Cover Image
close this bookViolence against Women (WB)
View the document(introduction...)
View the documentInformation
View the documentForeword
View the documentAcknowledgments
View the documentAbstract
View the document1. introduction
View the document2. The scope and evolution of the problem
View the document3. A primer on violence against women
View the document4. health consequences of gender-based violence
View the document5. Implications of gender violence for health and development
View the document6. Steps toward eliminating violence against women
View the document7. Research needs
View the document8. Conclusions
Open this folder and view contentsAppendix
View the documentNotes
View the documentBibliography

4. health consequences of gender-based violence

A growing body of research has emerged in recent years on the mental and physical health consequences of violence against women and the burden it places on health care Although much of this research is from the industrial world, clinicians and advocates in developing countries confirm that the U.S.-based literature corresponds well to their experience with battered women and survivors of sexual assault and abuse.

When this literature is considered together with estimates of the prevalence of gender-based violence in different parts of the world, the magnitude of the violence related health burden begins to become clear. Perhaps the best estimate of this burden comes from a modeling exercise undertaken by the World Bank to inform its policy annual, the World Development Report:, which in 1993 focused on health. For this effort, Bank staff and outside experts estimated the healthy years of life lost to men and women due to different causes. The exercise counted every year lost due to premature death as one disability-adjusted life year (DALY) and every year spent sick or incapacitated as a fraction of a DALY, with the value depending on the severity of the disability.

From this analysis, broken down by region and age group, rape and domestic violence emerge as a significant cause of disability and death among women of reproductive age in both the industrial and the developing world (sees appendix C). In established market economies gender based victimization accounts for nearly one in five healthy years of life lost to women age 15 to 44. On a per capita basis the health burden of rape and domestic violence affecting reproductive-age women is roughly the same in industrial and developing countries, but because the overall burden of disease is much greater in developing countries, a smaller percentage is attributable to gender based victimization.

The World Bank estimates that rape and domestic vim fence account for 5 percent of the healthy years of life lost to women of reproductive age in demographically developing countries. In developing countries such as China, where maternal mortality and poverty-related diseases are relatively under control, the healthy years of life lost due to rape and domestic violence again account for a larger share-16 percent of the total burden. At a global level the health burden from gender-based victimization among women age 15 to 44 is comparable to that posed by other risk factors and diseases already high on the world agenda, including the human immunodeficiency virus (HIV), tuberculosis, sepsis during childbirth, cancer, and cardiovascular disease (table 5).

The following section explores the long- and short-tern health effects of different types of abuse. Box 3 presents the range of health effects recorded in the literature.

Table 5 Estimated global health burden of selected conditions for women age 15 to 44


Disability-adjusted life years lost( millions)

Maternal conditions




Obstructed labor


STDs [excluding HIV)

15. 8

Pelvic inflammatory disease






Cardiovascular disease


Rape and domestic violence


All cancers






Motor vehicle accidents






· Rape and domestic violence are included here for illustrative purposes. They are risk factors for disease conditions, such as STDs, depression, and injuries, not diseases in themselves of

Source: World Bank 1993

Health effects of abuse by intimate partners

Abuse of women by intimate male partners has both physical and mental health consequences. The physical consequences include injury and a host of less-defined somatic complaints. In the United States 22 to 35 percent of women presenting to urban emergency rooms exhibit symptoms related to ongoing abuse (Council on Scientific Affairs, American Medical Association 1992). A representative survey of women in Texas found that 24 percent of women who had ever been abused required medical treatment because of the abuse (Teske and Parker 1983).

Abuse-related injuries include bruises, cuts, black eyes, concussions, and broken bones. Abuse also leads to miscarriages and to permanent injuries, such as damage to joints, partial loss of hearing or vision, and scars from burns, bites, and knife wounds. In Bangladesh and parts of Latin America acid throwing, a form of abuse increasingly perpetrated by vengeful lovers, leads to permanent disfigurement. In addition to injuries, battered women often suffer chronic headaches, abdominal pains, muscle aches, recurrent vaginal infections, and sleep and eating disorders. Two studies, one covering 390 randomly selected women in the United States and the other 2, 000 randomly selected women in New Zealand, found that abused women had significantly worse physical and mental health than non abused women (Koss, Ross, and Woodruff 1991; Mullen and others 1988). Recent research reported in the Journal of the American Medical Association suggests that abuse can also be associated with delayed physical effects, particularly arthritis, hypertension, and heart disease (Council on Scientific Affairs 1992).

Quantitative data on health consequences in developing countries are less available, but abundant evidence from crisis centers, police reports, and ethnographic research shows that in these countries, too, violence is a significant cause of injury and ill health. The case study presented by China to the United Nations Expert Group Meeting on Violence in the Family reports that domestic violence caused 6 percent of d Baths and serious injuries in Shanghai in 1984 (Wu 1986). A three-month surveillance survey in Alexandria, Egypt, indicated that domestic violence was the leading cause of injury to women, accounting for 27.9 percent of all visits by women to area trauma units (Graitcer, personal communication, 1994). And 18 percent of married women surveyed in urban areas of Papua New Guinea (PNG) had received hospital treatment for injuries inflicted by their husbands. As Christine Bradley of the country's law reform committee observes: "In PNG, where many women have enlarged spleens due to malaria, a single blow can kill them. (1988, p. 3).

For many women, however, the psychological effects of abuse are more debilitating than the physical effects. Fear, anxiety, fatigue, post-traumatic stress disorder (PTSD), and sleeping and eating disturbances are common long-term reactions to violence. Abused women may become dependent and suggestible, and they may find it difficult to make decisions done. Compounding the psychological consequences that women suffer from abuse is their relationship to the abuser. The legal, financial and emotional ties that the victims of marital violence often have to the perpetrator enhance their feelings of vulnerability, loss, betrayal and hopelessness. Abused women frequently become isolated and withdrawn as they try to hide the evidence of their abuse.

Box 3 Health consequences of genderbased violence

Physical health consequences

Pelvic inflammatory disease
Unwanted pregnancy
Chronic pelvic pain
Gynecological problems
Alcohol/drug abuse
Irritable bowel syndrome
Injurious health behaviors
(smoking, unprotected sex)
Partial or permanent disability
Mental health consequences

Post-traumatic stress disorder Depression Anxiety Sexual dysfunction Eating disorders Multiple personality disorder Obsessive compulsive disorder


Suicide Homicide

Not surprisingly, these effects make wife abuse a primary context for many other health problems. Battered women in the United States are four to five times more likely than nonbattered women to require psychiatric treatment and five times more likely to attempt suicide (Stark and Flitcraft 1991). About a third of battered women suffer major depressions, and some go on to abuse alcohol or drugs. Miller(1990) reports that spousal violence is the strongest predictor of alcoholism in women, even after controlling for income, violence in the family of origin, and having an alcoholic husband. Moreover, studies show that most battered women who drink begin drinking excessively only after the onset of abuse (Amaro and others 1990; Stark and others 1981).

The relation between battering and psychological dysfunction also has important implications for women's mortality, because of increased risk of suicide. After reviewing evidence from the United States, Stark and Flitcraft concluded that "abuse may be the single most important precipitant for female suicide attempts yet identified (1991, p. 141). One-fourth of suicide attempts by American women-and half of all attempts by African American women-are preceded by abuse (Stark 1984).

A cross cultural survey of suicide by Counts draws the same conclusion, positing that in some African, Oceanic, and South American societies, female suicide operates as a culturally recognized behavior that enables the "politically revenge themselves on those who have made their lives intolerable. (1987, p. 195). Counts finds support for her argument in cultures from Africa, Peru, Papua New Guinea, and the Melanesian islands. Among Fijian Indian families in which someone has committed suicide, 41 percent cite marital violence ax the cause (Haynes 1984).

Suicide is not an inconsequential form of death. The World Bank estimates that of the healthy years of life lost to women in rural Chins, 30 percent are lost due to suicide (Bobadilla, personal communication, 1993). This finding is consistent with reports of mass suicides in rural China among women forced or sold into unwanted (and often violent) marriages (name withheld 1991). In Sri Lanka, a country with reasonably accurate mortality statistics, the rate of death due to suicide among young women age 15 to 24 is five times that due to infectious diseases and 55 times the rate due to obstetric-related causes (WHO 1985).

Three studies from India almost suggest a similar link between marital violence and female suicide. A one-year study of completed suicides in Delhi revealed that 46 percent were committed by males and 54 percent by females. It cited marital discord and ill treatment by husbands and in-laws as the most common precipitating factor among women. Another study analyzed all cases of suicide in 1978 known to the Madras Police Department. The peak ages for women committing suicide were 15 to 20, Among the two-thirds of the women who were married, the principal cause cited for suicide was "maladjustment with an alcoholic or drug-addict husband. "The third study, on suicide deaths in Daspur, found that the peak a" for women were 15 to 24 and the most common precipitating factor was "quarrel with spouse" (as cited in Paltiel 1987).

The relation between domestic violence and homicide may be even more profound. Data from a wide range of countries demonstrate that domestic violence is a mayor risk factor for murder of and by women. A recent review of spousal homicide in the United States, published in the American Journal of Public Health, reports that "studies of homicides between intimates show that they are often preceded by a history of physical abuse directed at the women and several studies have documented that a high proportion of women imprisoned for killing a husband had been physically abused by their: spouses" (Mercy and Saltzman 1989, p. 597). In Canada 62 percent of women murdered in 1987 died at the hands of an intimate male partner (Canadian Centre for Justice Statistics 1988). In the first 11 months of 1992, 415 women were murdered in the Brazilian state of Pernambuco, 70 percent by a male intimate (Dimenstein 1992). Of the 100 murders in Israel (not including the territories) in 1991, 42 involved women killed by a husband or lover (Nevo 1993). And in Papua New Guinea almost 73 percent of adult women murdered between 1979 and 1982 were killed by their husbands (Bradley 1988). Studies from a variety of culture-including (Canada, Papua New Guinea, and the United States-confirm that when women kill men, it is often in self-defense and usually after years of persistent and escalating abuse (Browne 1987; Walker 1989; Canadian Centre for Justice Statistics 1988; Bradley 1988; Kellerman and Mercy 1992).

The link between intimate violence and homicide is particularly evident in India, where women's deaths due to burns have been increasing since 1979, a development that can be tied to the commercialization of dowry demands (Pawar 1990). A young bride may be subject to severe abuse from her husband and in-laws if their continuing demands for money or goods from her family are not met. A frequent subterfuge is to set the woman on fire with kerosene and then claim that she died in a kitchen accident hence the term bride burning. In 1990 the police officially recorded 4,835 dowry deaths in India, but government sources readily acknowledge that this is a gross underestimate (Kelkar 1992). In both urban Maharashtra and greater Bombay, one of every five deaths among women age 15 to 44 is due to "accidental burns" For the younger age group 15 to 24, the proportion is one of four Karkal 1985).

Health effects of rape and sexual assault

Sexual assaults can cause both physical injury and profound emotional trauma. A study of rape in urban and rural areas of Bangladesh reports that 84 percent of victims suffered severe injuries or unconsciousness, mental illness, or death following the rape (Shamim 1985). Rape survivors exhibit a variety of trauma-induced symptoms-nightmares, depression, inability to concentrate, sleep and eating disturbances, and feelings of anger, humiliation, and selfblame. In addition, 50 to 60 percent of victims experience severe sexual problems, including fear of sex, problems with arousal, and decreased sexual functioning (Burnam and others 1988; Becker and others 1986; Becker and others 1982).

The malignant effects of rape are not surprising given the physical, psychological, and moral violation of the person that it represents (Breslau and others 1991; Herman 1992). A study from the United States found that rape victims were nine times more likely than nonvictims to have attempted suicide, and twice as likely to experience a major depression (Kilpatrick 1990). Follow-up studies have shown that rape survivors have higher rates of persistent post-traumatic stress disorder (PTSD) than victims of other traumas (Norris 1992). Some experts consider female victims of sexual abuse and assault to be the largest single group of PI SD sufferers, and rape the single most likely event to cause PTSD (Foe, Olasov, and Steketee 1987).

Studies that follow victims over time show that the traumatic consequences of rape can persist for many years. A study to validate the Rape Aftermath Symptom Test (RAST) demonstrated that the instrument could distinguish the symptoms of rape victims from those of nonvictims at intervals up to three years after a rape (Kilpatrick 1988). According to studies in the United States, one in four women who have been raped still exhibits dysfunctional symptoms four to six years after the assault (Hanson 1990; Burgess and Holmstrom 1979). In another sample 60 percent of sexual assault victims reported sexual dysfunction three years after the assault (Becker and others 1986). Even after many years, women who have been sexually assaulted are significantly more likely to qualify for 10 different psychiatric diagnoses, including major depression, alcohol abuse, PTSD, drug abuse, obsessive compulsive disorder, generalized anxiety, eating disorders, multiple personality disorder, and borderline personality syndrome. The relative risk ratio for these diagnoses for survivors of rape and sexual assault is about two times greater risk (Koss 1990).

Beyond physical injury and emotional trauma, rape survivors face the risk of sexually transmitted diseases (STDs), including the acquired immunodeficiency syndrome (AIDS). A support center for rape victims in Bangkok, Thailand, reports that 10 percent of its clients contract a sexually transmitted disease as a result of the rape (Archavanitkui and Pramualratana 1990). In the United States almost a dozen women and twice as many children had contracted AIDS through rape and child sexual abuse by July 1992 (Dattel 1992).

The possibility of unwanted pregnancy is also substantial. Mexican rape crisis centers report that 15 to 18 percent of their clients become pregnant because of rape, a figure consistent with data from Korea and Thailand (COVAC 1990; CAMVAC 1985; Arehavanitlcui and Pramnalratana 1990; Shim 1992). Rape victims in Mexico have more options than victims in many other countries: a new law requires judges to rule on a rape survivor's request for an abortion within five working days. But in countries in which abortion is illegal even in eases of rape, or where safe abortion services are inaccessible or prohibitively expensive, thousands of women must suffer the double humiliation of being raped and then being compelled to bear the rapist's child.

The consequences of rape for victims in societies that place a high value on women's virginity are severe. Many African, Asian, and Middle Eastern cultures equate a young woman's worth with her virginity. As Fauveau and Blanchet describe in their ground-breaking study of injury in rural Bangladesh, "even when women are victims, a premarital sexual relation is said to spoil something intrinsic in their physical and moral person....Their ruined reputation cannot be mended (1989, p. 1125). The study cites numerous ease studies of women who were beaten, murdered, or driven to suicide because of the dishonor that rape or illegitimate pregnancy brought upon their families. Their study found that there were 130 percent more deaths from injury (suicide, homicide, assault, and complications from induced abortion) among unmarried than among married teenage girls; this reinforces their qualitative data sub deliberate violence toward girls who are raped or who become pregnant outside of marriage (Acasadi and Jonnson Acasadi 1990). Likewise, in a study of female homicides in Alexandria, Egypt, 47.1 percent of women who had been killed had been murdered by a relative after they had been raped (Graitcer and Youssef 1993).

Health effects of child and adolescent sexual abuse

Research in the United States has shown that about one fifth of child sexual abuse victims evidence serious long-term psychological effects (Browne and Finkelhor 1986). These may include dis-associative responses and other PTSD indicators, such as chronic arousal, nightmares, flashbacks, and emotional numbing. Burnam and others (1988), using multivariate techniques, demonstrated that women in the Los Angeles Epidemiological Catchment Area survey who were sexually abused as children were more than twice as likely as peers who were not abused (58.6 percent versus 24.0 percent) to have at least one psychiatric diagnosis in their lifetime. (The L.A. Catchment survey is an on going mental health research project sponsored by the National Institutes of Mental Health.) victims who were abused by fathers or stepfathers, whose assaults involved genital contact, and whose molestation involved force appear to be at high risk of long-lasting effects (Browne and Finkelhor 1986).

Fully distinguishing the physical and emotional effects of sexual abuse is difficult because the long-term psychological complications are often manifested as physical complaints. Limited research into the somatic consequences of child sexual abuse indicates that it is linked to chronic pelvic pain, headaches, asthma, and gynecological problems (Koss and Heslet l992). Recent research has also linked sexual abuse with such gastrointestinal disorders as irritable bowel syndrome and chronic abdominal pain. These disorders, which have no clearly established pathogenesis, occur more frequently in women than in men, and they pose a considerable health and economic burden. In the United States they are the most common chronic gastrointestinal diagnoses seen in primary care and gastroenterology practices (Drossman and others 1990).

Early sexual victimization may also leave women less skilled at protecting themselves, less sure of their worth and their personal boundaries, and more apt to accept victimization as a part of being female. These effects may increase the chances of future victimization (Koss 1990). Early traumatic sexual experiences have been linked to increased risk for rape among college women (Koss and Dinero 1989). Likewise, in Briere's community-based sample, 49 percent of childhood sexual abuse victims reported being battered in adult relationships, compared with 18 percent of the nonvictim group (Briere 1984, as cited in Browne and Finkelhor 1986). Russell (1986) found that 68 percent of incest victims reported being the victim of rape or attempted rape (excluding incestuous rape) later in their lives, compared with 17 percent of nonabused controls.

Recent studies also link early sexual victimization with high-risk behaviors in adolescence and adulthood, including excessive drug and alcohol use, unprotected sex with multiple partners, prostitution, and teen pregnancy (Zierler and others 1991; Finkelhor 1987; James and Meyerding 1977; Boyer and Fine 1992). A variety of studies link childhood abuse to the later development of alcoholism, especially in women (Dembo 1987; Jellinek, Murphy, and Poitrast 1992; Blane, Miller, and Leonard 1988). A particularly well-designed multiple-regression analysis by Miller, Downs, and Testa (forthcoming) found that rates of childhood victimization were significantly greater for women in alcoholism treatment programs than for women in treatment for other mental health problems, battered women without alcohol problems, or nonalcoholic women in the general population These findings remained significant even after controlling for demographic and family background differences, including parental alcohol problems.

Researchers Debra Boyer and David Fine (1992)

I likewise found a significant link between childhood sexual abuse and teenage pregnancy among a sample of 535 adolescent mothers in Washington State. Compared with teens who became pregnant but had not been abused, sexually abused adolescents began intercourse a year earlier, were more likely to have used drugs and alcohol, and were less likely to practice contraception. Abused adolescents were also more likely to have been battered by an intimate partner and to have exchanged sex for money, drugs, or a place to stay. The average age at first intercourse for abused women was 13.8 years, compared with the national average of 16.2. Only 28 percent of the abused teens used birth control at first intercourse, compared with 49 percent of their peers.

Health effects of genital mutilation

The medical complications of genital mutilation can be severe, especially for women who are infibulated. A study from Sierra Leone found that X3 percent of all women circumcised required medical attention at some time for problems related to the procedure (Hosken 1988). The immediate risks of clitoridectomy or infibulation are similar: hemorrhage of the clitoral artery, infection, urine retention, and tetanus or blood poisoning from unsterile and often primitive cutting implements (knife, razor blade, broken glass). And the pain of the operation, often carried out without anesthesia, can cause young girls to go into shock

Over the long term women who are infibulated generally suffer more severe physical health consequences than women who are excised. Infibulation, because it involves more extensive cutting and stitching, poses significantly higher risks of hemorrhage end infection. And the partial closing of the vaginal and urethral openings leads to more problems relating to retention of urine and menstrual blood, such as chronic urinary tract infections, stones in the urethra or bladder, constant back and menstrual pain, irregularity, and repeated reproductive tract infections. In some cases these infections can lead to sterility, a devastating consequence for women whose worth is defined largely in terms of their ability to bear children.

Infibulation destines a woman to a cycle of pain, cutting, and restitching to accommodate sexual intimacy and childbirth. Infibulated women often must be cut on their wedding night to make intercourse possible, and again for the birth of a child. Intercourse is frequently perceived as painful, a perception that likely has both physical and psychological roots. And at the time of birth, infibulation puts both mother and child at risk. Among 33 infibulated mothers followed et Somalia's Benadir Hospital all required extensive episiotomies during childbirth, their second-stage labor was five times longer than normal, five of their babies died, and 21 suffered oxygen deprivation because of the long and obstructed labor (Warsame 1988). Most women are reinfibulated after childbirth to reconstruct a small vaginal opening; over time this repeated cutting and stitching transforms the genital area into tough, unyielding scart issue.

Although excised women normally have fewer long-term complications than women who are infibulated, clitoridectomy is not without serious risks. A significant share of excised women face a lifetime of unending infections, pain, bleeding, and abscesses. They also face the possibility of severe psychological repercussions. Little research has been done on the psychological impact of genital mutilation, but clinicians report serious long-term distress and psychological dysfunction in some cases. Based on her experience in Sudan, Dr. Nahid Toubia describes a pattern of vague physical complaints, depression, and lethargy among circumcised women very similar to that common among sexually abused or raped women in the United States:

Thousands of women [in Sudan] come to the Ob/Gyn outpatient clinics with vague chronic symptoms which they metaphorically interpret as originating from the pelvis. These women are perceived by doctors and the hospital authorities as a great nuisance and a drain on the system since they have no medically detectable pathology. Sitting for hours listening to them, it soon becomes clear that the vague symptoms of general fatigue, loss of sleep, backache, headache, pelvic congestion, uttered in a depressed, monotonal voice, are a muted cry for help for a much more deeply felt paint With a little probing, the women talk about fear of sex, the threat of infertility after infection, and fears about the state of their genitals (they have no way of assessing whether they are normal). (Toubia 1993, p. 19)

In 1982 the World Health Organization (WHO) issued a statement warning that genital mutilation should never be carried out by professional health staff. Despite this statement and many similar resolutions crafted by various medical bodies, delegates to the UN Human Rights Seminar on Traditional Practices held in Burkina Faso in 1991 reported that, for reasons of financial gain, medical personnel are performing circumcisions in hospitals in place of the midwives and traditional practitioners who normally carry out the procedure (Dorkenoo and Scilla 19°2). Although "medicalizing" circumcision may reduce the immediate risks of infection, it does not end the abuse of women's human rights represented by this unnecessary, mutilating surgery. As Aziza Kamil, leader of the Cairo Family Planning Association's project on female genital mutilation, points out:

No action will entrench genital mutilation more than legitimating it through the medical profession. If doctors and hospitals start to perform it, rather than condemn it, we will have no hope of over eradicating the practice. All the respect and authority given to doctors will be transferred to the practice and [activists] will lose [their] credibility. (Dorkenoo and Scilla 1992)

Effects of violence against women on the health care system

Violence affects women's health-and the health of society at large-by diverting scarce resources to the treatment of a largely preventable social ill. Considering the prevalence of abuse and the nature of its health effects, it is reasonable to conclude that victimization represents a significant drain on available health resources. A study at a major U.S. health maintenance organizations (HMO) found that a history of rape or assault was a stronger predictor of physician visits and outpatient costs than any other variable, including a woman's age or such health risks as smoking (Koss, Kohl;, and Woodruff 1991). This multiple-regression analysis, which included five demographic variables, four measures of health status, and five measures of lifetime stress, found the following:

- Women who had been raped or beaten had medical costs in the index year that were two and a half times higher than those of women who were not victimized ($401 versus $161).

- According to a temporal analysis based on a subset of victims, the biggest increase in use of health care services occurred in the second year following victimization, but four years after the incident it still had not returned to baseline.

- Women who have been assaulted or raped describe themselves as less healthy, experience more symptoms across virtually all body systems (except skin and eyes), and report higher rates of behaviors injurious to the health (such as smoking and failure to use seat belts).

A similar study by Felitti (1991) found that among women enrolled in an HMO plan, 22 percent of those who had a history of childhood molestation or rape had visited a physician 10 or more times a year, compared with only 6 percent of nonvictimized women. And in a random, population based survey of medical care use links Angeles, respondents with a history of sexual assault were nearly twice as likely to have sought mental health care and a third more likely to have visited a physician within the past six months than men and women who had not been sexually assaulted (Golding and others 1988). This effect persisted even after gender, ethnicity, and age were controlled for. The study's multivariate analysis suggests that assault has au indirect influence, affecting health care use by increasing psychiatric morbidity and reducing functional ability. The study also demonstrates that the prevalence of sexual assault is significantly higher among those who use health services than among those who do not, underlining the importance of using health care facilities to identify victims of violence for referral for appropriate advocacy and support.