A history of sexual abuse and health: a Nordic multicentre study

Authors


Dr M. Hilden, Centre for Victims of Sexual Assault, Rigshospitalet, 4031 Blegdamsvej 9, DK-2100 Copenhagen, Denmark.

Abstract

Objectives  To determine if a history of sexual abuse is associated with objective and subjective indicators of health and if certain abusive incidents had a stronger impact on health than others.

Design  A cross-sectional, multicentre study.

Setting  Five gynaecological departments in the five Nordic countries.

Sample  Three thousand five hundred and thirty-nine gynaecology patients.

Methods  The NorVold Abuse Questionnaire (NorAQ) on abuse history and current health was mailed to all patients who consented to participate.

Main outcome measures  Reason for index visit at the gynaeocological clinic as well as several questions on health were recorded. General health status was measured as self-estimated health, psychosomatic symptoms (headache, abdominal pain, muscle, weakness, dizziness), number of health care visits and number of periods on sick leave.

Result  A history of sexual abuse was reported by 20.7% of respondents. A history of sexual abuse was significantly associated with chronic pelvic pain as reason for index visit (P < 0.01), laparoscopic surgery (P < 0.01), psychosomatic symptoms (P < 0.01), self-estimated poor health (P < 0.01), many health care visits (P < 0.01) and high incidence of sick leave (P < 0.01). Several subgroups within the group of sexually abused women were more likely to report poor health: women abused as both children and adults, women who experienced additional emotional and/or physical abuse and women abused by a person they knew.

Conclusion  Sexual abuse has a profound impact on women's health. Taking a history of sexual abuse seems particularly warranted when the patient presents with chronic pelvic pain or symptoms of a vague and diffuse nature.

INTRODUCTION

Sexually abused patients frequently complain of symptoms emanating from the lower abdomen and the pelvis.1–14

However, not all patients with a history of sexual abuse develop long term sequelae. Characteristics of the abuse, such as timing, course and circumstances, as well as the relationship with the abuser, might influence the impact that such abuse has on health.

We investigated to what extent a history of sexual abuse was related to various health variables in a population of gynaecology patients.

In general, we expected to find that a history of sexual abuse was associated with poor overall health. In addition we expected that as the scale of the abuse increased, so the ill health of women would increase.

METHODS

The study is part of a multicentre study conducted at a total of five gynaecological departments in Denmark, Finland, Iceland, Norway and Sweden. The recruitment period was from October 1999 to January 2001. Consecutive in- and outpatients were invited to participate and received the NorVold Abuse Questionnaire (NorAQ). Of 4729 eligible patients, 3641 answered the NorAQ (77%) and 3,539 (75%) women answered the questions on history of sexual abuse.

Inclusion criteria were age 18 years or more and literacy in the national language. In the Nordic countries, the age of majority is attained at 18 years of age, the age of consent is either 15 or 16 years of age depending on the laws of the individual countries.

The local ethical committee in each country approved the study. More detailed descriptions of the samples and procedures are published elsewhere.15

The NorAQ was developed to quantify ‘abuse’ and has been validated in a separate study in a Swedish population.16 The concurrent validity and test–retest reliability for the questions regarding abuse were good. Complete versions of The NorAQ are available in English and in all Scandinavian languages.17

The NorAQ includes general questions on sociodemographic variables, reasons for the index visit, self-estimated health and medical history and questions about lifetime experiences of emotional, physical and sexual abuse and of abuse in the health care system.

The questions on sexual abuse range from experiences of mild to severe lifetime abuse and permit classification according to ‘degree of severity of the abusive act’(Fig. 1). If a woman reported several degrees of severity in her experiences of sexual abuse she was classified according to the most severe act of abuse. If a woman had ever experienced sexual abuse, she was invited to answer more detailed questions. In the analyses, the answers concerning sexual abuse were reduced to two categories; ‘no’ including no abuse and mild abuse and ‘yes’ including moderate and severe abuse. Subgroups of abusive experiences were identified: if the abuse occurred exclusively in childhood (<18 years), exclusively in adulthood or in both periods; whether penetration had occurred (vaginal, oral or anal); if any sexual abuse had taken place within the 12 months before participation in the study; and whether the woman had lifetime experiences of additional abuse (emotional and/or physical abuse).

Figure 1.

NorVold Abuse Questionnaire.

The victim's relation with the abuser(s) was categorised into (1) partner (including ex-partner), (2) family (including stepparents), (3) known person (excluding partner and family) and (4) unknown person. In cases with several abusers, the answers were prioritised according to the order above.

When evaluating overall health, the following health variables were used: self-estimated health, reported psychosomatic symptoms, number of health care visits, periods of sick leave and the reason for the index visit.

Patients were asked to indicate the reason for the index visit on a checklist where several options could be marked. The distributions are listed in Table 1. Reasons thought to be closely related to sexual abuse according to the literature were selected for further analyses: chronic pelvic pain, irregular and/or heavy menstruation, acute pelvic pain, infertility and termination of pregnancy.

Table 1.  Reasons for encounter at the index visit among gynaecology patients in five Nordic countries (N= 3479). Values are given as n (%).
 n (%*)
  • *

    Adds up to more than 100%, as several reasons could be marked.

Irregular and/or heavy menstruation593 (17.0)
Infertility438 (12.6)
Urogynaecological problems316 (9.1)
Cervical smear286 (8.2)
Pelvic pain (>3 months duration)284 (8.2)
Early pregnancy problems288 (8.3)
Termination of pregnancy274 (7.9)
Pelvic pain (<3 months duration)263 (7.6)
Ovarian cyst, fibroma and prolapsus uteri183 (5.3)
Gynaecological cancer152 (4.4)
Pre- and post-operative assessment161 (4.6)
Contraception117 (3.4)
Gynaecological check up105 (3.0)
Vaginal discharge91 (2.6)
Climacteric complaints44 (1.3)
Sexual transmitted disease (STD)44 (1.3)
Various289 (8.3)

Questions on how a patient rated her health within the last year were classified as ‘poor’ when the responder answered ‘poor’ or ‘very poor’ and ‘good’ when the answer was ‘good’ or ‘very good’. The patient was asked if, within the last year, she had experienced physical symptoms such as headache, abdominal pain, muscle weakness and dizziness to such an extent that daily life was affected. If she answered ‘yes, sometimes’ or ‘yes, often’ the answer was classified as ‘yes’ to psychosomatic symptoms and ‘no’ when the answer was ‘yes, rarely’ or ‘no’. Additionally, overall health status was evaluated by number of health care visits and periods of sick leave. Patients who reported more than seven health care visits within the latest year were classified as high consumers, and patients on sick leave for more than two weeks within the last year were classified as having a high incidence of sick leave.

No medical data from the patient's records were collected.

Statistical tests of differences were performed with the Pearson χ2 test; however, when the cell sizes were small, Fisher's exact test was used. P≤ 0.01 were referred to as statistically significant in the univariate analyses. All χ2 values and levels of significance are included in the tables. Logistic regressions were performed with backward elimination, including all factors significant in the bivariate analyses, and the sociodemographic variables: age, partner status and country of residence (significance level at P≤ 0.05).

RESULTS

Table 2 shows the sociodemographic characteristics of the total population and of the populations with and without a history of sexual abuse. One-fifth (20.7%, n= 731) of the patients reported a history of sexual abuse. The sexually abused patients were younger and more likely to have no regular partner and be on welfare support than the non-abused; however, welfare support was not significant at the 1% level (P= 0.03). Educational and occupational status did not differ between the two groups.

Table 2.  Sociodemographic characteristics of gynaecology patients in five Nordic countries in total and with and without a history of sexual abuse. Values are given as n (%).
 Total (n= 3539)History of sexual abuse
No (n= 2808)Yes (n= 731)
  • Comparisons were made between the groups yes and no to sexual abuse history.

  • *

    χ2(4)=35.95, P < 0.00.

  • **

    χ2(2)=25.97, P < 0.00.

  • ***

    χ2(4)=67.75, P < 0.00.

Age*
18–29 years856 (24.2)667 (23.8)189 (25.9)
30–39 years1084 (30.6)844 (30.1)240 (32.8)
40–49 years693 (19.6)518 (18.4)175 (23.9)
50 and above887 (25.1)762 (27.1)125 (17.1)
Missing19 (0.5)17 (0.6)2 (0.3)
 
Education
≤9 years560 (15.8)445 (15.8)115 (15.7)
10–12 years1266 (35.8)997 (35.5)269 (36.8)
≥13 years1687 (47.7)1345 (47.9)342 (46.8)
Missing26 (0.7)21 (0.7)5 (0.7)
 
Occupation
Working outside home2380 (67.3)1901 (67.7)479 (65.5)
Housewife/maternity leave242 (6.8)205 (7.3)37 (5.1)
On welfare support612 (17.3)463 (16.5)149 (20.4)
Student281 (7.9)222 (7.9)59 (8.1)
Other/missing24 (0.7)17 (0.6)7 (1.0)
 
Regular partner**
No792 (22.4)578 (20.6)214 (29.3)
Yes2715 (76.7)2206 (78.6)509 (69.6)
Missing32 (0.9)24 (0.9)8 (1.1)
 
Country***
Denmark795 (22.5)627 (22.3)168 (23.0)
Finland583 (16.5)457 (16.3)126 (17.2)
Iceland647 (18.3)449 (16.0)198 (27.1)
Norway692 (19.6)563 (20.0)129 (17.6)
Sweden822 (23.2)712 (25.4)110 (15.0)

Iceland and Denmark had the highest prevalence rates of sexual abuse, 27.1% and 23.0%, whereas Finland, Norway and Sweden had prevalence rates between 15% and 18%.

Table 3 shows health variables by sexual abuse history. Chronic pelvic pain as reason for index visit and having ever had laparoscopic surgery were the only gynaecologic variables associated with a history of sexual abuse. All indicators of poor overall health (poor self-estimated health, psychosomatic symptoms, high consumption of health care visits and high incidence of sick leave) were clearly associated with a history of sexual abuse.

Table 3.  Health variables by sexual abuse history among 3539 gynaecology patients in five Nordic countries. χ2 test P values displayed. Values are given as n (%).
 Sexual abuse historyP
No (n= 2808)Yes (n= 731)
  • *

    HCV = Health care visits.

Acute pelvic pain202 (7.2)61 (8.3)0.29
Chronic pelvic pain194 (6.9)90 (12.3)<0.00
Termination of pregnancy205 (7.3)69 (9.4)0.05
Irregular and/or heavy menstruation455 (16.2)138 (18.9)0.09
Infertility358 (12.7)80 (10.9)0.19
Ever had laparoscopic surgery863 (30.7)298 (40.8)<0.00
Ever had a laparotomy697 (24.8)194 (26.5)0.32
Poor self-estimated health498 (17.7)198 (27.1)<0.00
Psychosomatic symptoms831 (29.6)328 (44.9)<0.00
High consumer of HCV*508 (18.1)190 (26.0)<0.00
High incidence of sick leave794 (28.3)298 (40.8)<0.00

In women who had been sexually abused, we found no association between reasons for the index visit and the relationship with abuser(s), or being a child or an adult at the time of the abuse. Reporting sexual abuse within the last year was associated to requesting termination of pregnancy at the index visit (P= 0.02). Irregular and/or heavy menstruation was a more frequent reason for index visit (P= 0.01) among women who had experienced sexual abuse combined with physical or emotional abuse than among women with solely a sexual abuse history.

Table 4 shows that women who had experiences of sexual abuse both as children and as adults, who experienced penetrative abuse, who had a family relationship with the abuser or who experienced additional emotional and/or physical abuse were more likely to report poor overall health than abused women who did not belong to any of these groups. Whether any abuse had occurred within the last year or earlier was the only stratification not significant on any variables.

Table 4.  Selected health variables in relation to history of sexual abuse stratified in to subgroups according to circumstances of the sexual abuse experience among gynaecology patients reporting a sexual abuse history. Values are given as n (%) and χ2 test P values.
 Time of sexual abusePenetrative abuseAbuse within the last 12 monthsHistory of physical/emotional abuseRelation between abuser and victim
Child (n= 379)Adult (n= 246)Both (n= 103)PNo (n= 171)Yes (n= 516)PNo (n=681)Yes (n=39)PNo (n= 188)Yes (n= 540)PAbuser partner (n= 192)Abuser family (n= 187)Abuser known (n= 241)Abuser unknown (n= 81)P
  • *

    HCV=Health care visits.

  • **

    Fisher's exact test.

Chronic pelvic pain48 (12.7)25 (10.2)17 (16.5)0.2518 (10.5)72 (14.0)0.0485 (12.5)4 (10.3)1.00**20 (10.6)70 (13.0)0.4228 (14.6)21 (11.2)30 (12.5)8 (9.9)0.66
Poor self-estimated health87 (23.0)61 (24.8)50 (48.5)<0.00145 (26.3)153 (29.7)0.02182 (26.8)14 (35.9)0.2226 (13.8)172 (31.9)<0.00153 (27.6)65 (34.8)64 (26.6)7 (8.6)<0.001
Psychosomatic symptoms155 (40.9)108 (43.9)65 (63.1)<0.00172 (42.1)256 (49.6)<0.001301 (44.4)23 (59.0)0.0751 (27.1)277 (51.3)<0.00188 (45.8)102 (54.5)103 (42.7)24 (29.6)<0.001
High consumer of HCV*84 (22.2)64 (26.0)42 (40.8)<0.00142 (24.6)148 (28.7)0.01175 (25.7)12 (30.8)0.4830 (16.0)160 (29.6)<0.00145 (23.4)62 (33.2)59 (24.5)16 (19.8)0.06
High incidence of sick leave144 (38.0)101 (41.1)53 (51.5)0.0469 (40.4)229 (44.4)<0.001274 (40.2)19 (48.7)0.3247 (25.0)251 (46.5)<0.00183 (43.2)90 (48.1)96 (39.8)20 (24.7)0.01

As we presumed that interactions between the different health variables might be found, three logistic regressions were performed (Table 5). Women having experienced sexual abuse both as children and as adults were more likely to report poor self-estimated health (OR 2.29) and psychosomatic symptoms (OR 1.70) than women who reported abuse exclusively in childhood or as adults. Patients who reported penetrative sexual abuse experiences were more likely to have chronic pelvic pain (OR 1.80) than patients who had experienced non-penetrative abuse. A history of physical and/or emotional abuse in addition to a sexual abuse history was also highly predictive of both poor self-estimated health and of existence of psychosomatic symptoms (OR 1.96 and 2.04). Reporting the abuser as being a family member did not increase reporting of any of the three health variables when controlling for all other factors in the analyses. However, if the concept of ‘family’ was extended to ‘a known abuser’, we found that women who reported the abuser as being known to them were more likely to report poor self-estimated health (OR 3.26) than women reporting the abuser as being unknown.

Table 5.  Logistic regression analyses of subgroups of sexually abused women by selected health variables among gynaecology patients reporting a history of sexual abuse. Values given are odds ratio with 95% confidence interval, all adjusted for age, country of residence and partner status.
 Poor self-estimated health (n= 689)Psychosomatic symptoms (n=690)Chronic pelvic pain (n=692)
  1. NS = non-significant according to the 5% level.

Both child and adult sexual abuse2.29 (1.44–3.64)1.70 (1.06–2.71)NS
Penetrative sexual abuseNSNS1.80 (1.02–3.19)
History of physical/emotional abuse1.96 (1.18–3.23)2.04 (1.36–3.05)NS
Abuser within familyNSNSNS
Abuser known3.26 (1.45–7.33)NSNS

DISCUSSION

The study shows that a history of sexual abuse is associated with poor overall health in a Nordic sample of gynaecology patients.

However, our study has several limitations: the cross-sectional design provides associations and not causal relations and therefore does not allow us to determine to what extent these problems are a direct result of the abusive sexual encounter or of other circumstances in the victim's life. Neither do we know whether the symptomatology/poor health preceded or followed the abuse or whether both are a function of some third factor (e.g. dysfunctional family dynamics or substance abuse). However, as almost all of the exposed (94.7%) had been exposed before the 12 months prior to the study, it is most likely that their health status was reported after the sexual abuse incident(s).

Moreover, the women were recruited from a clinical population and the results might have been different if the same study had been conducted in a general population of women.

Women with poor health might seek causality and therefore try to find an explanation in their past. This recall bias could lead to overreporting of sexual abuse history, and in turn, skews the effect of sexual abuse history on health.

Why should a sexual abuse history affect health? Based on the stress–illness theory,18 victimisation might cause ill health by several pathways: emotional responses might be perceived as physical disease, pre-existing symptoms could be exacerbated or the tolerance for them lowered and the immune system may be suppressed due to stress. Several diagnoses are more frequently found among sexually abused patients than among non-abused patients, such as chronic pelvic pain, premenstrual syndrome, dysmenorrhoea, pelvic inflammatory disease, sexual dysfunction, irritable bowel syndrome, uncharacteristic abdominal pain, fibromyalgia, chronic fatigue and depression.2,7,10,12,14,19–22 An association between sexual abuse and increased utilisation of health care services has also been found.18,20,23 Many of these studies are based on small sample sizes. Only few have related the severity or circumstances of the abuse to other health outcomes.

We found a strong correlation between chronic pelvic pain as reason for the index visit and a history of sexual abuse. It has been suggested that chronic pelvic pain reflects a psychological repetition of the sexual trauma.8 Chronic pelvic pain has also been linked to physical abuse,19,24,25 which suggests that chronic pelvic pain is a symptom or an effect of traumatic abuse. It might be that abuse in general promotes chronicity of painful conditions.

Chronic pelvic pain may explain our finding of the higher frequency of laparoscopic procedures among the sexually abused women. It is estimated in the literature that chronic pelvic pain accounts for up to 10% of all gynaecological consultations, which agrees with the 8% found in our study and approximately 40% of all laparoscopies performed.5,24

We found a strong correlation between sexual abuse history and psychosomatic symptoms and poor self-estimated health. This finding suggests that the medical problems associated with sexual abuse are of a vague non-specific and multiorgan character. This is in agreement with some previous studies.1,3,12,13,20,26 Many of the proposed diagnoses related to sexual abuse can be thought of as functional disorders and could all be regarded as integrated in the term somatisation, as discussed by Berkowitz.27

We studied certain subgroups among the sexually abused women to learn whether the specific circumstances of the abuse could predict poor health. It transpired that women abused both as children and as adults reported significantly poorer self-estimated health than women abused solely as children or as adults. They also reported psychosomatic symptoms to such an extent that it interfered with their daily activities, had more health care visits and had more often been on sick leave, which supports earlier studies.3,18,21 This might reflect that revictimisation increases the risk of poor overall health.10,21

About two-thirds of women with a history of sexual abuse also reported having been physically and/or emotionally abused. A similar result has been found by Leserman et al.10 We found that when women had experienced additional types of abuse, it was more likely that they would report poor overall health, which is in accordance with previous studies.10,14,19,24,28

Some studies suggest that sexual abuse involving penetration is associated to worse somatic and psychological health than non-penetrative abuse.10,12,28 We could not confirm this finding. Women who had experienced penetrative abuse were more likely to consult the clinics because of chronic pelvic pain than women who had experienced non-penetrative abuse.

The relationship between abuser and victim is possibly also important when discussing the association between a sexual abuse history and poor health. Being abused by a family member, or someone known to the victim, is a serious breach of trust and may be associated with a greater tendency to feel shame and guilt over what happened, and in turn, more severe distress, than being abused by a stranger. In our material, there was a strong correlation between having been sexually abused by a known person (family, friend, ex-partner) and poor self-estimated health. The literature is inconclusive on this subject and there are contradictory reports.13,14,28

CONCLUSION

This study underlines the importance of eliciting a history of sexual abuse in women with unexplained chronic pelvic pain, and if a history is present, relevant treatment should be offered before surgical intervention. In addition, the presence of a sexual abuse history should be considered in women with multiple complaints. As only few patients spontaneously report prior experiences of abuse,15,29 the physician must ask the patient directly.

This study supports the notion that victims of sexual abuse are at risk of developing serious health problems over time. Chronic pelvic pain is the most important symptom in gynaecology, but symptoms of a vague and diffuse nature are frequent. It appears that some subgroups of sexually abused women are more likely to report poor health: women abused both as children and as adults, women who have also been physically and/or emotionally abused and women abused by someone known to them.

We hope that focussing on the health issues related to sexual abuse will improve the diagnosis, prediction, prevention and treatment of the consequences of sexual abuse.

Acknowledgements

The authors would like to thank the following organisations for their financial contributions: K. Vinders Foundation, Organon Gynaecological Research Grant 1999 and The Augustinus Foundation (Denmark); The Jahnsson Foundation, The Gyllenberg Foundation and The Helsinki University Central Hospital (Finland); The Department of Obstetrics and Gynaecology and The Research Fund, Landspitalinn-University Hospital, Reykjavik (Iceland); Norwegian Women Public Health Association (Norway); The Health Research Council in the Southeast of Sweden (Sweden).

Funding

The study was organised and carried through by members in NorVold, a research network supported by grants from the Nordic Minister Council.

Accepted 3 April 2004

Ancillary