Health Risk Behaviors and Medical Sequelae of Childhood Sexual Abuse
FERN E. SPRINGS, M.D.,* Mayo Medical School; WILLIAM N. FRIEDRICH, Ph.D., Section of Psychology The relationship between childhood sexual abuse and subsequent health risk behaviors and medical problems was examined in 511 women who had used a family practice clinic in a rural midwestern community during a 2-year period (1988 and 1989). These women completed a questionnaire that assessed various health risk behaviors-smoking, drinking, drug abuse, number of sexual partners, and age at first intercourse-and a medical symptom checklist that assessed 38 medical problems related to major systems of body function, the somatization scale from the SCL-90, a screen for sexual abuse, and a brief measure of social support. The results indicated that sexually abused women, who represented 22.1 % of the sample, reported significantly more medical problems, greater levels of somatization, and more health risk behaviors than did the nonabused women. More severe abuse (for example, penetration or multiple abusers) correlated with more severe problems. Extent of social support correlated inversely with the number of gynecologic problems reported in the sexually abused group. Fewer than 2 % of the sexually abused women had discussed the abuse with a physician. To identify and assist victims of sexual abuse, physicians should become experienced with nonthreatening methods of eliciting such information when the medical history is obtained.
The lasting psychologic effects of childhood sexual abuse have been documented. 1·9 Victims of such abuse have experienced increased frequencies of depression, anxiety, sexual concerns, and dissociation. An effect reported more recently is somatization.l-v'? Several studies have shown an association between medical problems-most notably, pelvic pain, lI-l3 gastrointestinal disorders," and general medical problems's-i-and prior sexual abuse. These studies, however, were conducted in patients from psychiatric or specialty clinics. We are unaware of any previous study, such as our current investigation, that was conducted in a general medical population. Early sexual activity, multiple sexual partners (generally referred to as promiscuity), and smoking are factors known to be relevant to the development of cervical intraepithelial neoplasms.v-'? These three factors also have been reported *Current address: Gundersen Clinic, La Crosse, Wisconsin. This study was supportedin part by a grant from Mayo Foundation, Addressreprint requests to Dr. W. N. Friedrich,Section of Psycho1ogy, Mayo Clinic, Rochester, MN 55905. Mayo Clin Proc 67:527-532, 1992
to be related to sexual abuse in studies of adolescents.v" If a correlation could be found between childhood sexual abuse and health risk behaviors and medical problems, implications would exist for the prevention of medical problems in those identified as victims of sexual abuse and also for the identification of victims who may need but have not received psychologic help. The literature on the response to sexual victimization has identified social support as a possible buffering variable that aids in decreasing the overall level of symptoms experienced by victims." Although the relationship of social support to psychologic problems has been studied, this relationship has not been examined in regard to medical problems specifically associated with prior sexual abuse. Empiric evaluation of the coping resource of social support would be important relative to the victim's response to sexual abuse. This study was designed to examine, in a large sample of adult women, the relationship of various health risk behaviors and of reported medical symptoms to the experience of sexual abuse during childhood. In addition, social support was examined as a potential buffer to the manifestation of medical problems and somatization. These adult women
527
528
SEQUELAE OF CHILDHOOD SEXUAL ABUSE
were neither college students nor specifically gynecologic patients; therefore, this study corrects for several biases present in recent reports on the effect of sexual abuse in adult women.
STUDY SUBJECTS AND METHODS Invitations to participate in this study were mailed to a simple random sample of 1,443 women who were 18 to 50 years old and had used a rural outpatient family practice clinic for any reason during the 2-year period beginning Jan. 1, 1988. Of these invitations, 122 were returned with no forwarding address. Eighty-two women declined to participate by returning the stamped, addressed postcard and providing information on age and education; 689 neither responded nor provided information on age and education. The 550 women who agreed to participate were sent a ninepage questionnaire; 511 completed the questionnaire and returned it. This response represents 38.7% of those who received an invitation to participate. The questionnaire was designed to obtain information on various demographic factors, health risks, and medical symptoms. The sexual abuse history was elicited by asking a series of questions derived from apreviously used questionnaire. 3•20 As in the study by Briere and Runtz,' the participants were asked to identify up to two important sexual experiences they had had before 18 years of age with persons at least 5 years older-for example, strangers, friends, or family members. The level of coercion involved in the experience also was elicited. All women who reported such experiences were considered to have been sexually abused with the following exception. We excluded 13 women from the sexually abused group who were between 15 and 17 years old at the time of the reported incident because each related that the person involved was a friend, that no force was involved, and that an ongoing relationship had been established. Three women who reported forceful encounters with someone who was less than 5 years older than they were at the time of the incident were included in the abused group; the ages of these women at that time were 13, 16, and 17 years, respectively. Somatization was evaluated by using the somatization scale from the SCL-90. 21 The SCL-90 is a widely used screening measure, and the somatization scale consists of 12 items (for example, headache, numbness, weakness, and faintness) whose frequencies for the previous week were assessed by using a 5-point scale. The somatization scale has acceptable reliability and validity. Alcohol and drug abuses were assessed by posing the following three questions: Have you ever felt a need to cut down on the amount of alcohol you drink? Do you think you are a normal drinker? Have you ever had a problem with drugs or addictive medications?
Mayo Clin Proc, June 1992, Vol 67
Smoking as a health risk factor was examined in four ways: whether the participant had ever smoked, age at which smoking was begun, heaviest use (in packs/day), and duration of smoking (in years). Other health risk behaviors were evaluated by asking about age at first intercourse, number of sexual partners before 18 years of age, total number of sexual partners, pregnancy before 18 years of age,current height and weight, and frequency of Papanicolaou tests (Pap smears). Medical problems were assessed by using a short review of systems for various medical disorders and operations. The checklist included a total of 38 medical problems; the participant was asked to indicate all those she currently or previously had. Social support was assessed with four questions about number of close friends, satisfaction with the current level of social support, whether the abuse experience had been told to anyone, and whether emotional support was offered by those who were told about the abuse experience. The relationship of a history of sexual abuse, health risk behaviors, medical problems, and somatization was examined with the X2 technique or t test, as appropriate. Because of the exploratory nature of our study, P values of 0.05 or less were considered significant findings.
RESULTS Sample Description.-Demographic information on the entire study sample, the nonabused and abused subsamples, and two groups of nonresponders was determined for age, education, income, and marital status (Table 1). The sexually abused subsample was significantly older than the nonabused group (t = 2.6; P
FERN E. SPRINGS, M.D.,* Mayo Medical School; WILLIAM N. FRIEDRICH, Ph.D., Section of Psychology The relationship between childhood sexual abuse and subsequent health risk behaviors and medical problems was examined in 511 women who had used a family practice clinic in a rural midwestern community during a 2-year period (1988 and 1989). These women completed a questionnaire that assessed various health risk behaviors-smoking, drinking, drug abuse, number of sexual partners, and age at first intercourse-and a medical symptom checklist that assessed 38 medical problems related to major systems of body function, the somatization scale from the SCL-90, a screen for sexual abuse, and a brief measure of social support. The results indicated that sexually abused women, who represented 22.1 % of the sample, reported significantly more medical problems, greater levels of somatization, and more health risk behaviors than did the nonabused women. More severe abuse (for example, penetration or multiple abusers) correlated with more severe problems. Extent of social support correlated inversely with the number of gynecologic problems reported in the sexually abused group. Fewer than 2 % of the sexually abused women had discussed the abuse with a physician. To identify and assist victims of sexual abuse, physicians should become experienced with nonthreatening methods of eliciting such information when the medical history is obtained.
The lasting psychologic effects of childhood sexual abuse have been documented. 1·9 Victims of such abuse have experienced increased frequencies of depression, anxiety, sexual concerns, and dissociation. An effect reported more recently is somatization.l-v'? Several studies have shown an association between medical problems-most notably, pelvic pain, lI-l3 gastrointestinal disorders," and general medical problems's-i-and prior sexual abuse. These studies, however, were conducted in patients from psychiatric or specialty clinics. We are unaware of any previous study, such as our current investigation, that was conducted in a general medical population. Early sexual activity, multiple sexual partners (generally referred to as promiscuity), and smoking are factors known to be relevant to the development of cervical intraepithelial neoplasms.v-'? These three factors also have been reported *Current address: Gundersen Clinic, La Crosse, Wisconsin. This study was supportedin part by a grant from Mayo Foundation, Addressreprint requests to Dr. W. N. Friedrich,Section of Psycho1ogy, Mayo Clinic, Rochester, MN 55905. Mayo Clin Proc 67:527-532, 1992
to be related to sexual abuse in studies of adolescents.v" If a correlation could be found between childhood sexual abuse and health risk behaviors and medical problems, implications would exist for the prevention of medical problems in those identified as victims of sexual abuse and also for the identification of victims who may need but have not received psychologic help. The literature on the response to sexual victimization has identified social support as a possible buffering variable that aids in decreasing the overall level of symptoms experienced by victims." Although the relationship of social support to psychologic problems has been studied, this relationship has not been examined in regard to medical problems specifically associated with prior sexual abuse. Empiric evaluation of the coping resource of social support would be important relative to the victim's response to sexual abuse. This study was designed to examine, in a large sample of adult women, the relationship of various health risk behaviors and of reported medical symptoms to the experience of sexual abuse during childhood. In addition, social support was examined as a potential buffer to the manifestation of medical problems and somatization. These adult women
527
528
SEQUELAE OF CHILDHOOD SEXUAL ABUSE
were neither college students nor specifically gynecologic patients; therefore, this study corrects for several biases present in recent reports on the effect of sexual abuse in adult women.
STUDY SUBJECTS AND METHODS Invitations to participate in this study were mailed to a simple random sample of 1,443 women who were 18 to 50 years old and had used a rural outpatient family practice clinic for any reason during the 2-year period beginning Jan. 1, 1988. Of these invitations, 122 were returned with no forwarding address. Eighty-two women declined to participate by returning the stamped, addressed postcard and providing information on age and education; 689 neither responded nor provided information on age and education. The 550 women who agreed to participate were sent a ninepage questionnaire; 511 completed the questionnaire and returned it. This response represents 38.7% of those who received an invitation to participate. The questionnaire was designed to obtain information on various demographic factors, health risks, and medical symptoms. The sexual abuse history was elicited by asking a series of questions derived from apreviously used questionnaire. 3•20 As in the study by Briere and Runtz,' the participants were asked to identify up to two important sexual experiences they had had before 18 years of age with persons at least 5 years older-for example, strangers, friends, or family members. The level of coercion involved in the experience also was elicited. All women who reported such experiences were considered to have been sexually abused with the following exception. We excluded 13 women from the sexually abused group who were between 15 and 17 years old at the time of the reported incident because each related that the person involved was a friend, that no force was involved, and that an ongoing relationship had been established. Three women who reported forceful encounters with someone who was less than 5 years older than they were at the time of the incident were included in the abused group; the ages of these women at that time were 13, 16, and 17 years, respectively. Somatization was evaluated by using the somatization scale from the SCL-90. 21 The SCL-90 is a widely used screening measure, and the somatization scale consists of 12 items (for example, headache, numbness, weakness, and faintness) whose frequencies for the previous week were assessed by using a 5-point scale. The somatization scale has acceptable reliability and validity. Alcohol and drug abuses were assessed by posing the following three questions: Have you ever felt a need to cut down on the amount of alcohol you drink? Do you think you are a normal drinker? Have you ever had a problem with drugs or addictive medications?
Mayo Clin Proc, June 1992, Vol 67
Smoking as a health risk factor was examined in four ways: whether the participant had ever smoked, age at which smoking was begun, heaviest use (in packs/day), and duration of smoking (in years). Other health risk behaviors were evaluated by asking about age at first intercourse, number of sexual partners before 18 years of age, total number of sexual partners, pregnancy before 18 years of age,current height and weight, and frequency of Papanicolaou tests (Pap smears). Medical problems were assessed by using a short review of systems for various medical disorders and operations. The checklist included a total of 38 medical problems; the participant was asked to indicate all those she currently or previously had. Social support was assessed with four questions about number of close friends, satisfaction with the current level of social support, whether the abuse experience had been told to anyone, and whether emotional support was offered by those who were told about the abuse experience. The relationship of a history of sexual abuse, health risk behaviors, medical problems, and somatization was examined with the X2 technique or t test, as appropriate. Because of the exploratory nature of our study, P values of 0.05 or less were considered significant findings.
RESULTS Sample Description.-Demographic information on the entire study sample, the nonabused and abused subsamples, and two groups of nonresponders was determined for age, education, income, and marital status (Table 1). The sexually abused subsample was significantly older than the nonabused group (t = 2.6; P
