Sexual and Physical Abuse and Gastrointestinal Illness: Review and Recommendations

Methods

Each member of the working team was assigned a topic by the primary author. He or she then did a MEDLINE search on that topic and submitted it to the primary author, who integrated the material into a manuscript that was then resubmitted to the working team and revised. The final document was agreed on by consensus.

Clinical and Epidemiologic Associations

Methodologic Considerations in Evaluating Studies of Abuse Reporting

The widely differing estimates of the prevalence of abuse (6% to 62%) in the United States [9] result from the varying definitions and methods used to assess abuse history. Furthermore, police records and confirmation with family or acquaintances grossly underestimate the frequency of abuse, leaving no gold standard of validation. For clinicians, merely the disclosure of this information is considered truthful unless proven otherwise. However, to evaluate epidemiologic estimates of abuse history in clinical or population-based studies, clinicians and investigators must consider several factors.

Changing Societal Values about Definitions of Abuse

Numerous studies have suggested that the number of reports of sexual and physical abuse is high and may be increasing. In a review comparing the frequency of sexual abuse reported in the United States from the 1940s to the late 1970s, Leventhal [10] concluded that the frequency of these reports has increased (from 24% to 48% by the broad definition of abuse and from 12% to 28% by the narrow definition). The increase relates in part to changing societal values: The disclosure of an abusive experience is now encouraged and supported, whereas it was previously considered secretive and shameful. For example, society now considers date rape to be a form of sexual abuse. Thirty or 40 years ago, however, this experience may not have been defined as such, and victims may have been more reluctant to report it to officials. Investigators therefore must consider that the frequency of abuse reports are higher in areas where there is increased public attention to these events.

Interview versus Questionnaire

Some evidence suggests that interview methods may yield more reports of abuse than questionnaires [9, 11]. This theory is difficult to assess because most studies that use interviews also use more questions and specific activity-based questions that are known to increase abuse reporting. A carefully administered interview in a supportive environment may be the best way to identify a history of abuse [11], but this theory has yet to be tested adequately.

The Operational Definition

There is evidence that more persons report abuse when definitions include questions based on behavior (for example, “Has anyone ever touched the sex organs of your body when you did not want this?”) rather than general or emotionally charged questions (for example, “Have you ever been sexually abused or molested?”) [9, 12, 13]. Furthermore, a broader definition of abuse that includes many types of forced or unwanted sexual encounters (such as noncontact abuse or fondling) results in higher estimates of abuse prevalence. Noncontact sexual abuse includes unsolicited sexual advances or encounters with exhibitionists during childhood [9]. Noncontact experiences include attempted or threatened rape or sexual touching in which force is used but sexual contact does not occur (such as when the victim escapes). Contact abuse can include both touch experiences (that is, being fondled or being made to touch the perpetrator) and penetration (that is, vaginal sex, anal sex, or oral sex [14, 15]).

Similarly, physical abuse is identified by several variables: being assaulted or attacked with a weapon, beaten up, hit with a fist or object, kicked, bit, burned, slapped, or threatened with a weapon. Because the life-threat associated with these experiences can differ, investigators can use groupings of physical abuse experiences [in descending order of threat]: 1) being assaulted or attacked with a weapon [such as a gun or knife]; 2) being attacked without a weapon but with the intent to kill; 3) being beaten up, hit with a fist or object, kicked, bit, burned, or slapped by another without intent to kill; 4) being threatened with a weapon but not actually attacked; or 5) being threatened with harm but without a weapon or threat to life [16, 17].

The degree of coercion indicated in the question can affect estimates of the prevalence of abuse. Some investigators consider any unwanted sexual experiences to be abuse [18]. Others define abuse more rigorously as using force or threatening harm to engage in sexual acts [19-21]. Because abused persons are more likely to acknowledge abuse using the first definition and because questionnaires may tend to underestimate abuse [12], defining abuse as unwanted sexual experiences may be a more sensitive measure. However, with interviews, defining abuse in terms of force or threat of harm may be more valid because the nature of the abuse can be further clarified.

Many researchers stipulate that there be a 5-year age difference between the perpetrator and a child [9] so that the possibility of consensual sexual activity with peers can be eliminated. However, this definition may exclude abusive encounters with peers or siblings. We believe that by stipulating “unwanted or forced” sexual experience in the definition, the requirement of a 5-year difference in age is unnecessary. Estimates of childhood sexual abuse have also varied because different age criteria have been used to define childhood. The definition of childhood has ranged from 13 to 18 years, with some studies not defining what is meant by “child.” Age 14 years has recently been used as a cut-off for child and adult sexual abuse [22, 23].

Nature of the Setting and Patient Sample

The clinical setting in which the information is obtained may be as important as the type of questions asked. Studies done in referral practices (such as pain centers or academic practices) yield much higher response rates than those done in primary care or nonclinical settings. Similarly, patients seeing mental health professionals for emotional difficulties may be more likely to report abusive experiences than patients attending medical practices [24]. Furthermore, the prevalence of abuse tends to be higher in younger samples and perhaps among persons in certain regional areas (for example, urban compared with rural) [13, 25]. Investigators should also consider that the positive predictive value of a screening evaluation will probably be greater in clinical settings in which the prevalence is high than in nonclinical settings in which the prevalence is low.

The Psychosocial Profile of the Patient

Patients with certain psychiatric disorders (such as somatization disorder) or personality disorders (such as borderline personality disorder) may set low thresholds for reporting medical or psychological symptoms. These patients might therefore overinterpret and over-report previous experiences as abuse. In contrast, patients with dissociation disorders who may not recall these experiences, or those who harbor intense feelings of shame or guilt, are less likely to report a history of abuse. Finally, patients who are experiencing ongoing abuse, but who have limited social support or poorly developed coping skills or who fear retribution from the perpetrator, are much less likely to report these experiences. Investigators should recognize that concurrent psychosocial difficulties affect reporting tendencies.

Relationship with the Interviewer

Finally, persons may be more likely to disclose sensitive information when confidentiality is assured and when the interviewer or test administrator can effectively communicate trust and support [11]. The person doing this assessment should be adequately trained to address the patient's emotional response, and mental health resources should be made available if needed.

Abuse History and Gastrointestinal Symptoms

The epidemiologic investigation of abuse history is still in the early stage of its development, and additional work is needed to standardize the assessment of a history of sexual and physical abuse. When interview methods are not feasible, a standardized questionnaire may be helpful as a screening instrument, particularly when it is administered in a trusting and relaxed environment. Given these caveats, the following is a review of existing studies about the relation between abuse history and gastrointestinal symptoms.

Frequency of Gastrointestinal Symptoms among Sexually Abused Patients

In a review of the psychological and medical consequences of abuse in victimized female children, Bachmann and colleagues [26] found that a prominent theme was the development of psychological and physical symptoms, particularly of the gastrointestinal and genitourinary tract. For example, Rimsza and colleagues [27] did a chart review and telephone interview of the mothers of 72 female children and adolescents who had experienced forced sexual activity with an adult, comparing the frequency of behavioral and physical problems with the frequency in a matched control group of 26 children. The investigators found that the abused group reported significantly more physical symptoms (P < 0.01) and that the duration of abuse affected symptom reporting (P < 0.005), specifically the reporting of gastrointestinal (P < 0.01) and genitourinary (P < 0.01) symptoms. Seventy-one percent of children who were abused for more than 24 months reported gastrointestinal symptoms.

Felice and colleagues [28] described the development and progression of reports of physical symptoms in a chart study of a cohort of 25 female adolescent rape victims who were followed in a rehabilitation program. Phobias developed in approximately half of the rape victims soon after the rape, and this led to a period of denial and loss of interest in the treatment program. Years later, many of the victims returned for medical treatment with “psychosomatic” symptoms of abdominal pain, headaches, and dizzy spells, and they did not report an association between the physical symptoms and the previous rape. This observation suggests that for some patients, physical symptoms tend to emerge when recollection of the abusive experience diminishes. Although these two studies were limited in the rigor with which the outcome measures were assessed, the validity of the abuse history was assured, and the findings are consistent with those of retrospective studies.

Similar associations were found in studies of adults who had been sexually abused as children. In a health maintenance organization (HMO) clinic study, Felitti [3] identified 131 adult patients who acknowledged a history of abuse (abuse had occurred an average of 30 years earlier) from a questionnaire (specifically, from the answer to the question “Have you ever been raped or molested?”) and compared the frequency of current somatic symptoms with the frequency in a matched clinic control group. The group with an abuse history described a higher frequency of depressive symptoms (83% compared with 32%, respectively; P < 0.001); the next most common symptom was gastrointestinal problems (64% compared with 39%; P < 0.01), followed by headaches (45% compared with 25%; P < 0.05). In another study, Lechner and colleagues [29] administered a questionnaire to a consecutive sample of adult female patients seen in the waiting room of a family practice clinic. A positive response to the screening question “As a child 16 years of age or younger, were you ever a victim of sexual abuse… (with a person more than two years older)?” was obtained in 26% of the sample and led to a series of more detailed questions on sexual abuse. After this, the patients were asked questions about respiratory, cardiovascular, gastrointestinal, musculoskeletal, gynecologic, and otorhinolaryngologic symptoms, as well as mental health treatment. When abuse victims were compared with the group that had not been abused, significant differences were noted for reports of gastrointestinal (30.1% compared with 10.9%; P < 0.001), respiratory (15.4% compared with 6.2%; P < 0.002), and neurologic (7.4% compared with 2.1%; P < 0.01) symptoms. Abuse history was also associated with more mental health treatment (60% compared with 28%; P < 0.001). Although these two studies identified abuse victims through self-reports, the findings indicate that female victims of early abuse are more likely than others to have physical (and, in particular, gastrointestinal) symptoms later in life.

Prevalence of Abuse History in Gastrointestinal Practices

A high frequency of reports of sexual and physical abuse in a cohort of gastrointestinal patients was first described in 1990 [22]. The investigators evaluated a consecutive sample of women referred to the University of North Carolina gastroenterology practice over 2 months. The women were asked to complete a self-report screening questionnaire of sexual and physical abuse history Appendix 1 that was derived from previous studies [16, 18] and that contained the criteria noted above. This questionnaire was later validated against a detailed psychological interview [11]. Forty-four percent of the 206 patients studied reported a history of sexual or physical abuse in childhood or later in life. Furthermore, sexual abuse was strongly associated with physical abuse, and child abuse was strongly associated with history of abuse in adults. In addition, patients with functional gastrointestinal disorders reported more abuse (Table 1). For example, the frequency of rape or incest was 31% for patients who had functional gastrointestinal disorders (for example, the irritable bowel syndrome, functional dyspepsia, constipation, chronic abdominal pain) compared with 18% for those with organic disorders (for example, acid-peptic disease, inflammatory bowel disease, liver disease) (odds ratio, 2.08 [95% CI, 1.03 to 4.21]). In general, this history of abuse was concealed; the history of only 17% of the abuse victims was known to their physicians, and 30% of the victims had not previously disclosed this history to anyone. The findings also suggest that a standardized and confidential screening questionnaire can identify abuse history with reasonable sensitivity.

It is important to determine the generalizability of these findings to other clinical settings. Talley and colleagues [30] used the same self-report questionnaire to identify abuse and found a frequency of 30% among a consecutive sample of 68 men and 149 women seen at the Mayo Clinic (compared with a frequency of 44% among only female patients at the University of North Carolina). Patients who reported sexual abuse history were 2.8 times more likely to have a functional bowel disorder, a finding similar to that of the previous study. In a study of female patients referred to the University of Alabama for investigation of esophageal disorders (a more select patient sample) in which the same questionnaire was used, the frequency of abuse history was 56% [31]. Abuse frequencies among patients with different functional gastrointestinal disorders have been compared in only one study [32]. The frequency of sexual abuse was greater in patients with functional lower gastrointestinal tract disorders than in those with functional upper gastrointestinal tract disorders. Abused patients were more likely to report constipation, diarrhea, or animus (pelvic-floor dyssynergia) [32].

The high frequency of these abuse reports may be related to the selective nature of the study sample (patients with gastrointestinal symptoms seen at referral centers). This finding is supported by the findings of Longstreth and Wolde-Tsadik [33], who evaluated 1264 persons presenting for a periodic health examination at an HMO. Using the same screening questionnaire as in the previous studies Appendix 1, these investigators found lower rates of abuse reports. Women had a higher frequency of sexual abuse history (19%) than men (5.7%; P < 0.001). Twenty percent of this sample met criteria for moderate or severe irritable bowel syndrome, and these patients were compared with the women who did not have bowel symptoms (Table 2). The investigators found that unwanted sexual intercourse was reported by 5.2% of patients without the irritable bowel syndrome, 9.6% of those with less severe symptoms of the syndrome, and 22.2% of those with severe symptoms (Table 2). These data can be compared to the 31% frequency of abuse in patients with the irritable bowel syndrome who were referred to the University of North Carolina gastroenterology clinic (Table 1) [22]. As shown in Figure 1, the frequency of abuse history progressively increases in relation to the presence of the syndrome (compared with patients with no bowel symptoms), its severity (mild or severe), and the clinical population (HMO compared with referral center).

Only one population-based study has been reported [34]. A self-report questionnaire that included the previous questions about abuse (Table 5) was mailed to a random sample of 919 persons in Olmsted County, Minnesota (age range, 30 to 49 years). Twenty-six percent of the population reported some form of abuse in the past (age-adjusted prevalence was 41% for women and 11% for men), and most in this sample (22%) reported sexual abuse. In addition, sexual abuse history was associated with functional gastrointestinal symptoms. When compared with normal persons, patients who were sexually abused had an increased risk (odds ratio) for the irritable bowel syndrome of 1.9 (CI, 1.2 to 3.0) and a risk of 1.9 (CI, 1.3 to 2.9) for functional dyspepsia.

Relation between Abuse History and Psychiatric Disturbance

A history of physical or sexual abuse during childhood is associated with psychiatric disturbances. In a survey of a nonclinical population (278 women attending a university), 15% reported having had unwanted sexual contact with a substantially older person before they had reached the age of 15 years [35]. Using a modified version of the Hopkins Symptom Checklist, the investigators found that the affected persons had significantly higher levels of dissociation (P < 0.009), somatization (P < 0.03), anxiety (P < 0.03), and depression (P < 0.05) than the women who were not abused. Similarly, a higher frequency of psychiatric diagnoses was reported among clinical patients with gastrointestinal disorders who reported abuse history. Walker and colleagues [36] evaluated a cohort of patients with the irritable bowel syndrome and inflammatory bowel disease using a 14-item questionnaire [37] that ascertained sexual abuse history, the severity of the abuse, and the victim's relationship with the perpetrator. Patients who had been severely abused were more likely than those with no history of abuse or those with a history of less severe abuse to have several lifetime DSM-III (Diagnostic and Statistical Manual, Third Edition) diagnoses: phobia (67% compared with 26% [odds ratio, 5.6; CI, 1.2 to 26.7]), panic disorder (44% compared with 11% [odds ratio, 6.8; CI, 1.3 to 36.7]), somatization disorder (56% compared with 11% [odds ratio, 10.6; CI, 2.0 to 56.7]), alcohol abuse (44% compared with 13% [odds ratio, 5.3; CI, 1.0 to 26.7]), functional dyspareunia (78% compared with 21% [odds ratio, 13.1; CI, 2.3 to 75.9]), and major depression (78% compared with 34% [odds ratio, 6.7; CI, 1.2 to 27.2]).

Dissociation, which is commonly seen in persons who have been abused [38], may allow the patient to psychosocially adapt to the experience. In these patients, the “depersonalization” and “blanking out” occur as painful memories materialize and help the patient to avoid reexperiencing emotional trauma. This process can range from isolated, brief flashbacks and nightmares to amnesia or full-blown maladaptive personality styles. Patients with dissociation tend to frequently and persistently report physical symptoms [39]. As previously observed by Felice and colleagues [28], dissociation and physical symptoms may be a late outcome that occurs when the emotional trauma from childhood abuse is “forgotten.” Dissociation and somatization may also be associated with other psychiatric diagnoses such as multiple-personality disorder and borderline-personality disorder [40], or it may exist as a behavior pattern independent of other psychiatric diagnosis among patients with gastrointestinal or other medical diagnoses.

Relation between Abuse History and Other Medical Diagnoses

Several studies have shown that abuse history is independently associated with certain other (nongastrointestinal) medical syndromes [7] and is strongly associated with increased symptom reporting (for example, somatization) in general [41-43]. At referral centers, diagnoses frequently associated with abuse history include chronic pain (such as pelvic pain [44-46], headache [47], and back and myofascial pain [48]) and eating disorders, particularly bulimia nervosa [49], morbid obesity [3], and alcohol and drug abuse [50, 51]. Abuse history probably has a general effect on increased symptom reporting and a poorer adjustment to illness. Well-designed epidemiologic studies are needed to determine whether the putative preferential association of abuse history with gastrointestinal illness is valid or whether these observations are related to the greater attention recently given to gastrointestinal disorders.

Relation between Abuse History and Health Status

The relation between abuse history and poor health status has implications for treatment. In the University of North Carolina study of gastroenterology patients [22], those with an abuse history were more likely than those without an abuse history to report pelvic pain (16% compared with 6% [odds ratio, 4.05; CI, 1.41 to 11.69]), multiple somatic symptoms (7.1 ± 0.28 symptoms compared with 5.8 ± 0.25 symptoms; P = 0.001) and more lifetime surgeries (2.8 ± 1.10 surgeries compared with 2.0 ± 1.09 surgeries; P = 0.009). Similar findings were found in an HMO study [33], in which abuse history was significantly associated with at least one nongastrointestinal symptom (P = 0.016) and at least one surgery (P = 0.034). Finally, in the population-based study [34], the odds of visiting a physician were significantly higher in persons with a history of any type of abuse (odds ratio, 1.4 [CI, 1.0 to 2.0]), sexual abuse (odds ratio, 1.5 [CI, 1.0 to 2.2]), and emotional or verbal abuse (odds ratio, 1.8 [CI, 1.1 to 3.0]), but not for those with a history of physical abuse alone. Therefore, regardless of practice type or diagnosis, abuse history is associated with poor adjustment to illness and adverse clinical outcomes, including an increased amount of health-care seeking and a greater risk for surgical procedures.

Summary of Clinical Data

We obtained the following information from our review of the literature: 1) Abuse history, psychiatric disturbance, and medical symptoms are significantly associated; 2) the apparent preferential association of abuse history with gastrointestinal symptoms requires epidemiologic confirmation; 3) abuse history is more commonly reported by women; 4) abuse history is more frequent in referral practices than in primary care; 5) patients with functional gastroenterologic disorders report abuse more frequently than patients with organic disorders; 6) abuse history is associated with an increased tendency to seek health care and poorer health status [psychological dysfunction, symptom reporting, more frequent surgery]; and 7) the patients' physicians are usually unaware of the history of abuse.

A Conceptual Model

We propose a conceptual model to explain the relation between abuse history and psychosocial disturbance, gastrointestinal illness, and health care utilization (Figure 2). Additional studies are needed to confirm the effect of the contributing factors.

1. Abuse can be a traumatic event that has long-lasting psychosocial consequences. Abusive experiences may produce long-standing symptoms of psychological distress and may predispose patients to psychiatric diagnoses (for example, post-traumatic stress disorder; anxiety; and depressive, somatoform, or personality disorders [39, 41, 43, 45, 52]). Furthermore, abusive experiences are often part of and contribute to a milieu of poor social support and the development of ineffective coping strategies [5, 31].

2. Susceptibility to gastrointestinal illness in combination with psychological disturbances may mediate the development or exacerbation of gastrointestinal symptoms. Psychological distress generally lowers symptom threshold, thereby increasing the number and severity of symptoms reported for many medical conditions. However, the functional gastrointestinal disorders are so common that many susceptible persons may develop gastrointestinal symptoms. This may occur through any of several plausible mechanisms [8]:

Psychophysiologic effects: Psychological distress can, through the central nervous system-enteric nervous system [53] or autonomic pathways [54], produce exaggerated intestinal motility [55] and abdominal discomfort [56]; this occurs to a greater degree in persons with functional bowel disorders [57-59].

Enhanced visceral sensitivity: Prolonged visceral stimulation from injury or inflammation evokes the activity of previously unresponsive silent nociceptors that, along with other sensory afferents, amplify peripheral input to the central nervous system to produce persistent pain even after peripheral afferent activity decreases [60, 61]. For example, repetitive noxious distention of the colon in humans produces a progressive increase in pain scores and an increased area of referred sensation to the abdomen [62]. Therefore, traumatic stimulation of the vagina or anus in children may lead to neural changes [63] that down-regulate the sensation thresholds of visceral nociceptors, thereby increasing perception of abdominal or pelvic pain or other bowel symptoms [8, 61].

Psychodynamic effects: Young children who have been sexually abused often believe that their sexual organs are dirty or bad; this can lead to feelings of guilt and shame. The psychological distress engendered by these feelings may be expiated through physical pain or suffering [64, 65]. The location of the pain in the abdomen, pelvis, or genitourinary area is logical to the psyche because this area represents the “bad” or offending part of the body that is to be punished.

Response bias: Patients with gastrointestinal disorders and abuse histories have significantly lower pain thresholds than normal persons, and this relates in part to the psychological tendency to set low standards for judging stimuli as painful (response bias) [31].

Psychiatric comorbid conditions: Patients with functional gastrointestinal disorders who are seen in medical centers have a high frequency of psychiatric comorbid conditions [66, 67] and are likely to communicate psychological distress through physical symptoms. Furthermore, they may not acknowledge a relation between psychological difficulties and their symptoms [68, 69], which reinforces their belief that they have a serious medical disorder.

Early-life reinforcement of illness behavior: Increased attention paid to reports of illness early in life may lead to reinforcement of illness behaviors [70-72] and the seeking of health care.

3. Gastrointestinal symptoms, combined with existing psychosocial difficulties, amplify the symptom experience and lead to health-care seeking, refractoriness, and, ultimately, referral [73-75]. Continued referral is reinforced because the health care system provides only incomplete relief of somatic symptoms, a socially acceptable and secure social support system, an idealized “parental” figure (physician) to take responsibility for “cure,” and possible reinforcement of pain and suffering through submission to unneeded diagnostic and therapeutic procedures.

4. A “vicious cycle” of psychological difficulties, increased symptom severity caused by refractory symptoms, and health-care seeking is perpetuated. Refractory symptoms amplify psychological disturbance, which in turn worsens symptoms and illness behavior. The cycle continues when the physician does not consider the psychosocial determinants of these disorders and responds to the patient's distress by administering unneeded testing or treatments.

Obtaining an Abuse History

Because abuse history may adversely affect illness outcome [22, 33, 76], it is important to elicit a history of abuse among patients at high risk and to initiate appropriate referral with the hope that the clinical condition improves. The clinician should consider several factors in deciding when and how to address this issue [77].

Risk Factors for Abuse History among Patients with Gastrointestinal Illness

Several clinical features increase the likelihood that a person has been abused (Table 4).

Psychological Issues

Patients with abuse histories have been violated in ways that erode their sense of trust in others, particularly if they depend on the perpetrator as a source of support. This may lead to a cycle of helplessness and lack of control over the actions in their life, which fosters dependence on those whom they cannot trust. If these experiences occur early in life, the child cannot discriminate right from wrong or personal responsibility from the responsibility of others. This may lead to a pervasive and unjustified sense of shame and guilt: They may blame themselves for actions they did not initiate. In clinical practice, patients demonstrate these difficulties by 1) distrusting physicians but feeling dependent on them for attention and security; 2) harboring unrealistic (magical) expectations for cure and overtly delegating decisions to physicians while engaging in passive conflicts over control [78]; 3) submitting to unneeded or painful procedures, which may expiate feelings of guilt; or 4) contributing to interpersonal difficulties that lead to anger or rejection by physicians, which in turn reinforces feelings of worthlessness and poor self-esteem. Such difficulties can only be ameliorated when the bases for these behaviors are understood.

Medical and Psychiatric Disorders

Certain medical and psychiatric disorders are strongly associated with a history of abuse. Medical disorders include chronic pain [48, 79]; severe constipation [80], particularly pelvic dyssynergia (obstructive defecation); the eating disorders [49]; morbid obesity [3]; unexplained vomiting; and sexual dysfunction [46, 81]. Several psychiatric disorders may develop concurrently: the somatoform disorders [39, 41, 43], the dissociation disorders and multiple-personality disorder [4, 82], post-traumatic stress disorder [83], and severe depression or panic disorder [39].

Illness-Related Behaviors

Illness behavior defines the degree to which symptoms are reported (its frequency, severity, and qualitative description), the manner in which it is communicated (with a stoic, suffering, or help-seeking demeanor), and its clinical effects (self-care, narcotic use, disability, health care utilization) [84]. When the physician observes that the patient's behavior surpasses the usual range of expectation, the illness behavior is considered abnormal [85]. This often occurs when the patient displays disability that is disproportionate to the clinical data, seeks to validate disease (often excluding a role for psychological factors), places responsibility for the treatment with the physician, and avoids health-promoting behaviors.

Patients with abuse histories may also have difficulties with certain procedures (such as rectal or pelvic examination and endoscopies), and patients may show “borderline” behaviors, such as forming intense and chaotic emotional attachments with physicians, “splitting” the health care team by playing one person against another, or being unable to accept the uncertainties of medical care by demanding diagnostic procedures or treatments for “cure.”

Unwanted Outcomes

Inevitably, the above features lead to unwanted effects. Patients may undergo multiple diagnostic procedures, treatments, and surgeries [86]; abuse alcohol, medications, and other substances [7]; seek disability and litigation; and overuse health care services [22, 33, 76].

When To Inquire about an Abuse History

Physicians should inquire about a history of abuse when the clinical data are suggestive (Table 4) and the information will help improve the patient's outcome. When the symptoms are refractory and disabling and an abuse history is suspected, determining current or previous abuse is important because the physician can then refer the patient to a mental health professional. The mental health specialist in turn can address and alleviate the psychological encumbrances of the experience or experiences. The patient's risk for becoming significantly distressed from discussing these experiences is low [22]. In an ongoing study by one of the authors (DD) that involves intensive interviews of more than 100 patients with abuse history, the ability to discuss the abuse history was considered beneficial or at least not harmful to the patient in almost all cases. Only 3 patients required immediate additional counseling.

Physicians should also inquire about an abuse history if they believe they can discuss the topic. Certainly, physicians may feel uncomfortable addressing these issues or may be unable to take the time to allow adequate discussion of the patient's thoughts or feelings. In these situations, it is understandable and appropriate to defer the inquiry or to refer the patient to a mental health colleague (see below).

When eliciting an abuse history, physicians should ensure that resources for referral are available. Obtaining this history is only the first step in what may be a long treatment process to help the patient work through the thoughts and feelings of the abuse experience. The physician should therefore be prepared to provide additional counseling services.

How To Ask about an Abuse History

The history should be obtained in a clinical setting that is free from interruption and after the physician establishes rapport and mutual trust with the patient. We recommend that, rather than directly ask these questions, the physician follow the patient's lead and look for an opportunity to address the issue of abuse [87]. If the patient refers to an experience suggestive of abuse (for example, “Things were pretty horrible then …”), then the physician should encourage the patient to elaborate. If the patient does not volunteer this information, the physician should provide further opportunity: “Is there anything else you would like to discuss that you think is important?” If the information is still not forthcoming, the physician can ask if any experiences not yet discussed have been particularly painful or difficult. Finally, if the physician is still concerned that the patient has been abused, he or she can ask about it more directly: “As you may know, it's not uncommon these days for persons to have been emotionally, physically, or sexually victimized at some time in their life and this can affect how people manage with their medical condition. Has this ever happened to you?” This approach communicates that the patient is not alone with this type of experience. It also allows the patient to define victimization in personal terms, and the physician can clarify the responses. Although a more direct inquiry has also been recommended, we believe our approach reduces the pressure to disclose. This improves the response rate and preserves the therapeutic relationship and the patient's right not to disclose this information if feeling unable or unwilling.

It is important to observe the manner in which the information is presented. If the patient denies this history but the nonverbal response is incongruent, the physician should register the information for future inquiry and say no more. If the patient acknowledges a history of abuse, the physician should remain nonjudgmental and encourage the patient to continue. Obtaining the details of the experience or experience is not as important as being supportive and empathic, an approach that allows the patients to express previously suppressed thoughts and feelings in his or her own way. Inquiring about this information in a supportive manner may be therapeutic. The information will not be disclosed if the patient cannot or is unwilling to do so.

After the history is obtained, the physician needs to monitor the patient's comfort level and decide whether to continue. Gentle encouragement provides the opportunity for the patient to say more. Periods of silence must be permitted so the patient can collect thoughts and feelings in order to coherently present them. Some emotional distress is expected and should be permitted. However, if the patient begins to change or evade the topic, turn away with arms folded, or present emotion-laden information in a disaffected or emotionally incongruent manner, the physician should end the discussion but provide the option to discuss this issue again in the future. The question items listed in Appendix, although developed primarily for research, can also be used by clinicians to determine the severity of abuse.

How To Make a Referral

Referral to a mental health professional should be consistent with the patient's needs and expectations. If the patient freely and congruently discloses the abuse experience, the physician should acknowledge that he or she has shared some very important, private, and meaningful experience and that these experiences have clearly affected the patient's feelings and ability to cope with the illness. For these reasons, it may be helpful to seek further psychological treatment.

In some cases, patients believe they have come to a physician for a medical problem and may be surprised or feel ashamed about discussing their emotional experiences. They may question the relevance of these experiences to the medical condition. In this case, the physician needs to help the patient accept the importance of referral. The physician can first acknowledge that sharing important personal information must have been difficult for the patient and that the physician will be available in the future if the patient wishes to discuss it further. The physician can also address the difficulties the patient has had with the illness and can refer to statements of feeling depressed, despondent, or unable to cope with the illness. The physician can then use the patient's comments to emphasize the importance of the mental health professional in helping him or her to reverse the vicious cycle of continued symptoms and psychological distress. Finally, the physician could mention that improvement in psychological distress can increase tolerance of pain and provide an overall better adjustment to medical illness. Because of feelings of shame or of an inability to deal with the emotions generated, the patient may be reluctant to see a mental health consultant. In this situation, the physician must accept the patient's wishes, continue in the care, and suggest that the topic can always be discussed again in the future. In all cases, the physician should maintain continuity of care. To refer the patient to a mental health professional and then remove oneself from the treatment would be viewed by the patient as a rejection.

If the patient is willing to be referred, a telephone call or letter should precede the visit, with an offer made to maintain ongoing communication. The type of mental health professional (such as psychiatrist, psychologist, or psychiatric social worker) is not as important as choosing someone who has experience in working with abused patients. Peer support groups or group therapy with other abused patients can also be recommended because they may accelerate the patient's ability to communicate shameful thoughts to an understanding and supportive audience.

The Role of the Mental Health Consultant

We recognize that addressing the psychological difficulties patients have in coping with chronic pain and illness can be intimidating for the nonpsychiatric physician. Distressed patients who are demanding, not well understood by their physicians, and refractory to treatment produce frustration, despair, and, at times, hostility on the part of the treating physician [78, 88, 89]. Frequently, however, these are the patients in whom a history of abuse is likely, and a physician's reluctance to inquire about this sensitive and emotion-laden topic is understandable. Nevertheless, ignoring this important history only further hampers the clinical progress of the patient and the physician–patient relationship.

When an abuse history is suspected to contribute to refractory illness, the medical physician may choose to defer exploration of psychological issues and recommend consultation with a mental health professional. This person can provide specialty care and also serve as an additional source of clinical support and emotional validation.

The mental health consultant can provide several services.

1. Identify psychological comorbid conditions. The mental health consultant can confirm whether a major psychological disturbance is present, and this may lead to specific forms of psychopharmacologic or psychological treatment. In a clinical interview, the consultant may screen for psychiatric diagnoses by using standardized criteria (Diagnostic and Statistical Manual, Fourth Edition [DSM-IV]) [90] or by administering validated questionnaires to identify certain psychological features. Some questionnaires, such as the Spielberger Anxiety State-Trait measure [91], the Beck Depression Inventory [92], the Sheehan panic disorder questionnaire [93], and the Eating Disorders Attitudes Test [94], can be administered by the mental health consultant or even the medical physicians at the time of the patient's visit. More comprehensive personality and symptom inventories such as the Minnesota Multiphasic Personality Inventory [95] and the Hopkins Symptom Checklist 90 (SCL-90) [96] give complex information that is usually obtained by psychologists or psychiatrists.

2. Identify a history of abuse. The mental health consultant can confirm the physician's findings of abuse history or can independently identify its existence. He or she will then determine with the patient whether additional treatment is needed. This is usually done if the abuse history is contributing to poor adjustment to illness or to significant psychological distress. The consultant might also determine that the abuse history does not require further action (for example, if the patient has previously sought help or has adjusted psychologically to the experience).

3. Decide on psychopharmacologic treatment. The psychiatric consultant determines whether a psychopharmacologic agent would be helpful either as primary treatment or as ancillary to psychological care. The consultant also considers factors such as the patient's medical condition, concurrent medications, and the nature of the psychological problem to decide on the best medication choice and dosage.

4. Initiate concurrent psychological treatment. When needed, the mental health consultant can either personally initiate or implement referral for psychological treatment along with the medical care. The goal of therapy is to encourage the patient to accept the reality of the abuse experience and explore and revise maladaptive thoughts and feelings. For example, patients abused early in life frequently feel responsible for the events, and this leads to pervasive feelings of guilt and shame. Clinical symptoms usually improve and self-esteem is usually heightened when patients can separate themselves from the experience and redirect the responsibility to the perpetrator. Including a psychopharmacologic agent in addition to the psychological care may be helpful [97].

5. Assume primary responsibility for the patient's care. In some circumstances, the psychological difficulties will be so great that the mental health consultant will take the primary role in the care of the patient either through psychiatric hospitalization or outpatient therapy. However, the medical physician should still continue treatment, albeit less frequently, to monitor the patient's medical symptoms. Because many patients consider mental health referral a rejection by the medical physician, their care is best managed by a combined medical and psychological approach [98, 99].

Patient Support Groups

Victims of abuse may find it difficult to communicate with physicians because of shame or because their trust in authority has been eroded. Referral to a patient support group (see Appendix 2) may help by promoting helpful exchanges among persons with similar experiences. Patients with gastrointestinal disorders are particularly stigmatized because, aside from their emotional difficulties, they have embarrassing and socially unacceptable physical symptoms.

Many communities have mental health resources that either have or are willing to establish a patient support group led by an experienced group facilitator. Patient support groups can help patients adjust to and gain control over chronic illness or the effects of abuse. Certain medical support groups (for example, the International Foundation for Bowel Dysfunction and the Crohn's and Colitis Foundation of America) organize local groups to address the personal experiences of having specific gastrointestinal disorders (the irritable bowel syndrome, ulcerative colitis, and Crohn disease). They can also provide referrals for persons who want to work primarily on issues related to abuse.

Physicians who are unfamiliar with their local support network can access them through several entry points:

1. Self-help groups for crime victims or victims' assistance organizations. Most communities have a local crime victims' assistance resource. These agencies provide free counseling, support, and referral for victims. They are usually familiar with issues relating to post-traumatic stress disorder, spousal abuse, and adult survivors of child sexual abuse.

2. Women's health centers. Independent or hospital-based women's centers that focus on women's health issues exist in most communities and either sponsor or provide referral to ongoing support groups.

3. Women's shelters. Most communities have shelters to which abused persons can go 24 hours a day. Persons working in these shelters also have access to ongoing support groups.

4. Young Women's Christian Association (YWCA) and Young Women's Hebrew Association (YWHA). Part of the national mandate of the YWCA and YWHA is to promote women's health. Although not all local YWCAs or YWHAs are active in this area, in many communities they provide a valuable resource and can refer patients to other peer or professional support systems active in women's mental health issues.

If local organizations are not available, several national organizations can provide information and referrals for physicians or their patients. Their addresses and telephone numbers are listed in Appendix 2.

Conclusion

We agree that a history of abuse is associated with certain chronic medical conditions, particularly in women seen at referral centers and those seen for functional gastrointestinal disorders. Additional studies are needed to determine the degree to which abuse history has general as opposed to disease-specific effects on symptom reporting. Finally, a history of abuse is associated with poorer health status. For this reason, we recommend that physicians become aware of the many risk factors among patients with chronic or severe refractory symptoms, and, when these symptoms are present, that they inquire about a history of sexual or physical abuse. If abuse has occurred, we believe that appropriate referral to a mental health professional may help alleviate psychological distress and symptom severity and possibly improve the outcome.

Dr. Talley: University of Sydney, Clinical Sciences Building, Nepean Hospital, Box 63, Penrith, New South Wales 2751, Australia.

Dr. Leserman: University of North Carolina, Department of Psychiatry, CB #7160, Chapel Hill, NC 27599-7160.

Dr. Olden: 1 Shrader Street, Suite 550, San Francisco, CA 94117.

Dr. Barreiro: United Medical Associates, 1159 Vestal Avenue, Binghamton, NY 13903.