GENDER MEDICINE/VOL. 8,NO. 2, 2011 NorVold Abuse Questionnaire for Men (m-noraq): Validation of New Measures of Emotional, Physical, and Sexual Abuse and Abuse in Health Care in Male Patients Katarina Swahnberg, PhD Department of Clinical and Experimental Medicine, Division of Gender and Medicine, Faculty of Health Sciences, Linköping University, Linköping, Sweden ABSTRACT Background: There are far more prevalence studies on abuse of females than on males as subjects of abuse. The NorVold Abuse Questionnaire (NorAQ) measures emotional, physical, and sexual abuse, as well as abuse in health care, in women and men. Objective: The aim of this study was to test the concurrent validity and test-retest reliability of the questions in the version of NorAQ administered to men (m-noraq) against the interview model. Methods: The validation was tested in a subsample (n 86) of a male patient sample who had filled out the m-noraq (N 1667). Respondents completed m-noraq twice and were then interviewed. Before the interview, respondents were instructed to answer questions based on personal experience. The interview consisted of 4 open-ended questions about lifetime experiences of emotional, physical, and sexual abuse and abuse in health care. Results: Results indicated that respondents in the subsample had discussed the experiences of abuse more often in both formal and informal settings than had subjects in the total patient sample. Measures of sensitivity for m-noraq were good to excellent (emotional abuse, 83%; physical abuse, 76%; sexual abuse, 68%; abuse in health care, 93%), as were those for specificity (emotional abuse, 72%; physical abuse, 92%; sexual abuse, 99%, abuse in health care, 90%); likelihood ratios were satisfactory (emotional abuse, 3; physical abuse, 9; sexual abuse, 46; abuse in health care, 9); and test-retest reliability measures were excellent (emotional abuse, 80% 95%; physical abuse, 77% 88%; sexual abuse, 91% 100%; abuse in health care, 84% 92%). Conclusion: m-noraq showed good to excellent concurrent validity for the different types of abuse and excellent reliability for all questions about abuse. In spite of methodological challenges, validation studies must be conducted as a minimum precaution to ensure that an instrument accurately measures abuse as intended. (Gend Med. 2011;8:69 79) 2011 Elsevier HS Journals, Inc. All rights reserved. Key words: abuse in health care, emotional abuse, NorAQ, NorVold Abuse Questionnaire, physical abuse, reliability, sexual abuse, validity. Accepted for publication February 28, 2011. doi:10.1016/j.genm.2011.03.001 2011 Elsevier HS Journals, Inc. All rights reserved. 1550-8579/$ - see front matter 69
Gender Medicine INTRODUCTION Validation ensures that an instrument accurately measures the phenomenon under study, as intended. However, few instruments measuring abuse in either male or female subjects have been validated. 1,2 The NorVold Abuse Questionnaire (NorAQ) a measure of emotional abuse, physical abuse, sexual abuse, and abuse in health care was developed in 1999 to measure prevalence rates of abuse in gynecologic patients in 5 European countries 2 4 (Appendix I). Since then, NorAQ has been used in independent female and male groups, 5 13 and its concurrent validity was tested in a female population in Sweden with good results (likelihood ratio: emotional abuse 38; physical abuse 6; sexual abuse 42; abuse in health care 43). 2 NorAQ was developed for research purposes, but it can also be used as a screening tool. Compared with other screening tools, NorAQ provides several examples of each type of abuse it is designed to measure. There are no perpetrator restrictions on how questions on abuse are phrased. Abuse in health care is not usually included in studies of abuse. Childhood emotional, physical, and sexual abuse have been recognized as risk factors for revictimization of adults in health care systems (Edholm S, Fredman K, Wijma B, Swahnberg K. Men s perceived experiences of abuse in health care: their relationship to childhood abuse. Submitted for publication). 12 The high prevalence of abuse in health care reported in previous studies on revictimization in health care motivated this investigation of questionnaire items about all 4 types of abuse. Because male victimization may be underreported, there is the potential for bias in validation studies. Therefore, one aim of this study was to examine how men communicate their own experiences as subjects of abuse. However, the primary aim was to investigate the concurrent validity and test-retest reliability of the questions about abuse in m-noraq, compared with the interview model used when the original version of NorAQ was validated in a female sample, and to discuss results from the 2 studies from a gender perspective. METHODS Measurement To create m-noraq, changes to questions about sexual abuse were made in the version of NorAQ administered to women, including the omission of 16 questions about reproductive health, 1 question about hospitalization, and the word vagina. Four new questions about ethnicity, income, and parental education level were added. m-noraq was first tested at Statistics Sweden (SCB) pretesting laboratory 13 on 6 male volunteers, aged 20 to 72 years. All volunteers read information letters and completed the m-noraq and were then interviewed. The aim was to measure how respondents comprehended and reasoned about the information letters and the questions in m-noraq. In response to results from the SCB test, information letters were shortened and simplified and structural and layout changes were made, including rephrasing and omitting some questions and clarifying the instructions in m-noraq. To determine whether male patients would likely answer the questions, m-noraq was tested further, using 24 subjects in a pilot study at an infectious diseases clinic. Questionnaires returned from participants in the pilot study were later incorporated into the full-scale study. The m-noraq is divided into 7 sections and consists of 67 questions. In addition to abuse, m- NorAQ addresses sociodemography and self-reported health status and medical history. Four identically structured sections covered experiences of emotional, physical, and sexual abuse and abuse in health care. Experiences of each type of abuse were defined based on 1 or more positive answers to the questions that specified the type of abuse. The questions included specific examples of abusive behavior (Appendix II), ranging from mild to severe, that occurred in childhood, adulthood, or both and allowed for an approximate classification according to degree of severity of the abusive act. If the participant reported several degrees of a specific type of abuse, he was classified according to the most severe. If the participant had experienced a specified type of abuse, he was instructed to answer more detailed questions about 70
Katarina Swahnberg the event, such as identity of the perpetrator, whether he had disclosed the information, and whether he had ever reported the abuse to police. Furthermore, the respondent was asked to estimate, separately for each type of abuse, how severely he was affected by the experience, based on an 11-point scale. The respondent was also asked whether he had ever sought any type of assistance as a result of the abuse. Material and Procedures in the Total Sample The Regional Ethical Review Board, Linköping, Sweden, approved the study, and recruitment continued from August 2004 to March 2005. Eligible subjects included 2363 consecutive inpatient and outpatient men, aged 18 or older, who understood Swedish and who were at the University Hospital in Linköping, Sweden. Six clinics participated in the study including the Center for Orthopedics, the Center for Surgery and Oncology, the Center for Reconstructive Medicine, the Dermatological and Venereology Clinic, the Heart Center, and the Infectious Diseases Clinic. On arrival, patients received an information letter about the study from the staff. One to 2 weeks later, m-noraq and another explanatory letter were mailed to each person who had not actively declined to participate in the study when he received the first explanatory letter, followed by 2 reminders. A total of 1767 (75%, N total sample) answered m-noraq. Material and Procedures in the Validation Sample Of the 1767 respondents who completed m-noraq, 3 did not answer questions about abuse, leaving 1764 eligible respondents. Of these, 876 (49.7%) reported experiencing at least 1 type of abuse and 888 (50.3%) reported no abuse. Respondents in each of the 2 groups (those with and without experiences of abuse) were randomly selected for the study by an assistant who was not otherwise involved in the research. Among the men selected, those who lived in or near Linköping, Sweden, were invited by telephone to participate in the validation session. Approximately every second man agreed to participate (n 93). Of these 93, 7 men (6 of whom had reported abuse) did not report for follow up, leaving 86 men (74 of whom had reported 1 or more types of abuse) in the study. Participants were not compensated in any way. Data were collected from answers to the questionnaire (m-noraq [I]) completed at home, answers to the questionnaire (m-noraq [II]) completed at Linköping University Hospital, and results from the interviews conducted at Linköping University Hospital. Test-retest reliability for all 67 questions in the m-noraq was estimated by comparison between answers given in m-noraq I and m-noraq II. On average, 8 weeks (mean) had passed between completion of m-noraq I and completion of m-noraq II (maximum 8 months). No participant reported experiencing abuse during the period between completing m-noraq I and m-noraq II. Validity was tested only for the 13 questions about experiencing abuse. m-noraq II was considered the screening test, and the interview as the gold standard, that is, the benchmark or diagnostic test categorizing participants into groups with or without experiences of emotional, physical, or sexual abuse or abuse in health care. 14 The author, who conducted all interviews, was blinded to the participant s answers to questions in m-noraq I and m-noraq II. In general, interviews with respondents who reported no abuse lasted a minimum of 20 minutes. The more types of abusive experiences reported, the longer the interview lasted; the maximum interview time was 90 minutes. To study the effect of the number of observations in relation to concurrent validity, the test characteristics of m-noraq were estimated at 3 points during data collection. The validation sample was sorted at random before computing analyses with the first 2 frequencies of participants (n 26 and n 56). Interview The interview procedure was standardized according to a protocol developed for the validation of NorAQ in a female population and used for selecting potential participants, contacting potential participants, and determining the order ques- 71
Gender Medicine tionnaires should be distributed in relation to the interview. Test interviews were conducted before the procedure was finally established. The interview was conducted according to an interview guide developed for the validation of NorAQ in a female sample. 2 Before each interview, respondents were instructed to answer questions based on personal experience. The interview included 4 open-ended questions about lifetime experiences of emotional, physical, and sexual abuse and abuse in health care, respectively. The respondent was asked, Have you ever experienced emotional abuse? If the respondent answered yes, more detailed questions followed about when the experience took place, how frequently, the perpetrator(s), and impact of the abuse experience. All 4 types of abuse were approached in the same manner. Statistics and Measures Analyses were computed using SPSS 12.0.1 for Windows. Demographic differences between the total sample and the validation samples were evaluated using the Mann-Whitney U test. Differences between the total sample and the validation samples about reporting/talking about abuse were tested using the Pearson 2 test. Significance was P 0.05 (95% confidence interval). Concurrent validity was assessed using sensitivity, specificity, positive and negative predictive value, and positive likelihood ratio performance measures. 14,15 Reliability was measured with test-retest reliability and Cohen s kappa test of concordance. Kappa was calculated according to the kappa measure of agreement for an R R table. 16 RESULTS Sample Sociodemographic characteristics of the total sample (n 1681) and the validation sample (n 86) are shown in Table I. In the total sample, 96.5% of respondents were born in Sweden, 79.9% lived with a partner, and 68.6% had 12 or fewer years of education (mean age, 59.5; range, 18 91). The validation sample consisted of men with more Table I. Background characteristics in the validation sample compared with the total sample (% of men in each sample). Total Sample n% n 1681 Validation Sample n% n 86 Age (y) P 0.194 18 34 172 (10.3) 8 (9.3) 35 49 233 (13.9) 19 (22.1) 50 1270 (75.8) 59 (68.6) Education (y) P 0.002 9 710 (42.5) 18 (20.9) 10 12 436 (26.1) 35 (40.7) 13 525 (31.4) 33 (38.4) Civil status P 0.331 Single 336 (20.1) 21 (24.4) Partner 1336 (79.9) 65 (75.6) Native country P 0.249 Sweden 1610 (96.5) 80 (94.1) Other Nordic country 18 (1.1) 2 (2.4) Other 40 (2.4) 3 (3.5) Income P 0.034 149,000 362 (22.0) 10 (11.8) 150,000 449,000 1154 (70.3) 67 (78.8) 450,000 126 (7.7) 8 (9.4) Parents years in school Mother P 0.844 9 1368 (83.6) 70 (82.4) 10 12 178 (10.9) 12 (14.1) 13 91 (5.6) 3 (3.5) Father P 0.374 9 1315 (80.1) 71 (84.5) 10 12 194 (11.8) 6 (7.1) 13 132 (8.0) 7 (8.3) Occupation P 0.045 Employed 764 (45.6) 48 (55.8) Unemployed 29 (1.7) 4 (4.7) Student 57 (3.4) 1 (1.2) Parental leave 3 (0.2) 1 (1,2) Sick lv./soc. w./retir. 803 (47.9) 32 (37.2) Other 20 (1.2) 0 (0) retir. retired (temporary disability pension, disability pension); sick lv. on sick leave over a long period; soc. w. recipient of social welfare. All data from m-noraq I. Internal drop out ranged from n 5 44 (0.3 2.6) in the total sample, and n 0 2 (0.0% 2.3%) in the validation sample. years of education (P 0.01) and a higher income (P 0.05), and fewer were retired, on sick leave, or receiving social support (P 0.045) than in the total sample. All questions about abuse were an- 72
Katarina Swahnberg Table II. Exploration of concurrent validity of the abuse questions in m-noraq with an interview as gold standard. Performance of m-noraq in relation to number of participants when n 26, n 56, n 86. Validation Concepts Emotional Abuse Physical Abuse Sexual Abuse Abuse in Health Care n 26 Sensitivity 100 71.4 66.7 50.0 Specificity 75.0 91.7 95.7 87.5 Pretest probability (prevalence) 38.5 53.8 11.5 7.7 PV 71.4 90.9 66.7 25.0 PV 100 73.3 95.7 95.5 LR 4 9 16 4 n 56 Sensitivity 95.5 75.7 71.4 94.1 Specificity 73.5 89.5 97.6 89.7 Pretest probability (prevalence) 39.3 66.1 25.0 30.4 PV 70.0 93.3 90.9 80.0 PV 96.2 65.4 91.1 97.2 LR 4.0 7.0 30.0 9.0 n 86 Sensitivity 83.3 75.8 68.4 93.1 Specificity 72.0 91.7 98.5 89.5 Pretest probability (prevalence) 41.9 72.1 22.1 33.7 PV 68.2 95.9 92.9 81.8 PV 85.7 59.5 91.7 96.2 LR 3.0 9.0 46.0 9.0 LR positive likelihood ratio PV predictive value of a positive test; PV predictive value of a negative test. swered satisfactorily; internal dropout ranged from 0.9% to 2.7% for the 4 types of abuse. Concurrent Validity Table II presents m-noraq s ability to capture true positive and negative answers in relation to the numbers of participants. Of note is the relationship among prevalence, sensitivity, and positive predictive values. The lowest sensitivity and positive predictive values were estimated at the lowest prevalence: abuse in health care, 8%; sensitivity, 50%; and positive predictive value, 25% (n 26). At the next level of observations (n 56), where prevalence of abuse in health care was 30%, the sensitivity and positive predictive values rose to 94% and 80%, respectively. At the final level of observations (n 86), small differences in prevalence did not seem to affect performance measures (Table II). The ability of m-noraq to separate true positive answers (sensitivity) was most accurate for abuse in health care and emotional abuse, and the ability to separate true negative answers (specificity) was most accurate for physical abuse and sexual abuse (Table II). Test-Retest Reliability Test-retest reliability for emotional abuse was 80% to 95%; physical abuse, 77% to 88%; sexual abuse, 91% to 100%; and abuse in health care, 84% to 92%. Highest test-retest reliability was measured for severe abuse in each of the 4 types of abuse. Table III presents kappa values that confirm the high agreement for emotional abuse, sexual abuse, and the somewhat lower agreement for physical abuse. Despite the low value for severe abuse in health care (0.18), 77 of the 86 respondents answered the question identically (negatively) in NorAQ I and NorAQ II. Similarly, mild sexual abuse (touch) showed a low kappa value (0.41), and still 73 of the 86 respondents answered the question identically (negatively) in NorAQ I NorAQ and II. 73
Gender Medicine Table III. Test-retest reliability of the 13 abuse variables in m-noraq according to the kappa measure of agreement for an R R table (n 86). Emotional Abuse Physical Abuse Sexual Abuse Abuse in Health Care Mild 0.67 0.55 0.41 (touch) 0.65 0.85 (other) Moderate 0.52 0.67 0.95 0.63 Severe 0.79 0.52 1.00 0.18 Excellent agreement 17 0.75 1.0; good agreement 0.60 0.74; fair agreement 0.40 0.59; poor agreement 0.40. Prevalence (NorAQ I) Table IV presents experiences of abuse according to severity in the total sample. Physical abuse was the most commonly reported type of abuse. Mild or moderate severity was the most commonly reported degree of abuse for all types of abuse. Six men (0.3%) reported severe sexual abuse (rape), and 113 (6.5%) reported severe physical abuse (death threat). Most men (n 585, or 34%) reported only 1 type of abuse. Communication of Abusive Experiences The percentage of respondents who had ever disclosed the experience of abuse varied according to type of abuse (emotional abuse, 81%; physical abuse, 60%; sexual abuse, 51%; abuse in health care, 85%). Few (13%) had ever reported the abuse to police, and only 3% to 19% had ever sought assistance as a result of emotional abuse (n 41; 19%), physical abuse (n 45; 6%), sexual abuse (n 2; 3%), and abuse in health care (n 11; 8%). A total of 44 men (21%) reported that they had not sought assistance despite having endured considerable distress as a result of emotional abuse. Corresponding figures for the other types of abuse were for physical abuse (n 30; 4%), sexual abuse (4; 6%), and abuse in health care (n 28; 20%). Disclosure of abusive experiences to police was more frequent in the validation sample than in the total sample; almost twice as many had ever reported the abuse to police (P 0.02). Seeking assistance as a result of sexual abuse (P 0.002) and abuse in health care (P 0.03) also occurred more frequently in the validation sample. Among the abused subjects, between 44% and 94% had disclosed the experience (emotional abuse, 82%; physical abuse, 73%; sexual abuse, 44%; abuse in health care, 94%), which was similar to rates in the total sample; that is, emotional abuse and abuse in health care were disclosed more frequently than, especially, sexual abuse. Physical abuse was the single exception; more respondents in the validation sample had disclosed these experiences than in the total sample (P 0.03). DISCUSSION Questions about abuse in m-noraq were validated against an interview as gold standard and proved good-to-excellent concurrent validity. m-noraq s ability to distinguish true positive responses was most accurate for emotional abuse (83%) and abuse in health care (93%), whereas the ability to distinguish true negative answers was most accurate for physical abuse (92%), sexual abuse (99%), and abuse in health care (90%). A different pattern was found in the validation of the female version of NorAQ: the ability to distinguish true positive answers was most accurate for physical abuse (96%), whereas the ability to distinguish true negative answers was equally accurate for emotional abuse, sexual abuse, and abuse in health care (98%). 2 Table IV. Experiences of lifetime abuse according to severity, N 1767 (% of all men). Emotional Abuse Physical Abuse Sexual Abuse Abuse in Health Care Mild 101 (5.8) 286 (16.5) 22 (1.3) 57 (3.3) Moderate 48 (2.7) 396 (22.8) 37 (2.1) 58 (3.3) Severe 76 (4.3) 113 (6.5) 6 (0.3) 27 (1.5) 74
Katarina Swahnberg Test-retest reliability for both versions of NorAQ was excellent. In the version administered to men, test-retest reliability was very high for emotional abuse, sexual abuse, and abuse in health care and somewhat less so for physical abuse, whereas in the version of NorAQ administered to women testretest reliability was highest for sexual abuse, physical abuse, and abuse in health care and lowest for emotional abuse. However, the highest test-retest reliability was found for severe abuse in both males and females except for abuse in health care (moderate abuse in health care had higher test-retest reliability in females [unpublished results]). Based on these results, NorAQ and m-noraq could be a firsthand choice when measuring emotional, physical, and sexual abuse and abuse in health care in male and female subjects. However, testretest reliability and validity for both versions of NorAQ has to be measured in each different sample. Results from this or any other validation study cannot be transferred automatically to other contexts. Limitations There was a considerable range in the test-retest period in this study, but to retain sample size, outliers were kept in the study. The mean time, after all, was 8 weeks, which must be considered typical, and a longer time span would only be to the m-noraq s disadvantage, increasing the risk for report bias. Furthermore, different outcomes have to be evaluated critically before use. Likelihood ratio is the only test measure considered independent of prevalence, 15 but likelihood ratio calculation is based on sensitivity and specificity, which are based on prevalence rates. In this study, the relationship between prevalence and test values is illustrated by presenting m-noraq s performance relative to 3 sets of participants (Table II). The most important changes occurred between the first and second number of observations. Whether the prevalence was 22% or 25% or 30% or 34% did not seem to matter, but a prevalence that rose from 8% to 30% made a difference. Bias decreased with a growing number of observations, which, to some extent, also was true for likelihood ratio; therefore, prevalence should determine the number of observations, also if the test measure is likelihood ratio. Another limitation is kappa measure of agreement. The kappa values in Table IV show 2 low outliers. In these cases, test-retest reliability was good, but the kappa value pointed out an unequal distribution of positive and negative answers, that is, the prevalence of severe abuse in health care and mild sexual abuse (touch) was too low. Dependency of distribution in the cells is a methodological weakness when calculating kappa values, but it also indicates a drawback in this study: excessively low prevalence of severe abuse in health care and mild sexual abuse. With these 2 exceptions, kappa values confirmed high agreement for emotional abuse, sexual abuse, and abuse in health care and a somewhat lower agreement for physical abuse. Kappa value and test-retest reliability, however, should be evaluated together. Gold Standard In this study, the interview was used as the gold standard. One could argue that subjects with a history of abuse would be more likely to respond affirmatively to an anonymous questionnaire on abuse rather than to a person conducting an interview. In fact, research in female subjects has shown the opposite to be true 1,17 (literature study in cover story 1 ). Although an interview could be assumed to provide the most accurate participant responses 18 and be more reliable than responses from a self-reported questionnaire, 19 one could consider the use of independent reports of abuse from other sources, such as other questionnaires, hospital records, or police reports. 20 The problem is that these sources are limited in the same manner as interviews and questionnaires regarding the subject s inability or reluctance to disclose abusive experiences. 17,21 Furthermore, when using a questionnaire as the gold standard, there is a distinction between using a definition of abuse that is too similar to or too different from that used in the instrument tested. Silence Bias The recruitment of men to interview was challenging in this study. Their reluctance to participate in a study about experiences of abuse may be 75
Gender Medicine attributable to a preference to conceal. Perhaps a male subject may find discussing abuse uncomfortable. Possibly concealment is linked to type of abuse; it might be more acceptable for a man to admit to street violence than to partner violence. Shakil et al wrote...it took 2 years to find 17 men who would admit the fact that they were victims of domestic violence. 22 The sample used for the validation was representative of the original sample background variables such as age, civil status, and native and family background, but not education, employment, and income, which are the usual indicators of socioeconomic class. Aside from this selection bias, it cannot be ruled out that more types of selection bias are found in this study, especially because it attracted respondents who were more accustomed to discussing their abusive experiences in formal and informal settings. Formal and informal disclosure likely has different meanings for the subject. Reporting to police or undergoing treatment, or both (ie, formal disclosure), could be strong confirmations that corroborate the subject s experience of being abused compared with those of informal disclosure. But how is a questionnaire viewed in this sense by subjects? Formal or informal? Safe or unsafe? Confirming or questioning experiences of abuse? Respondents in the validation sample possibly could be expected to be more experienced in reporting or disclosing their experiences of abuse because they had sought assistance more often than did subjects in the total sample. It is also possible that men in the validation sample experienced more severe abuse because they sought assistance more often and may be easier to identify true positive answers in a survey. This selection bias might have lead to overestimation of the capacity of the m-noraq to differentiate victims from nonvictims among the subjects. Thus, the result is primarily valid for Swedish male patients who have experience discussing abuse. This study cannot determine how well m- NorAQ would perform in male Swedish patients who have not discussed experiences of abuse previously. This kind of potential selection bias, or silence bias, is seldom addressed in the literature. CONCLUSION m-noraq is a sound measure in this study, showing good-to-excellent concurrent validity for the different types of abuse and excellent reliability for all questions about abuse. Despite the methodological difficulties in this study, validation studies must be considered a minimum precaution to ensure that the instrument accurately measures abuse as intended. For decades, prevalence studies in female samples have indicated an urgent need for interventions against abuse. Refined methodology, including validated instruments, will be needed to evaluate future interventions against abuse in male and female samples along with other outcome measures. Validated instruments are required for determining whether interventions are effective. Continued methodological discussions and development are therefore needed. ACKNOWLEDGMENTS The NorVold Abuse Questionnaire (NorAQ) was developed by NorVold, a research network established in 1997, to explore the prevalence of violence against women and its effects on women s health. The NorVold research network was supported by grants from the Nordic Minister Council. Principal investigators were Barbro Wijma and Berit Schei, and the coordinator was Katarina Swahnberg. Local investigators were as follows: in Denmark: Katrine Sidenius and Malene Hilden; in Finland: Erja Halmesmäki and Ulla Pikkarinen; in Iceland: Tora Steingrimsdottir; in Norway: Berit Schei, Hildegunn Stoum- Hinsverk, and Kristin Offerdal; and in Sweden: Barbro Wijma and Katarina Swahnberg. Financial support was provided by the Swedish Research Council and the Health Research Council in the Southeast of Sweden. There are no conflicts of interest to report. REFERENCES 1. Swahnberg K. Prevalence of Gender Violence. Studies of Four Kinds of Abuse in Five Nordic Countries [dissertation]. Linköping, Sweden: Faculty of Health Sciences, Linköping University; 2003. 76
Katarina Swahnberg 2. Swahnberg IM, Wijma B. The NorVold Abuse Questionnaire (NorAQ): validation of new measures of emotional, physical, and sexual abuse, and abuse in the health care system among women. Eur J Public Health. 2003;13:361 366. 3. Wijma B, Schei B, Swahnberg K. NorAQ. The Nor- Vold Abuse Questionnaire. Linköping: Division of Gender and Medicine, Faculty of Health Sciences, Linköping University. Report No 2; 2004. 4. Wijma B, Schei B, Swahnberg K, et al. Emotional, physical, and sexual abuse in patients visiting gynaecology clinics: a Nordic cross-sectional study. Lancet. 2003;361:2107 2113. 5. Hilden M, Schei B, Swahnberg K, et al. A history of sexual abuse and health: a Nordic multicentre study. BJOG. 2004;111:1121 1127. 6. Hilden M, Sidenius K, Langhoff-Roos J, et al. Women s experiences of the gynecologic examination: Factors associated with discomfort. Acta Obstet Gynecol Scand. 2003;82:1030 1036. 7. Pikarinen U, Saisto T, Schei B, et al. Experiences of physical and sexual abuse and their implications for current health. Obstet Gynecol. 2007;109:1116 1122. 8. Swahnberg K, Hearn J, Wijma B. Prevalence of perceived experiences of emotional, physical, sexual, and health care abuse in a Swedish male patient sample. Violence Vict. 2009;24:265 279. 9. Swahnberg K, Schei B, Hilden M, et al. Patients experiences of abuse in health care: a Nordic study on prevalence and associated factors in gynecological patients. Acta Obstet Gynecol Scand. 2007;86:349 356. 10. Swahnberg K, Wijma B, Liss PE. Female patients report on health care staff s disobedience of ethical principles. Acta Obstet Gynecol Scand. 2006; 85:830 836. 11. Swahnberg K, Wijma B, Schei B, et al. Are sociodemographic and regional and sample factors associated with prevalence of abuse? Acta Obstet Gynecol Scand. 2004;83:276 288. 12. Swahnberg K, Wijma B, Wingren G, et al. Women s perceived experiences of abuse in the health care system: their relationship to childhood abuse. BJOG. 2004;111:1429 1436. 13. Bergman LR. Pretesting procedures at statistics Sweden s measurement, evaluation and development laboratory. J Off Stat. 1995;11:309 23. 14. Fletcher RH, Fletcher SW, Wagner EH. Clinical Epidemiology: The Essentials. 3rd ed. Baltimore, Md: Williams & Wilkins; 1996. 15. Bowling A. Measuring Health. A Review of Quality of Life Measurement Scales. 2nd ed. Buckingham, England: Open University Press; 1997. 16. Bowling A. Measuring Disease. 2nd ed. Buckingham, England: Open University Press; 2001. 17. Ellsberg M, Heise L, Pena R, et al. Researching domestic violence against women: methodological and ethical considerations. Stud Fam Plann. 2001;32:1 16. 18. Leserman J, Drossman DA, Li Z. The reliability and validity of a sexual and physical abuse history questionnaire in female patients with gastrointestinal disorders. Behav Med. 1995;21:141 150. 19. Wyatt GE, Peters SD. Methodological considerations in research on the prevalence of child sexual abuse. Child Abuse Negl. 1986;10:241 251. 20. Kooiman CG, Ouwehand AW, ter Kuile MM. The Sexual and Physical Abuse Questionnaire (SPAQ). A screening instrument for adults to assess past and current experiences of abuse. Child Abuse Negl. 2002;26:939 953. 21. Langeland W, Draijer N, van den Brink W. Assessment of lifetime physical and sexual abuse in treated alcoholics. Validity of the Addiction Severity Index. Addict Behav. 2003;28:871 881. 22. Shakil A, Donald S, Sinacore JM, Krepcho M. Validation of the HITS domestic violence screening tool with males. Fam Med. 2005;37:193 198. (continued on next page) Address correspondence to: Katarina Swahnberg, PhD, Division of Gender and Medicine, Department of Clinical and Experimental Medicine, Faculty of Health Sciences, Linköping University, S-581 83 Linköping, Sweden. E-mail: katarina.swahnberg@liu.se 77
Gender Medicine Appendix I. Development of the NorVold Abuse Questionnaire (NorAQ). The first (female) version of NorAQ was developed by Professor Barbro Wijma, Linköping University, and Professor Berit Schei, University of Trondheim, between 1997 and 1999. The development of the new questionnaire was based on reviews of old instruments and theories and professional experiences of working with victims of abuse. Aspects of posttraumatic stress disorder diagnostics and revictimization served as theoretical landmarks. Revictimization by health professionals of patients who were former victims of abuse motivated the inclusion of questions about abuse in health care. Questions were formulated according to the existing knowledge on how to help respondents remember abusive experiences, that is, by asking more than 1 question or by giving several discrete examples. Unlike the questionnaires that ask respondents to judge whether the episode was abusive, NorAQ defined for the respondent what was considered an abusive experience. (continued) 78
Katarina Swahnberg Appendix II. Questions about abuse in the NorVold Abuse Questionnaire for men (m-noraq). 2,3 Mild abuse Mod. abuse Severe abuse Mild abuse Mod. abuse Severe abuse Mild abuse, no gen. cont. Mild abuse, emot. / sexual humiliation Mod. abuse, gen. cont. Severe abuse, penetration Mild abuse Mod. abuse Severe abuse EMOTIONAL ABUSE Have you experienced anyone systematically and for any period trying to repress, degrade, or humiliate you? Have you experienced anyone systematically and by threat or force trying to limit your contacts with others or control what you may and may not do? Have you experienced living in fear because someone systematically and for a long period has threatened you or someone close to you? PHYSICAL ABUSE Have you experienced anyone hitting you, smacking your face, or holding you firmly against your will? Have you experienced anyone hitting you with his/her fist(s) or with a hard object, kicking you, pushing you violently, beating you, thrashing you, or doing anything similar to you? Have you experienced anyone threatening your life by, for example, trying to strangle you, showing a weapon or knife, or by any other similar act? SEXUAL ABUSE Has anyone against your will touched parts of your body other than the genitals in a sexual way or forced you to touch other parts of his or her body in a sexual way? Have you in any other way been sexually humiliated; eg, by being forced to watch a pornographic movie or something similar against your will, participate in a pornographic movie or something similar, show your body naked, or watch when someone else showed his/her body naked? Has anyone against your will touched your genitals, used your body to satisfy him/herself sexually or forced you to touch anyone else s genitals? Has anyone against your will tried to or put his penis into your mouth or rectum; tried to or put an object or other part of the body into your mouth or rectum? ABUSE IN HEALTH CARE Have you ever felt offended or grossly degraded while accessing health services, felt blackmailed, or felt that someone did not show respect for your opinion in such a way that you were later disturbed by or suffered from the experience? While visiting health services, have you ever experienced a normal event suddenly become a really terrible and insulting experience, without your fully knowing how this could have happened? Have you experienced anyone in health services purposely as you understood hurting you physically or mentally, grossly violating you, or using your body and your subordinated position to your disadvantage for his/ her own satisfaction? ANSWER ALTERNATIVES (THE SAME FOR ALL QUESTIONS) 1 No 2 Yes, as a child ( 18 years) 3 Yes, as an adult ( 18 years) 4 Yes, as a child and as an adult emot., emotional; gen. cont., genital contact; mod., moderate. 79
本文献由 学霸图书馆 - 文献云下载 收集自网络, 仅供学习交流使用 学霸图书馆 (www.xuebalib.com) 是一个 整合众多图书馆数据库资源, 提供一站式文献检索和下载服务 的 24 小时在线不限 IP 图书馆 图书馆致力于便利 促进学习与科研, 提供最强文献下载服务 图书馆导航 : 图书馆首页文献云下载图书馆入口外文数据库大全疑难文献辅助工具