pelvic inflammatory disease, and chronic pelvic pain (Campbell, Lichty, Sturza, & Raja, 2006), cervical cancer (Coker, Hopenhayn, DeSimone, Bush, & Crofford, 2009; Coker, Sanderson, Fadden, & Pirisi, 2000), and HIV infection (Sareen, Pagura, & Grant, 2009). Not only are physical and sexual abuse a substantial health consequence by the nature of its direct effects, but emotional abuse also contributes to women’s mental health, including fear, anxiety, stress, self-esteem, depressive symptoms, posttraumatic stress disorder, suicidal thoughts, and suicide attempts (Afifi et al., 2009; Lacey, McPherson, Samuel, Sears, & Head, 2013). These findings suggest that IPV has serious deleterious effects on women’s health that needed further investigation. Although IPV is associated with a range of adverse health outcomes, few studies have explored the relationship between IPV and cervical cancer. For example, Coker and colleagues (2000) found that women with cervical cancer reported experiencing more frequent physical and sexual assaults than did controls. Consequently, Coker et al. (2009) reported that experiencing violence was associated with an increased prevalence of invasive cervical cancer
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
Thananowan and Vongsirimas
3
(adjusted odds ratio = 2.6, 95% confidence interval [CI] = [1.7, 3.9]). Recently, Cesario, McFarlane, Nava, Gilroy, and Maddoux (2014) indicated that the prevalence of cervical cancer reported by abused women was 10 times higher than the general population. Taken together, IPV may influence cervical cancer risk indirectly through increasing chronic stress, which may suppress immune response to human papilloma virus (HPV) infection (Coker et al., 2009). In fact, stress has been conceptualized as an imbalance between environmental demands and individual resources (Chang et al., 2005). Such an imbalance can enhance stress perception and maladaptive emotional response, leading either directly or indirectly to adverse health outcomes (Beydoun & Saftlas, 2008). In addition, the high level of stress may increase likelihood of poor help-seeking behaviors (Fanslow & Robinson, 2010) and unhealthy behaviors, including cigarette smoking, drug use, and alcoholism, which are well-established risk factors for cervical cancer development (Ruback, Clark, & Warner, 2014). Previous research with the impact of psychosocial factors on cervical intraepithelial neoplasia reported that stress increased developing squamous intraepithelial lesions by seven-fold (Pereira et al., 2003). Given the established association between IPV and stress, and the association between stress and cervical cancer, it is reasonable to hypothesize that stress is a mediator of the relationship between IPV and cervical cancer. Besides stress, information about the significant mediators affecting the relationship between IPV and cervical cancer is limited. First, women with history of IPV were more common in those separated or divorced, with economic difficulties, and with low level of social support (Arco et al., 2013). A lack of social and economic resources inhibits women from seeking health care (Fanslow & Robinson, 2010). As a result, cervical cancer may go undetected because of low rates of health care use among victimized women (Modesitt et al., 2006). Thus, providing social support can buffer the negative effects of IPV, which in turn reduces the risk of cervical cancer among abused women. Second, previous research considered self-esteem to be a mediating factor that may affect women’s health because interpersonal victimization has been associated with low levels of self-esteem (Shen, 2009). To support this hypothesis, the result from a local study in Thailand indicated that abused women significantly reported low level of self-esteem and social support, but high level of stress and depressive symptoms (Thananowan & Vongsirimas, 2014). Although previous research examined the mediating effect of selfesteem in postpartum mothers (Hall, Kotch, Browne, & Rayens, 1996) and young adults (Shen, 2009), to date, research has yet to examine the mediator role of self-esteem among women with cervical cancer. Finally, abused women were 3.8 times more likely to experience depression when compared with women in general (Golding, 1999). In a survey of 82 women with depressive symptoms, the severity of IPV was significantly and positively Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
4
Journal of Interpersonal Violence
correlated with the severity of depression (Dienemann et al., 2000). Consequently, women with depressive symptoms were less likely to receive cervical cancer screening than those without the diagnosis (Weitlauf et al., 2013). Although abused women are at risk of depression, the nature of this relationship is unclear. Although there is evidence documenting the link between IPV and increased risk for cervical cancer, the mechanism of this link has not been fully explored. In addition, most studies have been undertaken in Western and developed countries. To our knowledge, no study has explored the multivariate relationships among the factors mediating IPV and cervical cancer in Thai women. Therefore, this study sought to examine psychosocial factors (e.g., stress, social support, self-esteem, and depressive symptoms) as mediators of the relationship between IPV and cervical cancer among Thai women with gynecological problems. Our hypothesis was that the effect of IPV on cervical cancer would be mediated by psychosocial factors. Such an understanding may inform intervention development to reduce the impacts of IPV on cervical cancer. Nurses and physicians can use this information to understand the underlying link between IPV and cervical cancer, thus facilitating more targeted care.
Method Design and Setting This study is a part of a larger study investigating IPV, health consequences, and coping patterns among Thai women with gynecological problems. A cross-sectional design was used to examine stress, social support, self-esteem, and depressive symptoms as mediators of IPV and cervical cancer. Convenience sampling was used to recruit potential participants from two gynecology wards of a university hospital in Bangkok, Thailand. These settings were selected because they served a large number of women with gynecological problems, including STIs, benign tumors, cervical cancer, and so on. A total of 562 women with gynecological problems were recruited by the principal investigator (PI); all of them responded to the survey, and therefore the response rate was 100%.
Participants The inclusion criteria were as follows: inpatient Thai women aged between 15 and 65 years, currently lived with a recent partner or separated from an intimate relationship, had a diagnosis with gynecological problems, did not
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
Thananowan and Vongsirimas
5
have a complicated health care need, and could read and write in Thai. Those who were unable to communicate in Thai or had the presence of their partners or other family members were excluded.
Procedures Ethical approval for conducting this study was obtained from the Institutional Review Board of the university hospital. This study was conducted from July 2011 to December 2012. To avoid situations that create susceptibility to harm or coercion, recruitment strategies included a staff nurse at each gynecology ward to identify eligible inpatient Thai women. In addition, the staff nurse had an opportunity to discuss any concerns with the patients individually prior to their participation. Then the potential participant was met by the PI and given a detailed description of the study and the rights of research participants, including voluntary participation, confidentiality, and the potential benefits and risks. For participants younger than 18, formal consent was sought from their guardians. To guarantee safety and confidentiality, the survey was introduced for those requiring approval by their guardians as a study of women’s health and life experiences. However, the participant herself has to be fully informed about the nature of the questions and the opportunity to either withdraw from the research project or not to answer the questionnaires (World Health Organization [WHO], 2001). Serial code numbers instead of participants’ names were used on each questionnaire. Given the sensitive nature of the issue, asking for a signature to confirm the informed consent will identify someone and possibly place that individual at risk; therefore, the participants could either sign an informed consent or give a verbal consent for participating in the study (WHO, 2007). Without the presence of intimate partners or other family members, the participants were invited to stay at their bedsides or in the health consulting room at each site to complete the set of questionnaires. After completion, participants put their questionnaires in a sealed box provided at nurses’ station. All participants received a list of services related to IPV and information about mental health services if they wished. Questionnaire administration averaged 30 min in length.
Measures Sample characteristics. The Demographic Characteristics Questionnaire (DCQ) included age (women and their partners), marital status, number of marriage, length of marriage, education, career, income, socioeconomic status, family structure, family relationship, and risk behaviors (e.g., drinking, smoking, and substance use).
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
6
Journal of Interpersonal Violence
Sexual risk behaviors. This questionnaire included age at first sex, number of sexual partners, and frequency of sexual intercourse, condom use, STIs, vaginal douche, and forced sexual intercourse by oral, anal, or during menstruation. IPV. Experiences of IPV were measured by two instruments: the Abuse Assessment Screen (AAS) and the Index of Spouse Abuse (ISA). This is because a participant may feel uncomfortable to identify their abuse status on the AAS questions rather than the ISA questions that assess for the severity of violence. First, the AAS (McFarlane, Parker, Soeken, & Bullock, 1992) was used to screen participants for IPV by answering “yes” (1) and “no” (0) for the following three questions: (a) “Within the last year, have you ever been hit, slapped, kicked, or otherwise physically hurt by someone?”; (b) “Within the last year, has anyone forced you to have sexual activities?”; and (c) “Are you afraid of your partner or anyone you listed above?” If a participant answered “yes” to having been physically, sexually, or emotionally abused in the past year, she was screened positive for IPV. The Cronbach’s alpha coefficient for this study was .84. Second, the ISA (Hudson & McIntosh, 1981) was used to measure the severity of past IPV while a woman was living with the intimate partner. Participants were asked to indicate the frequency of abusive acts over the past year using a 5-point Likert-type scale ranging from 1 (never) to 5 (very frequently). Summing scores range from 30 to 150. The Cronbach’s alpha coefficient was .97. Stress. The Stress Test (ST; Ministry of Public Health, 2000) was used to evaluate women at risk of stress by assessing signs, behaviors, or feelings during the last 2 months. The ST is a 20-item self-report instrument. Participants were asked to rate their stress level on a 4-point Likert-type scale ranging from 0 (none of the time) to 3 (almost or all of the time). Examples of items were presented in a previous publication (Thananowan & Vongsirimas, 2014). The higher the score, the more the woman is stressful. The Cronbach’s alpha coefficient for this study was .94. Social support. Women’s social support received from family, friends, and significant others was assessed using the Multidimensional Scale of Perceived Social Support (MSPSS; Zimet, Dahlem, Zimet, & Farley, 1988). The MSPSS is a 12-item self-report instrument. Participants were asked to rate their social support subscale items on a 7-point Likert-type scale ranging from 1 (very strongly disagree) to 7 (very strongly agree). A total score is calculated by summing across all 12 items with a possible range from 4 to 28. The higher the scores, the more the woman has a higher level of social support. The Cronbach’s alpha coefficient was .92.
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
Thananowan and Vongsirimas
7
Self-esteem. The Rosenberg’s Self-Esteem Scale (RSE) is a self-report instrument that assesses one’s own evaluation using 10 different items: a feeling of self-worth and self-respect (8 items) and a feeling of competence and ability (2 items; Rosenberg, 1989). Participants were asked to indicate the strength of their agreement in the last month on a 4-point Likert-type scale ranging from 1 (absolutely disagree) to 4 (absolutely agree). A total score is obtained by summing across each item score ranging from 10 to 40. The higher the score, the higher the woman’s self-assessed self-esteem. The Cronbach’s alpha coefficient for this study was .94. Depressive symptoms. The Center for Epidemiologic Studies Depression Scale (CES-D; Radloff, 1977) is a 20-item self-report instrument examining the frequency and duration of depressive symptoms. Participants were asked to rate their depressive symptoms over the past week on a 4-point Likert-type scale ranging from 0 (rarely or none of the time) to 3 (almost or all of the time). A total score is obtained by summing across each item score ranging from 0 to 60. Higher scores indicate greater depressive symptoms. In this study, a score of 16 was used as the cutoff for depressive symptoms. The Cronbach’s alpha coefficient for this study was .96. All instruments used in this study (except for the ST) were translated from English to Thai using a forward–backward translation method. The instruments were tested for content validity using sample group of Thai women with gynecological problems to ensure that they were culturally appropriate.
Statistical Analyses Sample size calculation was based on power analysis for structural equation modeling (SEM). The sample size tables yielded a power of at least 80%, with 80 degrees of freedom (df), the root mean square error of approximation (RMSEA) is .02 with .05 level of significance (Hancock & Freeman, 2001). The required sample size is at least 250 participants each for nonabused and abused groups. With 10% attrition rate, 550 potential participants were expected. Prior to analysis, all variables in the model satisfied the assumptions of normality and multicollinearity. That is, the linearity was assessed through the inspection of scatterplots among prior measured variables by SPSS program. Then the multicollinearity could be checked by inspecting the determinant of the covariance matrix. An extremely small determinant indicated a problem with multicollinearity (Tabachnick & Fidell, 2001). Descriptive analysis was conducted with SPSS version 13 to determine the prevalence of IPV, demographic characteristics, and sexual risk behaviors as well as gynecological problems. Pearson’s correlation analyses were
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
8
Journal of Interpersonal Violence
performed to investigate the relationship between the severity of IPV (i.e., ISA) and the proposed mediators (i.e., stress, social support, self-esteem, and depressive symptoms). According to the hypothesis, the hypothesized model was tested using SEM, a multivariate analysis technique, to determine how well the model can be explained by the empirical data. By using SEM, the hypothesized model was tested to estimate the relationships between observed and latent variables (the measurement model) and among latent variables themselves (the construct model), as well as direct and indirect effects. This process involved an evaluation of the hypothesis that the indicated measured items or scales reflected the latent constructs including stress, social support, self-esteem, depressive symptoms, IPV, and cervical cancer. Models for each construct were defined by permitting each of the relevant test or scales to load on a single factor representing the latent construct that it was hypothesized to measure. Maximum likelihood using LISREL version 8.72 was used for data analysis.
Results Overall, the 562 Thai women were recruited to the present study. Thirty questionnaires (5.6%) were excluded due to missing data, leaving a final sample size of 532 for data analysis. The prevalence of IPV, demographic characteristics, and sexual risk behaviors are displayed in Table 1. In summary, about 21.1% of participants (n = 112) had ever experienced any type of IPV in the past year. From the AAS, the prevalence of physical, sexual, and emotional abuse in the past year was 17.3% (n = 92), 11.5% (n = 61), and 13.2% (n = 70), respectively. The prevalence of gynecological problems was sorted from high to low as follows: benign tumors (37.6%), cervical cancer (22.2%), ectopic pregnancy (10.2%), ovarian cancer (9.8%), STIs (7.5%), endometrial cancer (5.5%), vulva cancer (3.6%), pelvic organ prolapses (2.8%), and menstrual disorders (0.8%). A correlation matrix of the major study variables is displayed in Table 2. The severity of IPV was significantly positively correlated with stress, depressive symptoms, and cervical cancer, but negatively correlated with social support and self-esteem. Stress was significantly positively correlated with depressive symptoms (r = .871, p < .01) and cervical cancer (r = .243, p < .01), but negatively correlated with social support (r = −.335, p < .01) and self-esteem (r = −.637, p < .01). In addition, depressive symptoms was highly correlated with self-esteem (r = −.693, p < .01). Several criteria were used to evaluate fit of the proposed model including the χ2 test, the comparative fit index (CFI), and the RMSEA. The generally agreed upon critical value for assessing model fit are CFI > 0.95, RMSEA < 0.06,
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
9
Thananowan and Vongsirimas Table 1. The Prevalence of IPV, Demographic Characteristics, Sexual Risk Behaviors. Characteristics Experienced IPV in the past year Physical abuse Sexual abuse Emotional abuse Women’s age (M, SD) Partners’ age (M, SD) Marital status Married Divorced/separated Widowed Cohabiting Number of marriages (M, SD) Once remarried Length of marriage (M, SD) Education ≤High school >High school Employment Unemployed Employed Housewife Household income (Thai baht) 30,001 Socioeconomic status Sufficient Insufficient In debt Family structure Living with husband and children Living with family members Living with others/friends Living alone
n
%
112 92 61 70 42.46 (12.174) 44.50 (12.937)
21.1 17.3 11.5 13.2
380 91 43 18 1.51 (1.142) 351 181 15.24 (11.021)
71.4 17.1 8.1 3.4 66 34
319 213
60 40
76 393 63
14.4 73.8 11.8
262 132 59 79
49.3 24.7 11.1 14.9
289 122 121
54.4 22.9 22.7
416 99 10 7
78.2 18.6 1.9 1.3 (continued)
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
10
Journal of Interpersonal Violence
Table 1. (continued) Characteristics Family relationship Good relationship Poor communication Some arguments Marital conflict Drink alcohol Smoking Substance use Age at first sex (M, SD) Number of sexual partners (M, SD) Forced sexual intercourse Having anal sex intercourse Having oral sex intercourse Having sex during menstruation
n
%
115 9 339 78 119 50 12 21.28 (5.309) 1.53 (1.212)
21.6 1.7 62 14.7 22.4 9.4 2.3
10 15 80
1.9 2.8 15
Note. IPV = intimate partner violence.
Table 2. Correlation Matrix of the Major Study Variables (n = 532).
1. ISA 2. ST 3. MSPSS 4. RSE 5. CES-D 6. Cervical cancer
1
2
1.000 .410** −.234** −.269** .351** .153**
1.000 −.335** −.637** .871** .243**
3
1.000 .567** −.341** −.073
4
1.000 −.693** −.184**
5
6
1.000 .257**
1.000
Note. ISA = Index of Spouse Abuse; ST = Stress Test; MSPSS = Multidimensional Scale of Perceived Social Support; RSE = Rosenberg’s Self-Esteem Scale; CES-D = Center for Epidemiologic Studies Depression Scale. *p < .05. **p < .01.
and p > .05 (Hu & Bentler, 1999). When the hypothesized model was examined by SEM, fit indices showed that the model fitted very well to the empirical data (χ2 = 140.77, df = 119, p = .084, CFI = 1.00, RMSEA = .019, goodness of fit index = .98, adjusted goodness of fit index = .95). The parameter estimates of the model are shown in Table 3 and Figure 1. Overall, this model accounted for 9% of the variance in cervical cancer. This is possible because
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
11
IE
TE
DE
IE
TE
R2 = .12
R2 = .45
−.26** — −.26** −.09 −.17* .25** .64* — .64*
DE
Self-Esteem (RSE) IE
TE
DE
IE
TE
Depressive Symptoms (CES-D) DE
IE
Cervical Cancer TE
.27** .16** .42** −.07* .41** .34** — .12** .12** .05 −.43* −.38* — −.41** −.41** .07 −.13 −.06 −.67** — −.67** −.10* −.61** −.71** −.03 −.20** −.23* .91** — 91** .30* — .30* R2 = .57 R2 = .91 R2 = .09
DE
Stress (ST)
Note. DE = standardized direct effect; IE = standardized indirect effect; TE = standardized total effect; MSPSS = Multidimensional Scale of Perceived Social Support; RSE = Rosenberg’s Self-Esteem Scale; ST = Stress Test; CES-D = Center for Epidemiologic Studies Depression Scale. *p < .05. **p < .01.
IPV MSPSS RSE ST Structural equation fit
Causal Variables
Social Support (MSPSS)
Effected Variables
Table 3. Standardized DE, IE, TE of Latent Variables in the Model.
12
Journal of Interpersonal Violence Social Support -0.26**
-0.09
IPV
0.27**
-0.07*
0.07
0.64*
Self-Esteem -0.67**
Stress
0.05
-0.03
Cervical Cancer
0.30* -0.10*
0.91**
Depressive Symptoms
Figure 1. Proposed mediators of the relationship between IPV and cervical cancer. Note. IPV = intimate partner violence.
the relationship between IPV (i.e., ISA) and cervical cancer is very low (r = .153, p < .01; Table 2). As shown in Table 3, results indicated that not only did IPV exhibit direct effects on social support (β = −.26; p < .01), stress (β = .27; p < .01), depressive symptoms (β = −.07; p < .05) and indirect effects on self-esteem (β = −.17; p < .05), but it also had a significant, positive, total effect on cervical cancer (β = .12; p < .01). Importantly, the indirect effect of IPV through the mediating variables (i.e., social support, selfesteem, stress, and depressive symptoms) on cervical cancer was statistically significant (β = .12; p < .01).
Discussion The prevalence of IPV in the past year among Thai women with gynecological problems (21.1%) was similar to the U.S. study (25%; Tjaden & Thoennes, 2000). However, the prevalence of physical abuse (17.3%) and sexual abuse (11.5%) in this study was higher than a population-based study that reported the prevalence of physical and sexual abuse to be about 5.7% and 3.7%, respectively (Nur, 2012). A possible explanation may involve sampling bias because abused women may come to the hospital with gynecological problems, resulting from physical or sexual abuse, and thus the prevalence of IPV in women in clinical settings would be expected to be higher than in women
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
Thananowan and Vongsirimas
13
in population settings. In addition, the use of a self-administered anonymous approach allowed women to feel more comfortable in revealing sensitive information about their abusive experiences. The prevalence of physical abuse in this study was also higher than the prevalence of sexual or emotional abuse. This was because physical abuse can result in injuries, while sexual or emotional abuse may lead to other health problems that are not directly linked to acts of violence. The findings in Table 2 highlighted the complex interrelationships among IPV, psychosocial factors, and cervical cancer. Therefore, women who experience IPV tend to have high stress, which in turn increases the likelihood of depressive symptoms but decreases social support and selfesteem. As a result, abused women may have limited access to health care resources and/or to undergo cervical cancer screening (Johnson et al., 2007). However, further investigation is needed to determine the underlying mechanisms that explain the relationship between IPV and inadequate cervical cancer screening. The SEM results generally provide a set of relationships in which the causal connections between several variables are examined simultaneously. Such results are important to understand the complex interrelationship among IPV, psychosocial factors, and cervical cancer. In addition, the SEM results provide the direction and strength of cause-and-effect relationships hypothesized by the researchers and also calculate the direct and indirect effects that independent variables have on the dependent variables. Although this model explained a small proportion of the variance (9%), IPV exhibited the significant indirect effect on cervical cancer through social support, selfesteem, stress, and depressive symptoms. Thus, our hypothesis was supported because the SEM results showed the mediation effects of the mediators on the outcome variables (i.e., cervical cancer). Although the direct effect between IPV and cervical cancer did not exist, the indirect effect and total effects reveals significant. It is probable that an increase in IPV could aggravate stress and depressive symptoms, which in turn could lead to worsening of cervical cancer, whereas social support and self-esteem can buffer the negative effects of victimization. Therefore, it is important for health care providers not to overlook IPV among inpatient Thai women with stress and depressive symptoms. Health care protocols for abused Thai women should include screening for and treatment of IPV-related psychosocial factors to prevent the risk of cervical cancer. In addition, screening women in health care settings for a history of IPV and psychosocial factors can link women to resources and treatment for assault-related health symptoms. And final, interventions designed to provide social support and protect
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
14
Journal of Interpersonal Violence
self-esteem should reduce stress and depressive symptoms among abused Thai women, thereby reducing the risk of cervical cancer. This study had some limitations that need to be acknowledged. First, the use of convenience samples in clinical settings may limit the generalizability of the findings to other populations. For future research, the model should be tested with more diverse and more representative samples. Second, because of the cross-sectional design, no inferences could be made with respect to causality. Therefore, longitudinal studies are essential for verifying the causal link between IPV and cervical cancer. Third, potential confounders in Table 1 (e.g., demographic characteristics and sexual risk behaviors) were not controlled and as a result they may impact women’s psychosocial factors and/or cervical cancer. Finally, all data were collected using self-report measures that may reflect biased and/or limited recall. Future research might consider the use of multiple data collection methods, including salivary cortisol when appropriate.
Conclusion The present study served as a starting point for future research of mediating effects in Thailand. Overall, the findings revealed the significant mediating effects of psychosocial factors on the relationship between IPV and cervical cancer, thus affirming the complex interrelationships among IPV, psychosocial factors, and cervical cancer. Because women in their reproductive years make many visits to health care providers, an important opportunity to screen for and intervene in either IPV or stress is available. The findings were also clinically important for health care providers when designing psychosocial interventions (e.g., stress reduction program with problem-focused coping) for abused Thai women with cervical cancer. Acknowledgment The authors would like to thank Thai women who participated in this study and Dr. Det Kedcham for his assistance on data analysis.
Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was granted by China Medical Board (CMB), Faculty of Nursing, Mahidol University, Bangkok, Thailand.
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
Thananowan and Vongsirimas
15
References Afifi, T. O., MacMillan, H., Cox, B. J., Asmundson, G., Stein, M. B., & Sareen, J. (2009). Mental health correlates of intimate partner violence in marital relationships in a nationally representative sample of males and females. Journal of Interpersonal Violence, 24, 1398-1417. Arco, D. A., Amo, J., Garcia-Pina, R., Garcia-Fulgueiras, A. M., Rodrigrez-Arenas, M. A., Ibaňez-Rojo, V., & Llacer, A. (2013). Violence in adulthood and mental health: Gender and immigrant status. Journal of Interpersonal Violence, 28, 2203-2222. Bailey, B. A. (2010). Partner violence during pregnancy: Prevalence, effects, screening, and management. International Journal of Women’s Health, 2, 183-197. Beydoun, H. A., & Saftlas, A. F. (2008). Physical and mental health outcomes of prenatal maternal stress in human and animal studies: A review of recent evidence. Paediatric and Perinatal Epidemiology, 22, 438-466. Campbell, R., Lichty, L. F., Sturza, M., & Raja, S. (2006). Gynecologic health impact of sexual assault. Research in Nursing & Health, 29, 399-413. Cesario, S. K., McFarlane, J., Nava, A., Gilroy, H., & Maddoux, J. (2014). Linking cancer and intimate partner violence: The importance of screening women in the oncology setting. Clinical Journal of Oncology Nursing, 18, 65-73. Chang, J. C., Cluss, P. A., Ranieri, L., Hawker, L., Buranosky, R., Dado, D., . . .Scholle, S. H. (2005). Health care interventions for intimate partner violence: What women want. Women’s Health Issues, 15, 21-30. Coker, A. L., Hopenhayn, C., DeSimone, C. P., Bush, H. M., & Crofford, L. (2009). Violence against women raises risk of cervical cancer. Journal of Women’s Health, 18, 1179-1185. Coker, A. L., Sanderson, M., Fadden, M. K., & Pirisi, L. (2000). Intimate partner violence and cervical neoplasia. Journal of Women’s Health & Gender-Based Medicine, 9, 1015-1023. Dienemann, J., Boyle, E., Baker, D., Resnick, W., Wiederhorn, N., & Campbell, J. (2000). Intimate partner abuse among women diagnosed with depression. Issues in Mental Health Nursing, 21, 499-513. Fanslow, J. L., & Robinson, E. M. (2010). Help-seeking behaviors and reasons for help seeking reported by a representative sample of women victims of intimate partner violence in New Zealand. Journal of Interpersonal Violence, 25, 929-951. Golding, J. M. (1999). Intimate partner violence as a risk factor for mental disorders: A meta-analysis. Journal of Family Violence, 14, 99-132. Hall, L. A., Kotch, J. B., Browne, D., & Rayens, M. K. (1996). Self-esteem as a mediator of the effects of stressors and social resources on depressive symptoms in postpartum mothers. Nursing Research, 45, 231-238. Hancock, G. R., & Freeman, M. J. (2001). Power and sample size for the Root Mean Square Error of Approximation Test of not close fit in Structural Equation Modeling (SEM). Educational and Psychological Measurement, 61, 741-758.
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
16
Journal of Interpersonal Violence
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternative. Structure Equation Modeling: A Multidisciplinary Journal, 6, 1-55. Hudson, W. W., & McIntosh, S. R. (1981). The assessment of spouse abuse: Two quantifiable dimensions. Journal of Marriage and Family, 43, 873-888. Humphreys, J. (2011). Sexually transmitted infections, pregnancy, and intimate partner violence. Health Care for Women International, 32, 23-38. Johnson, J. K., John, R., Humera, A., Kukreja, S., Found, M., & Lindow, S. W. (2007). The prevalence of emotional abuse in gynecology patients and its association with gynecological symptoms. European Journal of Obstetrics & Gynecology and Reproductive Biology, 133, 95-99. Lacey, K. K., McPherson, M. D., Samuel, P. S., Sears, K. P., & Head, D. (2013). The impact of different types of intimate partner violence on the mental and physical health of women in different ethnic groups. Journal of Interpersonal Violence, 28, 359-385. McFarlane, J. M., Parker, B., Soeken, K., & Bullock, L. (1992). Assessing for abuse during pregnancy: Severity and frequency of injuries and associated entry into prenatal care. Journal of the American Medical Association, 267, 3176-3178. Ministry of Public Health. (2000). The Stress Test. In Handbook of stress reduction. Bangkok, Thailand: Ministry of Public Health. Modesitt, S. C., Gambrell, A. C., Cottrill, H. M., Hays, L. R., Walker, R., Shelton, B. J., & Ferguson, J. E. (2006). Adverse impact of a history of violence for women with breast, cervical, endometrial, or ovarian cancer. Obstetrics & Gynecology, 107, 1330-1336. Moracco, K. E., Runyan, C. W., Bowling, J. M., & Earp, J. A. L. (2007). Women’s experiences with violence: A national study. Women’s Health Issues, 17, 3-12. Nur, N. (2012). The effect of intimate partner violence on mental health status among women of reproductive ages: A population-based study in a middle Anatolian city. Journal of Interpersonal Violence, 27, 3236-3251. Pereira, D. B., Antoni, M. H., Danielson, A., Simon, T., Efantis-Potter, J., Carver, C. S., & O’Sullivan, M. J. (2003). Life stress and cervical squamous intraepithelial lesions in women with human papillomavirus and human immunodeficiency virus. Psychosomatic Medicine, 65, 427-434. Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychology Measurement, 1, 385-401. Rosenberg, M. (1989). Society and the adolescent self-image. Middletown, CT: Wesleyan University Press. Ruback, R. B., Clark, V. A., & Warner, C. (2014). Why are crime victims at risk of being victimized again? Substance use, depression, and offending as mediators of the victimization–revictimization link. Journal of Interpersonal Violence, 29, 157-185. Sareen, J., Pagura, J., & Grant, B. (2009). Is intimate partner violence associated with HIV infection among women in the United States? General Hospital Psychiatry, 31, 274-278.
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
Thananowan and Vongsirimas
17
Shen, A. C. (2009). Self-esteem of young adults experiencing interparental violence and child physical maltreatment: Parental and peer relationship as mediators. Journal of Interpersonal Violence, 24, 770-794. Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistic (4th ed.). Boston, MA: Allyn & Bacon. Thananowan, N., & Vongsirimas, N. (2014). Association between intimate partner violence and women’s mental health: Survey evidence from Thailand. Pacific Rim International Journal of Nursing Research, 18, 3-15. Tiwari, A., Fong, D., Chan, C., & Ho, P. (2013). Factors mediating the relationship between intimate partner violence and chronic pain in Chinese women. Journal of Interpersonal Violence, 28, 1067-1087. Tjaden, P., & Thoennes, N. (2000). Extent, nature, and consequences of intimate partner violence: Findings from the National Violence Against Women Survey. Washington, DC: U.S. Department of Justice. Weitlauf, J. C., Jones, S., Xu, X., Finney, J. W., Moos, R. H., Sawya, G. F., & Frayne, S. M. (2013). Receipt of cervical cancer screening in female veterans: Impact of posttraumatic stress disorder and depression. Women’s Health Issues, 23, e153-e159. World Health Organization. (2001). Putting women first: Ethical and safety recommendations for research on domestic violence against women. Geneva, Switzerland: Author. World Health Organization. (2007). Ethical and safety recommendations for researching, documenting, and monitoring sexual violence in emergencies. Geneva, Switzerland: Author. Wu, V., Huff, H., & Bhandari, M. (2010). Pattern of physical injury associated with intimate partner violence in women presenting to the emergency department: A systematic review and meta-analysis. Trauma, Violence, & Abuse, 11, 71-82. Zimet, G. D., Dahlem, N. W., Zimet, S. G., & Farley, G. K. (1988). The Multidimensional Scale of perceived social support. Journal of Personal Assessment, 52, 30-41.
Author Biographies Nanthana Thananowan, RN, PhD, is an associate professor in the Department of Obstetrics and Gynecological Nursing at Faculty of Nursing, Mahidol University, Bangkok, Thailand. Her research interest includes intimate partner violence and women’s health. Nopporn Vongsirimas, RN, PhD, is an assistant professor in the Department of Mental Health and Psychiatric Nursing at Faculty of Nursing, Mahidol University, Bangkok, Thailand. Her research focuses on depressive symptoms among adolescents and women exposed to violence.
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
JIVXXX10.1177/0886260514556108Journal of Interpersonal ViolenceThananowan and Vongsirimas
Article
Factors Mediating the Relationship Between Intimate Partner Violence and Cervical Cancer Among Thai Women
Journal of Interpersonal Violence 1–17 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/0886260514556108 jiv.sagepub.com
Nanthana Thananowan, RN, PhD,1 and Nopporn Vongsirimas, RN, PhD1
Abstract Previous research suggests that intimate partner violence (IPV), particularly physical or sexual violence, was associated with cervical cancer. However, there is less work examining the mechanism of the relationship between IPV and cervical cancer. The purpose of this cross-sectional study was to examine psychosocial factors (e.g., stress, social support, self-esteem, and depressive symptoms) as mediators of the relationship between IPV and cervical cancer among 532 Thai women with gynecological problems. About 21.1% of participants reported any type of IPV (e.g., physical, sexual, or emotional violence) in the past year and 22.2% had cervical cancer. IPV was significantly positively associated with stress, depressive symptoms, and cervical cancer but negatively correlated with social support and self-esteem. Results from structural equation modeling indicated that not only did IPV exhibit significantly direct effects on social support, stress, and depressive symptoms, and indirect effects on self-esteem, but it also had a significant, positive, total effect on cervical cancer. IPV exhibited the significant indirect effect on cervical cancer through social support, self-esteem, stress, and 1Mahidol
University, Bangkok, Thailand
Corresponding Author: Nanthana Thananowan, Department of Obstetrics and Gynecological Nursing, Faculty of Nursing, Mahidol University, 2 Siriraj Hospital, Prannok Rd, Bangkok 10700, Thailand. Email: [email protected]
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
2
Journal of Interpersonal Violence
depressive symptoms. The model fitted very well to the empirical data and explained 9% of variance. The findings affirmed that those psychosocial factors were mediators of the relationship between IPV and cervical cancer. Health care protocols for abused women should include screening for and treatment of IPV-related psychosocial factors. Interventions that provide social support and protect self-esteem should reduce stress and depressive symptoms among abused women, thereby reducing the risk of cervical cancer. Keywords abuse, cervical cancer, intimate partner violence, mediating, Thai women
Introduction Intimate partner violence (IPV), a pattern of physical, sexual, and/or emotional violence by an intimate partner in the context of coercive control (Tjaden & Thoennes, 2000), is a worldwide public health concern that predominantly affects women of reproductive age (Moracco, Runyan, Bowling, & Earp, 2007). Results from a growing body of research have consistently shown a significant association between IPV and women’s health. For example, physical and sexual assaults may result in fatal and nonfatal injuries (Wu, Huff, & Bhandari, 2010), chronic pain (Tiwari, Fong, Chan, & Ho, 2013), unwanted pregnancies (Bailey, 2010), sexually transmitted infections (STIs; Humphreys, 2011), vaginal irritation and discharge, pelvic inflammatory disease, and chronic pelvic pain (Campbell, Lichty, Sturza, & Raja, 2006), cervical cancer (Coker, Hopenhayn, DeSimone, Bush, & Crofford, 2009; Coker, Sanderson, Fadden, & Pirisi, 2000), and HIV infection (Sareen, Pagura, & Grant, 2009). Not only are physical and sexual abuse a substantial health consequence by the nature of its direct effects, but emotional abuse also contributes to women’s mental health, including fear, anxiety, stress, self-esteem, depressive symptoms, posttraumatic stress disorder, suicidal thoughts, and suicide attempts (Afifi et al., 2009; Lacey, McPherson, Samuel, Sears, & Head, 2013). These findings suggest that IPV has serious deleterious effects on women’s health that needed further investigation. Although IPV is associated with a range of adverse health outcomes, few studies have explored the relationship between IPV and cervical cancer. For example, Coker and colleagues (2000) found that women with cervical cancer reported experiencing more frequent physical and sexual assaults than did controls. Consequently, Coker et al. (2009) reported that experiencing violence was associated with an increased prevalence of invasive cervical cancer
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
Thananowan and Vongsirimas
3
(adjusted odds ratio = 2.6, 95% confidence interval [CI] = [1.7, 3.9]). Recently, Cesario, McFarlane, Nava, Gilroy, and Maddoux (2014) indicated that the prevalence of cervical cancer reported by abused women was 10 times higher than the general population. Taken together, IPV may influence cervical cancer risk indirectly through increasing chronic stress, which may suppress immune response to human papilloma virus (HPV) infection (Coker et al., 2009). In fact, stress has been conceptualized as an imbalance between environmental demands and individual resources (Chang et al., 2005). Such an imbalance can enhance stress perception and maladaptive emotional response, leading either directly or indirectly to adverse health outcomes (Beydoun & Saftlas, 2008). In addition, the high level of stress may increase likelihood of poor help-seeking behaviors (Fanslow & Robinson, 2010) and unhealthy behaviors, including cigarette smoking, drug use, and alcoholism, which are well-established risk factors for cervical cancer development (Ruback, Clark, & Warner, 2014). Previous research with the impact of psychosocial factors on cervical intraepithelial neoplasia reported that stress increased developing squamous intraepithelial lesions by seven-fold (Pereira et al., 2003). Given the established association between IPV and stress, and the association between stress and cervical cancer, it is reasonable to hypothesize that stress is a mediator of the relationship between IPV and cervical cancer. Besides stress, information about the significant mediators affecting the relationship between IPV and cervical cancer is limited. First, women with history of IPV were more common in those separated or divorced, with economic difficulties, and with low level of social support (Arco et al., 2013). A lack of social and economic resources inhibits women from seeking health care (Fanslow & Robinson, 2010). As a result, cervical cancer may go undetected because of low rates of health care use among victimized women (Modesitt et al., 2006). Thus, providing social support can buffer the negative effects of IPV, which in turn reduces the risk of cervical cancer among abused women. Second, previous research considered self-esteem to be a mediating factor that may affect women’s health because interpersonal victimization has been associated with low levels of self-esteem (Shen, 2009). To support this hypothesis, the result from a local study in Thailand indicated that abused women significantly reported low level of self-esteem and social support, but high level of stress and depressive symptoms (Thananowan & Vongsirimas, 2014). Although previous research examined the mediating effect of selfesteem in postpartum mothers (Hall, Kotch, Browne, & Rayens, 1996) and young adults (Shen, 2009), to date, research has yet to examine the mediator role of self-esteem among women with cervical cancer. Finally, abused women were 3.8 times more likely to experience depression when compared with women in general (Golding, 1999). In a survey of 82 women with depressive symptoms, the severity of IPV was significantly and positively Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
4
Journal of Interpersonal Violence
correlated with the severity of depression (Dienemann et al., 2000). Consequently, women with depressive symptoms were less likely to receive cervical cancer screening than those without the diagnosis (Weitlauf et al., 2013). Although abused women are at risk of depression, the nature of this relationship is unclear. Although there is evidence documenting the link between IPV and increased risk for cervical cancer, the mechanism of this link has not been fully explored. In addition, most studies have been undertaken in Western and developed countries. To our knowledge, no study has explored the multivariate relationships among the factors mediating IPV and cervical cancer in Thai women. Therefore, this study sought to examine psychosocial factors (e.g., stress, social support, self-esteem, and depressive symptoms) as mediators of the relationship between IPV and cervical cancer among Thai women with gynecological problems. Our hypothesis was that the effect of IPV on cervical cancer would be mediated by psychosocial factors. Such an understanding may inform intervention development to reduce the impacts of IPV on cervical cancer. Nurses and physicians can use this information to understand the underlying link between IPV and cervical cancer, thus facilitating more targeted care.
Method Design and Setting This study is a part of a larger study investigating IPV, health consequences, and coping patterns among Thai women with gynecological problems. A cross-sectional design was used to examine stress, social support, self-esteem, and depressive symptoms as mediators of IPV and cervical cancer. Convenience sampling was used to recruit potential participants from two gynecology wards of a university hospital in Bangkok, Thailand. These settings were selected because they served a large number of women with gynecological problems, including STIs, benign tumors, cervical cancer, and so on. A total of 562 women with gynecological problems were recruited by the principal investigator (PI); all of them responded to the survey, and therefore the response rate was 100%.
Participants The inclusion criteria were as follows: inpatient Thai women aged between 15 and 65 years, currently lived with a recent partner or separated from an intimate relationship, had a diagnosis with gynecological problems, did not
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
Thananowan and Vongsirimas
5
have a complicated health care need, and could read and write in Thai. Those who were unable to communicate in Thai or had the presence of their partners or other family members were excluded.
Procedures Ethical approval for conducting this study was obtained from the Institutional Review Board of the university hospital. This study was conducted from July 2011 to December 2012. To avoid situations that create susceptibility to harm or coercion, recruitment strategies included a staff nurse at each gynecology ward to identify eligible inpatient Thai women. In addition, the staff nurse had an opportunity to discuss any concerns with the patients individually prior to their participation. Then the potential participant was met by the PI and given a detailed description of the study and the rights of research participants, including voluntary participation, confidentiality, and the potential benefits and risks. For participants younger than 18, formal consent was sought from their guardians. To guarantee safety and confidentiality, the survey was introduced for those requiring approval by their guardians as a study of women’s health and life experiences. However, the participant herself has to be fully informed about the nature of the questions and the opportunity to either withdraw from the research project or not to answer the questionnaires (World Health Organization [WHO], 2001). Serial code numbers instead of participants’ names were used on each questionnaire. Given the sensitive nature of the issue, asking for a signature to confirm the informed consent will identify someone and possibly place that individual at risk; therefore, the participants could either sign an informed consent or give a verbal consent for participating in the study (WHO, 2007). Without the presence of intimate partners or other family members, the participants were invited to stay at their bedsides or in the health consulting room at each site to complete the set of questionnaires. After completion, participants put their questionnaires in a sealed box provided at nurses’ station. All participants received a list of services related to IPV and information about mental health services if they wished. Questionnaire administration averaged 30 min in length.
Measures Sample characteristics. The Demographic Characteristics Questionnaire (DCQ) included age (women and their partners), marital status, number of marriage, length of marriage, education, career, income, socioeconomic status, family structure, family relationship, and risk behaviors (e.g., drinking, smoking, and substance use).
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
6
Journal of Interpersonal Violence
Sexual risk behaviors. This questionnaire included age at first sex, number of sexual partners, and frequency of sexual intercourse, condom use, STIs, vaginal douche, and forced sexual intercourse by oral, anal, or during menstruation. IPV. Experiences of IPV were measured by two instruments: the Abuse Assessment Screen (AAS) and the Index of Spouse Abuse (ISA). This is because a participant may feel uncomfortable to identify their abuse status on the AAS questions rather than the ISA questions that assess for the severity of violence. First, the AAS (McFarlane, Parker, Soeken, & Bullock, 1992) was used to screen participants for IPV by answering “yes” (1) and “no” (0) for the following three questions: (a) “Within the last year, have you ever been hit, slapped, kicked, or otherwise physically hurt by someone?”; (b) “Within the last year, has anyone forced you to have sexual activities?”; and (c) “Are you afraid of your partner or anyone you listed above?” If a participant answered “yes” to having been physically, sexually, or emotionally abused in the past year, she was screened positive for IPV. The Cronbach’s alpha coefficient for this study was .84. Second, the ISA (Hudson & McIntosh, 1981) was used to measure the severity of past IPV while a woman was living with the intimate partner. Participants were asked to indicate the frequency of abusive acts over the past year using a 5-point Likert-type scale ranging from 1 (never) to 5 (very frequently). Summing scores range from 30 to 150. The Cronbach’s alpha coefficient was .97. Stress. The Stress Test (ST; Ministry of Public Health, 2000) was used to evaluate women at risk of stress by assessing signs, behaviors, or feelings during the last 2 months. The ST is a 20-item self-report instrument. Participants were asked to rate their stress level on a 4-point Likert-type scale ranging from 0 (none of the time) to 3 (almost or all of the time). Examples of items were presented in a previous publication (Thananowan & Vongsirimas, 2014). The higher the score, the more the woman is stressful. The Cronbach’s alpha coefficient for this study was .94. Social support. Women’s social support received from family, friends, and significant others was assessed using the Multidimensional Scale of Perceived Social Support (MSPSS; Zimet, Dahlem, Zimet, & Farley, 1988). The MSPSS is a 12-item self-report instrument. Participants were asked to rate their social support subscale items on a 7-point Likert-type scale ranging from 1 (very strongly disagree) to 7 (very strongly agree). A total score is calculated by summing across all 12 items with a possible range from 4 to 28. The higher the scores, the more the woman has a higher level of social support. The Cronbach’s alpha coefficient was .92.
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
Thananowan and Vongsirimas
7
Self-esteem. The Rosenberg’s Self-Esteem Scale (RSE) is a self-report instrument that assesses one’s own evaluation using 10 different items: a feeling of self-worth and self-respect (8 items) and a feeling of competence and ability (2 items; Rosenberg, 1989). Participants were asked to indicate the strength of their agreement in the last month on a 4-point Likert-type scale ranging from 1 (absolutely disagree) to 4 (absolutely agree). A total score is obtained by summing across each item score ranging from 10 to 40. The higher the score, the higher the woman’s self-assessed self-esteem. The Cronbach’s alpha coefficient for this study was .94. Depressive symptoms. The Center for Epidemiologic Studies Depression Scale (CES-D; Radloff, 1977) is a 20-item self-report instrument examining the frequency and duration of depressive symptoms. Participants were asked to rate their depressive symptoms over the past week on a 4-point Likert-type scale ranging from 0 (rarely or none of the time) to 3 (almost or all of the time). A total score is obtained by summing across each item score ranging from 0 to 60. Higher scores indicate greater depressive symptoms. In this study, a score of 16 was used as the cutoff for depressive symptoms. The Cronbach’s alpha coefficient for this study was .96. All instruments used in this study (except for the ST) were translated from English to Thai using a forward–backward translation method. The instruments were tested for content validity using sample group of Thai women with gynecological problems to ensure that they were culturally appropriate.
Statistical Analyses Sample size calculation was based on power analysis for structural equation modeling (SEM). The sample size tables yielded a power of at least 80%, with 80 degrees of freedom (df), the root mean square error of approximation (RMSEA) is .02 with .05 level of significance (Hancock & Freeman, 2001). The required sample size is at least 250 participants each for nonabused and abused groups. With 10% attrition rate, 550 potential participants were expected. Prior to analysis, all variables in the model satisfied the assumptions of normality and multicollinearity. That is, the linearity was assessed through the inspection of scatterplots among prior measured variables by SPSS program. Then the multicollinearity could be checked by inspecting the determinant of the covariance matrix. An extremely small determinant indicated a problem with multicollinearity (Tabachnick & Fidell, 2001). Descriptive analysis was conducted with SPSS version 13 to determine the prevalence of IPV, demographic characteristics, and sexual risk behaviors as well as gynecological problems. Pearson’s correlation analyses were
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
8
Journal of Interpersonal Violence
performed to investigate the relationship between the severity of IPV (i.e., ISA) and the proposed mediators (i.e., stress, social support, self-esteem, and depressive symptoms). According to the hypothesis, the hypothesized model was tested using SEM, a multivariate analysis technique, to determine how well the model can be explained by the empirical data. By using SEM, the hypothesized model was tested to estimate the relationships between observed and latent variables (the measurement model) and among latent variables themselves (the construct model), as well as direct and indirect effects. This process involved an evaluation of the hypothesis that the indicated measured items or scales reflected the latent constructs including stress, social support, self-esteem, depressive symptoms, IPV, and cervical cancer. Models for each construct were defined by permitting each of the relevant test or scales to load on a single factor representing the latent construct that it was hypothesized to measure. Maximum likelihood using LISREL version 8.72 was used for data analysis.
Results Overall, the 562 Thai women were recruited to the present study. Thirty questionnaires (5.6%) were excluded due to missing data, leaving a final sample size of 532 for data analysis. The prevalence of IPV, demographic characteristics, and sexual risk behaviors are displayed in Table 1. In summary, about 21.1% of participants (n = 112) had ever experienced any type of IPV in the past year. From the AAS, the prevalence of physical, sexual, and emotional abuse in the past year was 17.3% (n = 92), 11.5% (n = 61), and 13.2% (n = 70), respectively. The prevalence of gynecological problems was sorted from high to low as follows: benign tumors (37.6%), cervical cancer (22.2%), ectopic pregnancy (10.2%), ovarian cancer (9.8%), STIs (7.5%), endometrial cancer (5.5%), vulva cancer (3.6%), pelvic organ prolapses (2.8%), and menstrual disorders (0.8%). A correlation matrix of the major study variables is displayed in Table 2. The severity of IPV was significantly positively correlated with stress, depressive symptoms, and cervical cancer, but negatively correlated with social support and self-esteem. Stress was significantly positively correlated with depressive symptoms (r = .871, p < .01) and cervical cancer (r = .243, p < .01), but negatively correlated with social support (r = −.335, p < .01) and self-esteem (r = −.637, p < .01). In addition, depressive symptoms was highly correlated with self-esteem (r = −.693, p < .01). Several criteria were used to evaluate fit of the proposed model including the χ2 test, the comparative fit index (CFI), and the RMSEA. The generally agreed upon critical value for assessing model fit are CFI > 0.95, RMSEA < 0.06,
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
9
Thananowan and Vongsirimas Table 1. The Prevalence of IPV, Demographic Characteristics, Sexual Risk Behaviors. Characteristics Experienced IPV in the past year Physical abuse Sexual abuse Emotional abuse Women’s age (M, SD) Partners’ age (M, SD) Marital status Married Divorced/separated Widowed Cohabiting Number of marriages (M, SD) Once remarried Length of marriage (M, SD) Education ≤High school >High school Employment Unemployed Employed Housewife Household income (Thai baht) 30,001 Socioeconomic status Sufficient Insufficient In debt Family structure Living with husband and children Living with family members Living with others/friends Living alone
n
%
112 92 61 70 42.46 (12.174) 44.50 (12.937)
21.1 17.3 11.5 13.2
380 91 43 18 1.51 (1.142) 351 181 15.24 (11.021)
71.4 17.1 8.1 3.4 66 34
319 213
60 40
76 393 63
14.4 73.8 11.8
262 132 59 79
49.3 24.7 11.1 14.9
289 122 121
54.4 22.9 22.7
416 99 10 7
78.2 18.6 1.9 1.3 (continued)
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
10
Journal of Interpersonal Violence
Table 1. (continued) Characteristics Family relationship Good relationship Poor communication Some arguments Marital conflict Drink alcohol Smoking Substance use Age at first sex (M, SD) Number of sexual partners (M, SD) Forced sexual intercourse Having anal sex intercourse Having oral sex intercourse Having sex during menstruation
n
%
115 9 339 78 119 50 12 21.28 (5.309) 1.53 (1.212)
21.6 1.7 62 14.7 22.4 9.4 2.3
10 15 80
1.9 2.8 15
Note. IPV = intimate partner violence.
Table 2. Correlation Matrix of the Major Study Variables (n = 532).
1. ISA 2. ST 3. MSPSS 4. RSE 5. CES-D 6. Cervical cancer
1
2
1.000 .410** −.234** −.269** .351** .153**
1.000 −.335** −.637** .871** .243**
3
1.000 .567** −.341** −.073
4
1.000 −.693** −.184**
5
6
1.000 .257**
1.000
Note. ISA = Index of Spouse Abuse; ST = Stress Test; MSPSS = Multidimensional Scale of Perceived Social Support; RSE = Rosenberg’s Self-Esteem Scale; CES-D = Center for Epidemiologic Studies Depression Scale. *p < .05. **p < .01.
and p > .05 (Hu & Bentler, 1999). When the hypothesized model was examined by SEM, fit indices showed that the model fitted very well to the empirical data (χ2 = 140.77, df = 119, p = .084, CFI = 1.00, RMSEA = .019, goodness of fit index = .98, adjusted goodness of fit index = .95). The parameter estimates of the model are shown in Table 3 and Figure 1. Overall, this model accounted for 9% of the variance in cervical cancer. This is possible because
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
11
IE
TE
DE
IE
TE
R2 = .12
R2 = .45
−.26** — −.26** −.09 −.17* .25** .64* — .64*
DE
Self-Esteem (RSE) IE
TE
DE
IE
TE
Depressive Symptoms (CES-D) DE
IE
Cervical Cancer TE
.27** .16** .42** −.07* .41** .34** — .12** .12** .05 −.43* −.38* — −.41** −.41** .07 −.13 −.06 −.67** — −.67** −.10* −.61** −.71** −.03 −.20** −.23* .91** — 91** .30* — .30* R2 = .57 R2 = .91 R2 = .09
DE
Stress (ST)
Note. DE = standardized direct effect; IE = standardized indirect effect; TE = standardized total effect; MSPSS = Multidimensional Scale of Perceived Social Support; RSE = Rosenberg’s Self-Esteem Scale; ST = Stress Test; CES-D = Center for Epidemiologic Studies Depression Scale. *p < .05. **p < .01.
IPV MSPSS RSE ST Structural equation fit
Causal Variables
Social Support (MSPSS)
Effected Variables
Table 3. Standardized DE, IE, TE of Latent Variables in the Model.
12
Journal of Interpersonal Violence Social Support -0.26**
-0.09
IPV
0.27**
-0.07*
0.07
0.64*
Self-Esteem -0.67**
Stress
0.05
-0.03
Cervical Cancer
0.30* -0.10*
0.91**
Depressive Symptoms
Figure 1. Proposed mediators of the relationship between IPV and cervical cancer. Note. IPV = intimate partner violence.
the relationship between IPV (i.e., ISA) and cervical cancer is very low (r = .153, p < .01; Table 2). As shown in Table 3, results indicated that not only did IPV exhibit direct effects on social support (β = −.26; p < .01), stress (β = .27; p < .01), depressive symptoms (β = −.07; p < .05) and indirect effects on self-esteem (β = −.17; p < .05), but it also had a significant, positive, total effect on cervical cancer (β = .12; p < .01). Importantly, the indirect effect of IPV through the mediating variables (i.e., social support, selfesteem, stress, and depressive symptoms) on cervical cancer was statistically significant (β = .12; p < .01).
Discussion The prevalence of IPV in the past year among Thai women with gynecological problems (21.1%) was similar to the U.S. study (25%; Tjaden & Thoennes, 2000). However, the prevalence of physical abuse (17.3%) and sexual abuse (11.5%) in this study was higher than a population-based study that reported the prevalence of physical and sexual abuse to be about 5.7% and 3.7%, respectively (Nur, 2012). A possible explanation may involve sampling bias because abused women may come to the hospital with gynecological problems, resulting from physical or sexual abuse, and thus the prevalence of IPV in women in clinical settings would be expected to be higher than in women
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
Thananowan and Vongsirimas
13
in population settings. In addition, the use of a self-administered anonymous approach allowed women to feel more comfortable in revealing sensitive information about their abusive experiences. The prevalence of physical abuse in this study was also higher than the prevalence of sexual or emotional abuse. This was because physical abuse can result in injuries, while sexual or emotional abuse may lead to other health problems that are not directly linked to acts of violence. The findings in Table 2 highlighted the complex interrelationships among IPV, psychosocial factors, and cervical cancer. Therefore, women who experience IPV tend to have high stress, which in turn increases the likelihood of depressive symptoms but decreases social support and selfesteem. As a result, abused women may have limited access to health care resources and/or to undergo cervical cancer screening (Johnson et al., 2007). However, further investigation is needed to determine the underlying mechanisms that explain the relationship between IPV and inadequate cervical cancer screening. The SEM results generally provide a set of relationships in which the causal connections between several variables are examined simultaneously. Such results are important to understand the complex interrelationship among IPV, psychosocial factors, and cervical cancer. In addition, the SEM results provide the direction and strength of cause-and-effect relationships hypothesized by the researchers and also calculate the direct and indirect effects that independent variables have on the dependent variables. Although this model explained a small proportion of the variance (9%), IPV exhibited the significant indirect effect on cervical cancer through social support, selfesteem, stress, and depressive symptoms. Thus, our hypothesis was supported because the SEM results showed the mediation effects of the mediators on the outcome variables (i.e., cervical cancer). Although the direct effect between IPV and cervical cancer did not exist, the indirect effect and total effects reveals significant. It is probable that an increase in IPV could aggravate stress and depressive symptoms, which in turn could lead to worsening of cervical cancer, whereas social support and self-esteem can buffer the negative effects of victimization. Therefore, it is important for health care providers not to overlook IPV among inpatient Thai women with stress and depressive symptoms. Health care protocols for abused Thai women should include screening for and treatment of IPV-related psychosocial factors to prevent the risk of cervical cancer. In addition, screening women in health care settings for a history of IPV and psychosocial factors can link women to resources and treatment for assault-related health symptoms. And final, interventions designed to provide social support and protect
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
14
Journal of Interpersonal Violence
self-esteem should reduce stress and depressive symptoms among abused Thai women, thereby reducing the risk of cervical cancer. This study had some limitations that need to be acknowledged. First, the use of convenience samples in clinical settings may limit the generalizability of the findings to other populations. For future research, the model should be tested with more diverse and more representative samples. Second, because of the cross-sectional design, no inferences could be made with respect to causality. Therefore, longitudinal studies are essential for verifying the causal link between IPV and cervical cancer. Third, potential confounders in Table 1 (e.g., demographic characteristics and sexual risk behaviors) were not controlled and as a result they may impact women’s psychosocial factors and/or cervical cancer. Finally, all data were collected using self-report measures that may reflect biased and/or limited recall. Future research might consider the use of multiple data collection methods, including salivary cortisol when appropriate.
Conclusion The present study served as a starting point for future research of mediating effects in Thailand. Overall, the findings revealed the significant mediating effects of psychosocial factors on the relationship between IPV and cervical cancer, thus affirming the complex interrelationships among IPV, psychosocial factors, and cervical cancer. Because women in their reproductive years make many visits to health care providers, an important opportunity to screen for and intervene in either IPV or stress is available. The findings were also clinically important for health care providers when designing psychosocial interventions (e.g., stress reduction program with problem-focused coping) for abused Thai women with cervical cancer. Acknowledgment The authors would like to thank Thai women who participated in this study and Dr. Det Kedcham for his assistance on data analysis.
Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was granted by China Medical Board (CMB), Faculty of Nursing, Mahidol University, Bangkok, Thailand.
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
Thananowan and Vongsirimas
15
References Afifi, T. O., MacMillan, H., Cox, B. J., Asmundson, G., Stein, M. B., & Sareen, J. (2009). Mental health correlates of intimate partner violence in marital relationships in a nationally representative sample of males and females. Journal of Interpersonal Violence, 24, 1398-1417. Arco, D. A., Amo, J., Garcia-Pina, R., Garcia-Fulgueiras, A. M., Rodrigrez-Arenas, M. A., Ibaňez-Rojo, V., & Llacer, A. (2013). Violence in adulthood and mental health: Gender and immigrant status. Journal of Interpersonal Violence, 28, 2203-2222. Bailey, B. A. (2010). Partner violence during pregnancy: Prevalence, effects, screening, and management. International Journal of Women’s Health, 2, 183-197. Beydoun, H. A., & Saftlas, A. F. (2008). Physical and mental health outcomes of prenatal maternal stress in human and animal studies: A review of recent evidence. Paediatric and Perinatal Epidemiology, 22, 438-466. Campbell, R., Lichty, L. F., Sturza, M., & Raja, S. (2006). Gynecologic health impact of sexual assault. Research in Nursing & Health, 29, 399-413. Cesario, S. K., McFarlane, J., Nava, A., Gilroy, H., & Maddoux, J. (2014). Linking cancer and intimate partner violence: The importance of screening women in the oncology setting. Clinical Journal of Oncology Nursing, 18, 65-73. Chang, J. C., Cluss, P. A., Ranieri, L., Hawker, L., Buranosky, R., Dado, D., . . .Scholle, S. H. (2005). Health care interventions for intimate partner violence: What women want. Women’s Health Issues, 15, 21-30. Coker, A. L., Hopenhayn, C., DeSimone, C. P., Bush, H. M., & Crofford, L. (2009). Violence against women raises risk of cervical cancer. Journal of Women’s Health, 18, 1179-1185. Coker, A. L., Sanderson, M., Fadden, M. K., & Pirisi, L. (2000). Intimate partner violence and cervical neoplasia. Journal of Women’s Health & Gender-Based Medicine, 9, 1015-1023. Dienemann, J., Boyle, E., Baker, D., Resnick, W., Wiederhorn, N., & Campbell, J. (2000). Intimate partner abuse among women diagnosed with depression. Issues in Mental Health Nursing, 21, 499-513. Fanslow, J. L., & Robinson, E. M. (2010). Help-seeking behaviors and reasons for help seeking reported by a representative sample of women victims of intimate partner violence in New Zealand. Journal of Interpersonal Violence, 25, 929-951. Golding, J. M. (1999). Intimate partner violence as a risk factor for mental disorders: A meta-analysis. Journal of Family Violence, 14, 99-132. Hall, L. A., Kotch, J. B., Browne, D., & Rayens, M. K. (1996). Self-esteem as a mediator of the effects of stressors and social resources on depressive symptoms in postpartum mothers. Nursing Research, 45, 231-238. Hancock, G. R., & Freeman, M. J. (2001). Power and sample size for the Root Mean Square Error of Approximation Test of not close fit in Structural Equation Modeling (SEM). Educational and Psychological Measurement, 61, 741-758.
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
16
Journal of Interpersonal Violence
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternative. Structure Equation Modeling: A Multidisciplinary Journal, 6, 1-55. Hudson, W. W., & McIntosh, S. R. (1981). The assessment of spouse abuse: Two quantifiable dimensions. Journal of Marriage and Family, 43, 873-888. Humphreys, J. (2011). Sexually transmitted infections, pregnancy, and intimate partner violence. Health Care for Women International, 32, 23-38. Johnson, J. K., John, R., Humera, A., Kukreja, S., Found, M., & Lindow, S. W. (2007). The prevalence of emotional abuse in gynecology patients and its association with gynecological symptoms. European Journal of Obstetrics & Gynecology and Reproductive Biology, 133, 95-99. Lacey, K. K., McPherson, M. D., Samuel, P. S., Sears, K. P., & Head, D. (2013). The impact of different types of intimate partner violence on the mental and physical health of women in different ethnic groups. Journal of Interpersonal Violence, 28, 359-385. McFarlane, J. M., Parker, B., Soeken, K., & Bullock, L. (1992). Assessing for abuse during pregnancy: Severity and frequency of injuries and associated entry into prenatal care. Journal of the American Medical Association, 267, 3176-3178. Ministry of Public Health. (2000). The Stress Test. In Handbook of stress reduction. Bangkok, Thailand: Ministry of Public Health. Modesitt, S. C., Gambrell, A. C., Cottrill, H. M., Hays, L. R., Walker, R., Shelton, B. J., & Ferguson, J. E. (2006). Adverse impact of a history of violence for women with breast, cervical, endometrial, or ovarian cancer. Obstetrics & Gynecology, 107, 1330-1336. Moracco, K. E., Runyan, C. W., Bowling, J. M., & Earp, J. A. L. (2007). Women’s experiences with violence: A national study. Women’s Health Issues, 17, 3-12. Nur, N. (2012). The effect of intimate partner violence on mental health status among women of reproductive ages: A population-based study in a middle Anatolian city. Journal of Interpersonal Violence, 27, 3236-3251. Pereira, D. B., Antoni, M. H., Danielson, A., Simon, T., Efantis-Potter, J., Carver, C. S., & O’Sullivan, M. J. (2003). Life stress and cervical squamous intraepithelial lesions in women with human papillomavirus and human immunodeficiency virus. Psychosomatic Medicine, 65, 427-434. Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychology Measurement, 1, 385-401. Rosenberg, M. (1989). Society and the adolescent self-image. Middletown, CT: Wesleyan University Press. Ruback, R. B., Clark, V. A., & Warner, C. (2014). Why are crime victims at risk of being victimized again? Substance use, depression, and offending as mediators of the victimization–revictimization link. Journal of Interpersonal Violence, 29, 157-185. Sareen, J., Pagura, J., & Grant, B. (2009). Is intimate partner violence associated with HIV infection among women in the United States? General Hospital Psychiatry, 31, 274-278.
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
Thananowan and Vongsirimas
17
Shen, A. C. (2009). Self-esteem of young adults experiencing interparental violence and child physical maltreatment: Parental and peer relationship as mediators. Journal of Interpersonal Violence, 24, 770-794. Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistic (4th ed.). Boston, MA: Allyn & Bacon. Thananowan, N., & Vongsirimas, N. (2014). Association between intimate partner violence and women’s mental health: Survey evidence from Thailand. Pacific Rim International Journal of Nursing Research, 18, 3-15. Tiwari, A., Fong, D., Chan, C., & Ho, P. (2013). Factors mediating the relationship between intimate partner violence and chronic pain in Chinese women. Journal of Interpersonal Violence, 28, 1067-1087. Tjaden, P., & Thoennes, N. (2000). Extent, nature, and consequences of intimate partner violence: Findings from the National Violence Against Women Survey. Washington, DC: U.S. Department of Justice. Weitlauf, J. C., Jones, S., Xu, X., Finney, J. W., Moos, R. H., Sawya, G. F., & Frayne, S. M. (2013). Receipt of cervical cancer screening in female veterans: Impact of posttraumatic stress disorder and depression. Women’s Health Issues, 23, e153-e159. World Health Organization. (2001). Putting women first: Ethical and safety recommendations for research on domestic violence against women. Geneva, Switzerland: Author. World Health Organization. (2007). Ethical and safety recommendations for researching, documenting, and monitoring sexual violence in emergencies. Geneva, Switzerland: Author. Wu, V., Huff, H., & Bhandari, M. (2010). Pattern of physical injury associated with intimate partner violence in women presenting to the emergency department: A systematic review and meta-analysis. Trauma, Violence, & Abuse, 11, 71-82. Zimet, G. D., Dahlem, N. W., Zimet, S. G., & Farley, G. K. (1988). The Multidimensional Scale of perceived social support. Journal of Personal Assessment, 52, 30-41.
Author Biographies Nanthana Thananowan, RN, PhD, is an associate professor in the Department of Obstetrics and Gynecological Nursing at Faculty of Nursing, Mahidol University, Bangkok, Thailand. Her research interest includes intimate partner violence and women’s health. Nopporn Vongsirimas, RN, PhD, is an assistant professor in the Department of Mental Health and Psychiatric Nursing at Faculty of Nursing, Mahidol University, Bangkok, Thailand. Her research focuses on depressive symptoms among adolescents and women exposed to violence.
Downloaded from jiv.sagepub.com at NORTH CAROLINA STATE UNIV on April 21, 2015
