 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
PEDIATRICS Vol. 104 No. 2 August 1999, pp. 249-257
,
From the * Division of Developmental and Behavioral Pediatrics
and the Department of Pediatrics, Boston University School of
Medicine, Boston Medical Center, Boston, Massachusetts.
 |
ABSTRACT |
|---|
Top
Abstract
Methods
Results
Discussion
Conclusion
References
|
|---|
Objective. Women with histories of interpersonal violence are poorly identified because of barriers in self-disclosure. This study identified differences on maternal health and child behavior between women who report filing a restraining order (RO) and those who do not among a nonreferred sample of women living in high-crime neighborhoods.
Methods. During a maternal interview mothers were asked whether they ever filed a RO, the victim/defendant relationship, the number of times, and the year of the filing. Four types of violence were coded independently based on maternal narratives: verbal harassment, verbal threats or intimidation, physical assault, and destruction of property. We controlled for differences between RO and non-RO groups regarding demographic background, partner characteristics, other types of maternal past victimization, and use of alcohol and illicit drugs. Outcomes for mothers include partner aggression (Conflict Tactics Scale-R), perception of health and bodily pain (SF-36 Health Survey), distress symptoms (SCL90-R), posttraumatic stress (PTS)-related symptoms, and partial posttraumatic stress disorder (PTSD) diagnosis (modified Diagnostic Interview Schedule PTSD-Module). Outcomes for the child include partner aggression (Conflict Tactic Scale-R), behavior problems (CBC 2-3 or Child Behavior Checklist), and PTS-related symptoms (PTS-related symptom checklist).
Patients. One hundred sixty patients between 3.0 to 6.1 years who resided within five residential ZIP codes with a high rate of local crime in the City of Boston were drawn from a pediatric care clinic practice. Patients were participants of a larger study about the impact of community violence on mother-child relations.
Results. Sixty-four (40%) of 160 mothers reported a history of filing a RO against a current boyfriend or husband (39%), ex-boyfriend or husband (44%), someone known (8%), or other (9%), with a mean of 3.9 years (standard deviation = 3.5 years) since RO filing. After controlling for covariates of marital status, immigrant status, public assistance, and lifetime sexual victimization, we found a significant multivariate analysis of covariance group effect on maternal outcomes. Analysis of covariance analyses indicated that mothers in the RO group experienced higher current partner verbal aggression and physical violence to mother, poorer health, and higher PTS-related symptoms, compared with mothers in the non-RO group. More mothers in the RO group met partial lifetime PTSD diagnosis. Unadjusted for maternal covariates, the multivariate analysis of variance analyses on child outcomes (partner aggression to child, behavior problems, and PTS-related symptoms) indicated a nonsignificant group effect.
Conclusions. Among dyads residing in high-risk crime areas, the incidence of RO histories is substantive considering this was a nonshelter, nonreferred sample. The inquiry about the history of a RO may provide a new and efficient marker to quality of current partner relationship, maternal health, and maternal stress-related symptomatology. Key words: interpersonal violence, restraining orders, maternal health, child problems, high-crime neighborhoods.
Interpersonal violence, generally defined as threats or
acts of physical attack or severe emotional abuse inflicted by intimate or known adults, has become a recognized public health problem with
serious implications for the health of women1-5 and for
the well-being of children.6-10 At high risk may be women
of child bearing years, and those who reside in low-income
neighborhoods.11,12
Despite the increasing awareness of the detrimental effects of
interpersonal violence on women's and children's health, the identification of women in violent relationships remains difficult because of an array of self-reporting issues such as: personal shame,
psychologic denial, cost-benefit self-assessment, fear of retaliation,
or dependency on the abuser.13,14 This article addresses
the self-reported use of past restraining orders (ROs) as a method to
assess maternal risk for current interpersonal violence.
Under the Massachusetts (where the study took place) Domestic Violence
statute (Chapter 209A), one legal recourse available to women who
identify themselves as victims of interpersonal violence or harassment
from an intimate or known others is to obtain a RO from a court. ROs
are granted in cases in which there has been physical abuse (attempted
or actual), threats of violence, or substantial emotional distress
because of the defendant's behavior. ROs are not issued in lieu of an
appropriate law enforcement response in cases of crimes against women,
such as an arrest. Instead, RO is an additional civil measure aimed at
protecting the women's future safety and that of her children. A RO
can order one person to stay away from another, move out and not return
to a given address, take personal clothing, or stay a specified number
of yards away from places defined in the RO. Previous literature suggests that a RO may be an effective deterrent of abuse in cases of
women with less serious histories of family violence,15 when the perpetrator had no criminal record,16 and in the
absence of perpetrators' substance abuse history.17
A database gathered by the Massachusetts Registry of Civil ROs for the
years 1992 to 1995 indicated that >145 000 ROs were issued against
108 000 defendants. A study using a sample of 1000 RO cases from this
database indicated that 86% of the defendants were males, 82% were
husbands, ex-husbands, boyfriends, or ex-boyfriends; 65% of victims
report previous episodes of violence; and although violating a RO is
considered a criminal offense in Massachusetts, 17% of RO are
violated.16
The incidence and the correlates related to filing ROs among
nonreferred samples are primarily unknown. Past literature identified several psychosocial correlates of ROs among battered women recruited from shelters or courtrooms. In a study of 89 women who filed a RO from
a total of 270 battered women,15 the RO group was younger,
more educated, employed, were involved in shorter, less violent
marriages, and had a history of previous separations, as compared with
battered women who did not file a RO. Women seemed less likely to
pursue a RO if they were more dependent on the abuser and if they
experienced more severe abuse.18 The relationship between
filing RO and the severity of abuse is not consistent across
studies.19 The presence of children in the home was also
positively associated with filing a RO.17 In a study of
287 women who sought a court RO, 50% believed that the abuse was
beginning to affect their children or that the batterer was assaulting
the children.20
It is likely that past victimization predicts current victimization
because interpersonal violence involves enduring patterns of behaving,
rather than isolated interpersonal events.13 Because of
the association between past and current victimization, we hypothesized
that victimized women (ie, with positive histories of RO) may be more
likely than nonvictimized women (ie, with negative histories of RO), to
experience current victimization. The purpose of the study was to test
this hypothesis by comparing current health, partner-to-mother
aggression, and psychologic distress, particularly stress-related
symptoms, in mothers; and partner-to-child aggression, behavior
problems, and stress-related symptoms in children of mothers who filed
RO compared with those who did not.
The sample of 160 dyads were participants of a larger research
project addressing the impact of community violence on mother-child relations. The participants in the large study were selected from patients who received pediatric primary care at a large urban teaching
hospital during 1996 to 1998 and who met eligibility criteria involving
a two-step eligibility process (see Fig
1).
METHODS
Top
Abstract
Methods
Results
Discussion
Conclusion
References
View larger version (23K):
[in a new window]
Fig. 1.
Screening process.
Level 1 Screening
Using a computerized hospital census database, a randomly drawn sample of 689 patients was identified from all preschool children between the ages of 3.0 and 6.1 years who resided in target residential areas (level 1 screening). The target areas were identified using five residential ZIP codes with the highest police district crime rates (twice the city-wide rate) for seven serious crimes (homicide, rape, robbery, aggravated and simple assault, burglary, and larceny) in the City of Boston for the years of 1991 to 1995.21 Residential ZIP codes corresponded to geographically contiguous census tracts with a high concentration of ethnic minority residents, female-headed households, adolescent pregnancy, and school dropout rates according to census data for 1990. Study announcements describing the study were placed in pediatric waiting rooms before recruitment. Letters were sent to the 689 mothers to inform them of the study and to offer them the opportunity to decline study participation by mailing back a postcard. There were 3% (n = 24) of mothers who mailed back postcards; and 8% (n = 50) were not located; no further contact was made with these mothers.
Level 2 Screening
A research assistant approached 89% (n = 615) via the telephone or a home visit to determine further eligibility (level 2 screening) and willingness to participate. A stringent entry criteria was imposed to screen out mothers or children at high risk for adverse psychologic outcomes because of other risk factors such as teenage parenting, foster care placement, residential instability, shelter placement, medical disability, child low birth weight, or chronic health problems. Mothers were included if they: were 19 years or older, the biological or adopted mother and primary caregiver of the index child, resided in a target ZIP code for at least 9 of the 12 months during the last year, did not reside in shelters or other residential settings, did not receive Social Security Supplemental Income, and their children weighed at least 4 pounds at birth, and had no serious health problems or disability. In addition, only mothers who were fluent in either English or Spanish were included. Based on above criteria, 49% (n = 338) of the initial subject pool did not meet level 2 eligibility criteria (Fig 1). Seventeen percent (n = 117) of mothers declined participation during level 2 screening after a phone call or a home visit. Mothers who declined to participate (overall 20%) did not differ from participating mothers in regard to residential ZIP code or child's age. More Hispanic (66%) than African-American mothers (45%) agreed to participate.
Home Protocol
Mothers were interviewed alone, face-to-face, by one of two female interviewers in their homes. In 89% of cases, the partner was not home at the time of the interview. The self-report protocol included structured, semi-structured questionnaires, and open-ended questions about mothers' perceptions of the community, themselves, their partners, and their children. A videotape of mother and child during a play session was also gathered in the home. The session lasted ~2.5 hours. Mothers were compensated $40 for their time.
Measures
ROs Lifetime history of RO was assessed via self-report, using structured questions regarding whether or not the study participant ever filed a RO (whether RO was temporary or permanent was not ascertained), the victim/defendant relationship, the number of times a RO was filed, and the year or years of filing. Also noted was whether or not the defendant in the RO was the current partner (same partner) or not (different partner). The type of abuse was assessed using an open-ended question in which study participants were asked to describe in detail the reason or reasons for obtaining the RO. Maternal narratives were written verbatim by trained research assistants. Based on the maternal narratives collected from the open-ended question, two independent blind coders (B.G. and J.G.) coded for the presence of one or more types of abuse. Four types of abuse were coded:
and he would not stop.').
Sample Characteristics
Because of links among demographic characteristics (such as low socioeconomic status and ethnic minority background), different types of victimization histories (such as maternal childhood physical abuse and lifetime sexual abuse), alcohol and substance use, and maternal health,22 it is important to control for the impact of these contributing factors in a study of interpersonal violence. Via the face-to-face maternal interview, demographic background information was gathered, including maternal and child age, number of children, ethnicity, immigration status, years of education, whether or not mother was currently employed, and whether or not mother was a recipient of public assistance (Aid to Families With Dependent Children). We gathered partner-related information including marital status, whether or not mother was involved with a current partner, length of involvement with current partner, and the number of partners in the past 5 years.
We also gathered information about other types of maternal past victimization, including history of childhood physical abuse, lifetime sexual victimization, and use of alcohol and illicit drugs (described in the "Appendix").
Measurement of Outcomesa
Maternal Outcomes
Partner Aggression to Mother. The Verbal Aggression,
Physical Violence, and Total Partner Aggression scales from the
Conflict Tactics Scale-Revised (CTS-R)23,24 were used to
measure partner-to-woman conflict during the past 30 days. Reliability
coefficients for these scales were 0.37, 0.92, and 0.87, respectively.
coefficients for these scales were 0.58 and 0.85, respectively.
Distress Symptomatology. Three global indices of the
SCL90-R26 were used to measure distress symptoms in the
past 7 days. Reliability coefficients for these scales ranged from 0.89 to 0.96.
Posttraumatic Stress (PTS)-related Symptomatology. The
Diagnostic Interview Schedule PTSD-Module modified version used by Kilpatrick and associates in the National Women Study27 was used to measure posttraumatic stress-related symptoms and partial
PTSD diagnosis. Reliability coefficient for the PTSD scale was
0.86.
Child Outcomes
Partner Aggression to Child. The
CTS-R23,24) described above, was also used to measure
partner verbal aggression, physical violence, and total aggression
toward the child during the past 30 days. Reliability coefficients
for these scales were 0.65, 0.78, and 0.21, respectively.
coefficients for these scales ranged from
0.78 to 0.94.
Posttraumatic Stress (PTS)-related Symptomatology. A
developmentally appropriate symptom checklist for preschool children, developed in our laboratory, was used to measure posttraumatic stress-related symptoms. Reliability coefficient for this scale was
0.62.
Data Analyses
Preliminary univariate analyses were performed to assess mean differences between the restraining and non-RO groups for demographic characteristics, other types of maternal past victimization, and alcohol and drug use. To avoid type 1 error and reduce the number of statistical comparisons, multivariate analyses of covariance (MANCOVA) were performed on maternal and child outcomes with RO group as the between-subject factor. In the MANCOVA analyses, outcome variables were entered separately for mother and child. MANCOVA outcome variables for mother included: partner verbal aggression and physical violence, general health, bodily pain, global severity index, positive symptom total, positive symptom distress index, and PTS-total score. If MANCOVA analyses were significant, they were followed by analyses of covariance using covariates identified in univariate tests. In a parallel analytical procedure using categorical variables, we performed logistic regression analyses to test for the association between partial PTSD diagnosis and RO group, controlling for the same maternal covariates. MANCOVA outcome variables for child included: partner verbal aggression and physical violence, internalizing, externalizing and total behavior problems, and PTS-related symptoms.
| |
RESULTS |
|---|
|
Top
Abstract
Methods
Results
Discussion
Conclusion
References
|
|---|
Demographics
The sample comprised of 160 mothers and their 3.0- to 6.1-year-old preschool children who resided in five high-crime residential ZIP codes. On the average, mothers were in their early thirties. There were 53% African-American mothers, 10% African-Caribbean, 9% of other African heritage, 22% Hispanic, and 6% of other ethnic groups; 63% were US born, whereas the remaining were born in a myriad of 21 different countries reflecting the international composition of residents in the Boston area. Among immigrant mothers, the average length of immigration was 11.0 years. Mothers were interviewed in English (n = 140) or Spanish (n = 20). Among mothers, 42% had completed high school or had a GED, 32% had some college or post-high school training, whereas the remaining had less than high school education. In the sample, 52% were employed out of the home, 45% were recipients of public assistance, 24% were married, and 71% were involved with a partner at the time of the interview.
ROs
Among the sample, 64 mothers (40%) reported a history of filing a RO. At the time of the offense, defendants included 25 (38%) boyfriends or husbands, 28 (43%) ex-boyfriends or ex-husbands, 9 (13%) a friend or relative, and 5 (7%) were someone known. Thirteen mothers (20%) filed more than one order. Year of filing RO ranged from <45 days to 15 years before the interview with a mean of 3.9 years (standard deviation = 3.5), and a median of 3 years. In 78% of cases, the RO was filed during the past 73 months (the upper age limit of the children in the sample). Among mothers who filed RO against partners (n = 52), 32 reported having a different partner, whereas 20 reported the same partner (current or most recent) at the time of the interview.
RO were filed because of the following reasons: verbal harassment (39%), verbal threats or intimidation (37%), physical assault (57%), or destruction of property (4%). Reasons for filing a RO involved single (72%) or multiple (28%) types of abuse (such as verbal harassment and physical assault). The majority of multiple abuse cases (77%) involved physical assault. The type of abuse was not related to the type of defendant.
Group Comparisons (see Table 1)
Mothers in the RO group were more likely to be recipients of public assistance (P < .04) and US born (vs foreign born) (P < .03). There were no group differences on maternal age, age of target child, number of children, ethnicity, years of school completed, or current employment status.
|
Mothers in the RO group were less likely to be married than those in the non-RO group (P < .004). There were no significant group differences in the number of mothers involved with a current partner, the length of the relationship with current partner, or the number of past partners in the past 5 years.
Mothers in the RO group were more likely to report lifetime sexual victimization than those in the non-RO group (P < .05). No significant mean differences were found between the two groups on whether they experienced childhood physical abuse (including harsh punishment or use of a belt).
There were no significant differences between the RO and non-RO groups regarding current use of alcohol or illicit drugs.
Maternal Outcomes
Partner Aggression to Mother Table 2 presents the number (and percentages) of mothers who reported the presence of aggressive acts perpetrated by current partner (or most recent partner) in the past 30 days (or last 30 days of the relationship) as measured by individual items of the CTS-R. For the RO group, 10 to 41 (15%-64%) of mothers reported verbal aggression acts, compared with 8 to 34 (8%-35%) of mothers in the non-RO group. Physical violence acts were reported by 0 to 17 (0%-26%) of mothers and 0 to 8 (0%-8%) of mothers in the RO and non-RO groups, respectively. The majority of mothers, 86% and 70%, reported at least one verbal aggression act, whereas 31% and 13% reported at least one physical violence act, in the RO and non-RO groups, respectively.
|
Maternal Multivariate Analyses Based on preliminary univariate analyses, we identified the following covariates which were entered in MANCOVA analyses: immigrant status (US vs foreign born), public assistance status (recipient vs not recipient of Aid to Families With Dependent Children), marital status (married, separated, divorced, widowed, or nonmarried), and lifetime sexual victimization (number of reported sexual victimization events). The MANCOVA analyses, testing for maternal outcomes and assessing for these covariates in the model, resulted in significant main effects for RO group F(7, 145) = 4.45, P < .001. Analyses of covariance which show group mean differences after adjusting for covariates in the model, are presented in Table 3. Mothers in the RO group reported higher partner verbal aggression to mother (P < .006), higher physical violence to mother (P < .008), poorer health (P < .01), and higher DIS PTSD-symptoms (P < .0001) than those in the non-RO group. There was a tendency for mothers in the RO group to report higher global distress symptoms (intensity of symptomatology) (P < .07), positive symptoms (breath of symptomatology) (P < .09), and bodily pain (P < .09), compared with mothers in the non-RO group.
|
|
Child Outcomes
Partner Aggression to Child Table 5 presents the number (and percentages) of mothers who reported individual aggressive acts perpetrated against their child by current partner (or most recent partner), as measured by items in the CTS-R. For children in the RO group, partner verbal aggression, such as insulting or threatening to hit, was reported for 4 to 10 (6%-15%) children, compared with 1 to 9 (2%-9%) children in the non-RO group. Partner physical violence, such as pushing, kicking, or slapping was reported for 14 children (ranging from 2%-6%) in both groups.
|
Child Multivariate Analyses Multivariate analyses of variance performed for child outcomes (partner verbal aggression to child; partner physical violence to child; internalizing, externalizing, and total behavior problems; and PTS-related symptoms) without the inclusion of maternal covariates. These multivariate analyses of variance resulted in a nonsignificant group effect (F[4, 145] = 2.34; P < .06). Table 6 reports the unadjusted mean differences for child outcomes. Children in the RO group were reported to experience higher partner verbal aggression (P < .02) and higher externalizing problems (P < .02) than children in the non-RO group. MANCOVAs performed with child outcome variables controlling for maternal covariates (immigration status, marital status, public assistance, and lifetime sexual victimization) diminished the likelihood of an independent RO group effect (P < .14).
|
| |
DISCUSSION |
|---|
|
Top
Abstract
Methods
Results
Discussion
Conclusion
References
|
|---|
The inquiry of ROs resulted in high rates of self-disclosure of interpersonal violence. The finding that 40% of the sample disclosed histories of filing a RO was unexpected as well as alarming for a nonreferred sample. The sampling procedure used in this study aimed at oversampling mothers exposed to community violence, that is, interpersonal violence involving nonintimate persons. The high prevalence of ROs in high-crime neighborhoods provides another example to existing literature regarding the association of different forms of violence.30 Under Bronfenbrenner's31 ecologic perspective of "embeddedness of social contexts," proximal social contexts (eg, family violence) are nested within broader social contexts (eg, community violence). Different forms of contextual violence interact with each other in the development or expression of child risk for psychopathology. Likewise, the associations found in this study between ROs and sexual victimization, once again, supports the notion that multiple forms of victimization co-occur in the same individuals, although the mechanism of this link is not fully understood.32 From a social advocacy perspective, these findings provide support for the need to address forms of violence as related public health problems using a comprehensive and multilevel approach.
Mothers who were US born, not married, and on public assistance were more likely to file a RO. We speculate that American women may be more familiar with the justice system and more comfortable calling the police than foreign born women. Public assistance gives women a degree (albeit meager) of financial independence that women without public assistance did not have. The association of these demographic characteristics and filing ROs suggest that familiarity or accessibility to local courts and law enforcement systems and the financial independence from potentially abusive partners may be important factors in the decision to file a RO.
The majority of RO were filed against a current boyfriend or an ex-boyfriend (83%) at the time of the events, a profile that is consistent with state and national RO statistics. The high number of defendants who were ex-boyfriends suggests that leaving a relationship may not prevent violence. For example one mother said: "After I broke up I wanted him to stay away from me, but he began fighting with me and harassing me... " Thus, unfortunately, filing a RO may not protect women from abusive relationships and in a few cases it may even exacerbate violence.
In this study the time of filing ROs ranged from a few weeks to several years before the interview. It would not be surprising to find higher rates of partner aggression and stress-related symptoms in mothers at the time of filing a RO. Yet, this study finds that years later (a median of 3 years after orders were filed), maternal groups created based on RO histories are distinctly different regarding current, not past, level of partner aggression (verbal aggression and physical violence), impaired health, and PTS-related symptoms. These group differences continued to be statistically significant after we controlled for a number of socioeconomic variables and other types of past victimization, known to affect women's health. For the mother, a RO history may be a marker for poor health and for an array of psychologic symptoms, particularly posttraumatic stress-related symptoms. Thus, the inquiry into RO may provide pediatric providers with a potentially efficient tool for identifying mothers who are at current risk for adverse health outcomes.
For children in the RO group, our results showed slightly higher rates of current partner verbal aggression toward the child (such as insulting and threatening to hit the child). Verbal aggression is of concern because of the potential for the verbal abuse to escalate to physical abuse.33 The association between mothers with histories of RO and the child's behavior was weaker than expected, in light of the increasing evidence of the detrimental effect of a child witnessing interpersonal violence on child adjustment.34 Two variables bear further exploration: first, the age of the child in this sample at the time of the mother's filing the RO was between 0 to 5 years of age for the majority (78%) of the children. Thus, it is likely that the experience of witnessing interpersonal violence varied widely in the RO children based on developmental level. Second, the child's relationship with the defendant also varied. In some cases the defendant was not well known to the child. These factors (ie, age of the child at the onset and the emotional proximity of the child to the perpetrator) are known to affect the child's response to a violent interpersonal environment.35 The heterogeneity of the RO children may have masked the true impact of violence for those exposed to chronic interpersonal violence perpetrated by a parental figure. Therefore, replication studies using groups of various exposure levels are necessary to interpret child results and to better understand the effects of RO on child outcomes during the preschool years.
Overall, mild levels of verbal aggression to the mother by current partners seemed normative in this selected sample of mothers residing in high-risk neighborhoods. For example, ~70% to 86% of mothers reported at least one instance of the partner insulting or threatening to hit in a 30-day period. Approximately 13% to 31% reported actual physical violence (such as pushing or hitting) within the same 30-day period. Furthermore, the sample, as a whole, showed elevation of PTS-symptoms and a substantial number of mothers met partial lifetime PTSD diagnosis (24%-42%). These findings suggest that mothers selected through the sampling procedure of high-crime residential ZIP codes suffer high levels of current interpersonal violence and traumatic stress regardless of their RO histories.
There are several limitations in this study. The inquiry about RO by a trained research assistant in the home may not be the same as a provider's inquiry of RO within a clinical practice. The efficacy of this approach in clinical practice remains to be seen. The sample was relatively small and included a selected group of mothers whose children received pediatric primary care, who resided in selected high-crime areas, and who met stringent study criteria. Only mothers that agreed to be interviewed in their homes were included. Thus, these results may not generalize to other mothers. Despite the sampling limitation, the results presented here may apply to a substantial number of indigent ethnic minority mothers residing in high-crime risk areas in the inner city. It is likely that these results may not apply to mothers who, although experiencing interpersonal violence, presently or in the past, did not file a RO. These women may be at a different stage in confronting the violence in their lives. This group, although not addressed in this study, may be at greater need for identification and clinical intervention.
| |
CONCLUSION |
|---|
|
Top
Abstract
Methods
Results
Discussion
Conclusion
References
|
|---|
In summary, the inquiry of ROs is one method to identify mothers with past interpersonal violence that may serve as a proxy to identify current interpersonal violence and maternal health problems. The findings of this study may have important implications for pediatric practitioners who deliver care to families from high-crime neighborhoods. Because mothers with young children interact frequently with pediatric providers for routine and sick child care, providers may play an important role in the identification and referral of women who are victims of interpersonal violence.
However, providers face a number of difficulties in identifying mothers who are victimized by interpersonal violence. The mother is not the patient; mothers may be particularly reluctant to disclose violence to a pediatrician for fear of involvement with child protective services; the pediatric provider may have limited training in interviewing women on sensitive topics, such as violence. Training for pediatric providers to enable them to more competently interview women about histories of interpersonal violence would be important.
This study supports the argument for pediatric screening for histories of domestic violence in all families. Filing a RO may be the culmination of a long history of violence and may represent a public statement to stop the violence in a woman's life. Once women disclose the violence to local law enforcement authorities, these interpersonal events become acts in the public domain. We speculate that RO may be easier to disclose than other violence-related events because they already entered the public domain. We recognize, nevertheless, that there may be other ways to assess past interpersonal violence in similarly efficient ways. Further research is needed to address the relative effectiveness of RO in clinical settings as compared with other methods of assessing partner relationships and related maternal and child health risk.
The inquiry of ROs may be a first step toward offering assistance and safety to mothers and to their children, thus, ensuring family-centered pediatric care.
| |
APPENDIX |
|---|
Childhood Physical Abuse
Based on the work of Famularo and colleagues,36 women were asked seven yes/no question about whether they experienced physical abuse by primary caregivers before age 16. Questions ranged from being hit "too hard" with a belt or any object, to being threatened or attacked with a weapon.
Lifetime Sexual Victimization
Modeled from the work of Koss and colleagues,37 women were asked five yes/no sexual questions using a general probe about "sexual advances that you did not want," including forced sexual intercourse, anal sex, oral sex, and object sexual penetration. One question was asked about inappropriate sexual touch by a parent or parent figure before age 16.
Addiction Severity Index (ASI)38
Current use (past 30 days) of alcohol and illicit drugs was measured using composite scores for alcohol and illicit substances of the ASI.
The CTS-R
Mothers who did not report a partner at the time of the interview (29%) were asked to report on most recent partner. The CTS-R yielded two scales: Verbal Aggression (7 items) which included the use of verbal and nonverbal acts which symbolically hurt or the use of threats to hurt; and Physical Violence (9 items), which included physical acts of aggression ranging from pushing to attack with a weapon. We added a separate item about the partner's demanding unwanted sex. The CTS-R is a widely used instrument with robust psychometric properties, which detects occurrences of severe and minor forms of violence in the home.23,24 The same instrument (minus the added item) was administered for the partner-to-child measure.
Health Survey
The General Health and the Bodily Pain scales were used to evaluate personal health (5 items) and pain or limitations because of pain (2 items) in the past 4 weeks.
The SCL90-R Scale
This is a 90-item scale that yields three indices of psychologic distress: the Global Severity Index is a measure of the number of items reported and the intensity of perceived distress (sum of all scores); the Positive Symptom Total (PST) is a measure of symptom breadth (number of items with a nonzero response); and the Positive Symptom Distress Index (PSDI) is a measure of subject intensity (sum of all item values by the PST). The three indices provided distinct indicators of psychologic disorder.39 Scores were expressed as t scores with a mean of 50 and a standard deviation of 10.
The Diagnostic Interview Schedule PTSD-Module
This is a 20-item scale which is prefaced with a
nonevent-specific probe and addresses symptom duration of 2 weeks or
more during the last month, last 6 months, and more than 6 months. Scores were summed across symptom duration. If a symptom was reported, we followed-up by asking about the nature of the event and subsequently rated it as a qualifying or a nonqualifying event. Only qualifying events were entered in the analyses. In previous studies, the DIS
PTSD-Module produced 
coefficients of agreement 0.71, sensitivity of
0.96, and a specificity of 0.80 when compared with the Structured Clinical Interview for DSM-IIIR.40 We used the DIS
PTSD-Module as a continuous variable to assess total symptom duration
across the three intervals. The PTS-Total score has produced adequate reliability ( 0.89). We categorized mothers on whether or not they
reported a cluster of PTS-related symptoms (presence of 1 reexperiencing, 3 avoidance, and 2 increased arousal symptoms), and
thus met partial criteria for lifetime PTSD diagnosis.
The Child Behavior Checklist (CBC)
The CBC 2-3 is a 99-item checklist completed by the child's mother yielding two broadband groupings: internalizing and externalizing, and a total problem score. This instrument has adequate test-retest reliability (r = 0.87), and moderate predictive relationship with other checklists 3 years later (r = 0.49). The CBCL, a 118-item checklist, is the earlier upward version of the CBC 2-3. In this article, we used the two broadband groupings of internalizing, externalizing, and the total behavior problem scores available for both versions of the CBC.
The PTS-related Checklist for Children (PTS-C)
Based on the Diagnostic Classification of Mental Health and
Developmental Disorders of Infancy and Early Childhood Zero to Three41 and on the work of Scheeringa and
colleagues,42 the PTS-C is a 3-point maternal
rating of 13 behaviors associated with PTSD symptoms and which are not
included in the CBC scales. In this measure, symptoms are not linked to
any specific trauma. If a symptom was present, interviewer queried
about the qualifying event. Behaviors were reported if they occurred in
the past 2 months. Moderate reliability was obtained in the PTS-C ( 0.52). The PTS-C was positively correlated with increased CBCL total and internalizing problems (r = 0.30, 0.45, and 0.44, respectively).
| |
ACKNOWLEDGMENTS |
|---|
This study was supported with funds from the National Institute of Mental Health and the National Institute of Drug Abuse R01DA/MH11157 (NIH Interagency Consortium on Violence Against Women Research).
We thank the mothers and children who participated in this study. We also acknowledge Tim Heeren, PhD, and Jean Cantey-Kaiser, PhD, from the Boston University School of Public Health who provided statistical assistance to the project and to Howard Bauchner, MD, from the Division of General Pediatrics at Boston University School of Medicine for his insightful comments on an earlier manuscript.
| |
FOOTNOTES |
|---|
a Outcome measures are fully described in the "Appendix."
Received for publication Nov 16, 1998; accepted Feb 16, 1999.
Reprint requests to (L.O.L.) New York University Child Study Center, Division of Child and Adolescent Psychiatry, 550 First Ave, New York, NY 10016. E-mail: bgervey{at}aol.com
| |
ABBREVIATIONS |
|---|
RO, restraining order; PTS, posttraumatic stress; CTS-R, Conflict Tactics Scale-R; PTSD, posttraumatic stress disorder; CBC, Child Behavior Checklist; MANCOVA, multivariate analysis of covariance; PST, positive symptom total; OR, odds ratio; CI, confidence interval.
| |
REFERENCES |
|---|
|
Top
Abstract
Methods
Results
Discussion
Conclusion
References
|
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. L. Postmus Challenging the Negative Assumptions Surrounding Civil Protection Orders: A Guide for Advocates Affilia, November 1, 2007; 22(4): 347 - 356. [Abstract] [PDF]
|
|
|
|
 |
|
 M. B. Phelan Screening for Intimate Partner Violence in Medical Settings Trauma Violence Abuse, April 1, 2007; 8(2): 199 - 213. [Abstract] [PDF]
|
|
|
|
 |
|
 S. B. Sorenson and H. Shen Restraining Orders in California: A Look at Statewide Data Violence Against Women, July 1, 2005; 11(7): 912 - 933. [Abstract] [PDF]
|
|
|
|
 |
|
 C. E. Jordan Intimate Partner Violence and the Justice System: An Examination of the Interface J Interpers Violence, December 1, 2004; 19(12): 1412 - 1434. [Abstract] [PDF]
|
|
|
|
|
|
M. B. Mechanic Beyond PTSD: Mental Health Consequences of Violence Against Women: A Response to Briere and Jordan J Interpers Violence, November 1, 2004; 19(11): 1283 - 1289. [Abstract] [PDF]
|
|
|
|
 |
|
 S. Maman, J. K. Mbwambo, N. M. Hogan, G. P. Kilonzo, J. C. Campbell, E. Weiss, and M. D. Sweat HIV-Positive Women Report More Lifetime Partner Violence: Findings From a Voluntary Counseling and Testing Clinic in Dar es Salaam, Tanzania Am J Public Health, August 1, 2002; 92(8): 1331 - 1337. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. W. BARNETT Why Battered Women Do Not Leave, Part 2: External Inhibiting Factors--Social Support and Internal Inhibiting Factors Trauma Violence Abuse, January 1, 2001; 2(1): 3 - 35. [Abstract] [PDF]
|
|
|
|
 |
|
 Uncovering Domestic Violence: Ask About Restraining Orders Journal Watch Women's Health, November 1, 1999; 1999(1101): 18 - 18. [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||
