Martin SL, Mackie L, Kupper LL, Buescher PA, Moracco KE. Physical Abuse of Women Before, During, and After Pregnancy. JAMA. 2001;285(12):1581-1584. doi:10.1001/jama.285.12.1581
Author Affiliations: Department of Maternal and Child Health, School of Public Health (Dr Martin), Department of Biostatistics (Ms Mackie and Dr Kupper), and Department of Health Behavior and Health Education and Injury Prevention Research Center (Dr Moracco), University of North Carolina, Chapel Hill; and the State Center for Health Statistics, Raleigh, NC (Dr Buescher).
Context Clinicians who care for new mothers and infants need information concerning
postpartum physical abuse of women as a foundation on which to develop appropriate
clinical screening and intervention procedures. However, no previous population-based
studies have been conducted of postpartum physical abuse.
Objectives To examine patterns of physical abuse before, during, and after pregnancy
in a representative statewide sample of North Carolina women.
Design, Setting, and Participants Survey of participants in the North Carolina Pregnancy Risk Assessment
Monitoring System (NC PRAMS). Of the 3542 women invited to participate in
NC PRAMS between July 1, 1997, and December 31, 1998, 75% (n = 2648) responded.
Main Outcome Measures Prevalence of physical abuse during the 12 months before pregnancy,
during pregnancy, and after infant delivery; injuries and medical interventions
resulting from postpartum abuse; and patterns of abuse over time in relation
to sociodemographic characteristics and use of well-baby care.
Results The prevalence of abuse before pregnancy was 6.9% (95% confidence interval
[CI], 5.6%-8.2%) compared with 6.1% (95% CI, 4.8%-7.4%) during pregnancy and
3.2% (95% CI, 2.3%-4.1%) during a mean postpartum period of 3.6 months. Abuse
during a previous period was strongly predictive of later abuse. Most women
who were abused after pregnancy (77%) were injured, but only 23% received
medical treatment for their injuries. Virtually all abused and nonabused women
used well-baby care; private physicians were the most common source of care.
The mean number of well-baby care visits did not differ significantly by maternal
patterns of abuse.
Conclusion Since well-baby care use is similar for abused and nonabused mothers,
pediatric practices may be important settings for screening women for violence.
Many women, including pregnant women and those soon to become pregnant,
have been physically abused by intimate partners or others.1- 12
Studies of abuse during the year preceding pregnancy report prevalence estimates
ranging from 4% to 26%,5,8- 11
while investigations of abuse during pregnancy have generally found prevalences
of 4% to 8%.9 Differences in these prevalence
figures are probably due to variations in violence across populations, as
well as use of differing study methods.
Although research on abuse before and during pregnancy is increasing,
a dearth of information exists concerning abuse that occurs after infant delivery,
a stressful time for many families.13,14
Only 3 clinically based investigations were identified that followed up patients
after delivery to examine various types of postpartum abuse. Two of these
studies found postpartum abuse prevalences of 19% to 24%, while the third
study (which focused on women who were abused during pregnancy) found a postpartum
abuse prevalence of 90%.10,15,16
The American Academy of Pediatrics17
has recently joined other organizations, including the American Medical Association18 and the American College of Obstetricians and Gynecologists,19 in endorsing screening for violence among female
patients. Thus, clinicians who care for new mothers and infants need additional
empirical information concerning postpartum abuse to provide a foundation
on which to develop and implement appropriate abuse screening, referral, and
intervention procedures. This article examines (1) the prevalence and patterns
of physical abuse before, during, and after pregnancy; (2) injuries and medical
interventions resulting from postpartum abuse; and (3) patterns of abuse over
time in relation to women's sociodemographic characteristics and use of well-baby
Data are from the North Carolina Center for Health Statistics/Centers
for Disease Control and Prevention North Carolina Pregnancy Risk Assessment
Monitoring System (NC PRAMS). This ongoing project (described elsewhere)20 is a population-based mailed and telephone survey
of a representative group of North Carolina women who recently delivered live-born
infants. The North Carolina birth certificate file serves as the sampling
frame source for NC PRAMS, with oversampling of women with low-birth-weight
infants. Seventy-five percent of the 3542 women invited to participate in
NC PRAMS between July 1, 1997, and December 31, 1998, responded to the survey
(n = 2648); the mean postpartum period (ie, the time between infant delivery
and survey completion) was 3.6 months. Comparison of birth certificate information
for respondents vs nonrespondents showed that nonrespondents were significantly
more likely to be young, unmarried, and black and have less than a high school
The survey assessed physical abuse by asking women if they had been
pushed, hit, slapped, kicked, or physically hurt in some other way during
each of 3 periods: the 12 months before becoming pregnant, during pregnancy,
and after infant delivery. Women abused during each period were asked about
their social relationship with the perpetrator of the abuse (eg, husband/partner),
and women who experienced postpartum abuse were asked a series of questions
about particular types of violence-related injuries (eg, pain experienced
the day after the abuse, weapon wounds) and medical care for these injuries.
The survey also documented women's sociodemographic characteristics and well-baby
care use (defined as taking the infant to a physician or nurse for routine
care, excluding visits due to illness).
Abuse prevalences before, during, and after pregnancy were computed.
Odds ratios (ORs) and 95% confidence intervals (CIs) quantified associations
between abuse that occurred during a previous period and that which occurred
at a subsequent time. Descriptive and bivariate analyses examined relationships
between patterns of abuse and women's sociodemographic characteristics, as
well as use of well-baby care. SUDAAN software21
was used to take the survey sampling methods into account.
The statewide abuse prevalence during the 12 months before pregnancy
was 6.9% (95% CI, 5.6%-8.2%); 68% of perpetrators were current or former husbands/partners,
14% were family members, 8% were multiple persons, 5% were friends, and 5%
were someone else. Similarly, abuse prevalence during pregnancy was 6.1% (95%
CI, 4.8%-7.4%); 67% of perpetrators were current or former husbands/partners,
14% were family members, 4% were multiple persons, 3% were friends, and 12%
were someone else. Postpartum abuse prevalence was 3.2% (95% CI, 2.3%-4.1%);
76% of perpetrators were current or former husbands/partners, 10% were multiple
persons, 9% were family members, 5% were friends, and less than 1% were someone
else. Seventy-seven percent of women who were abused after delivery were injured,
experiencing pain the day after the abuse (73%), sprains/bruises/small cuts
(57%), head/internal/permanent injuries (9%), weapon wounds (8%), and broken
bones/severe cuts/burns (6%). Although three quarters of these women had multiple
types of injuries, only 23% received medical care for them.
Abuse during a previous period was a strong risk factor for subsequent
abuse (Figure 1). A strong, significant
association was found between abuse before and during pregnancy (OR, 67.6;
95% CI, 27.3-167.2). There also was a strong, significant association between
abuse during pregnancy and postpartum abuse among women who were abused before
pregnancy (OR, 38.0; 95% CI, 5.8-247.3). Among women who were not abused before
pregnancy, the positive association between abuse during pregnancy and postpartum
abuse also was strong; however, this association did not reach statistical
significance (OR, 4.4; 95% CI, 0.9-22.0). Less than 1% of all women experienced
abuse for the first time after infant delivery (ie, they were abused after
delivery but not before or during pregnancy). Furthermore, only 29% of women
who were abused after delivery had not been abused during the year before
pregnancy, so absence of abuse before pregnancy was strongly protective against
postpartum abuse (OR, 0.02; 95% CI, 0.01-0.06). Similarly, only 18% of women
who were abused after delivery had not been abused before or during pregnancy,
so absence of any previous abuse was strongly protective against postpartum
abuse (OR, 0.01; 95% CI, 0.001-0.05).
Respondents' sociodemographic characteristics did not vary significantly
among the 8 patterns of abuse (Table 1).
However, comparison of women with a history of abuse during any period with
never-abused women found that abused women were significantly more likely
to be unmarried (χ21 = 43.5; P = .01), poor (χ21 = 21.7; P = .02), and younger (χ21 = 21.6; P = .02), with less than a high school education (χ21 = 20.2; P = .02).
Virtually all respondents took their infants to health care practitioners
for well-baby care (Table 2).
Both abused and nonabused women used a variety of well-baby care sources,
with private physicians being the predominant care source in 7 of the 8 abuse
pattern groups. There was no statistically significant association between
abuse patterns and primary source of well-baby care, nor did the mean number
of well-baby care visits differ significantly between infants of never-abused
women and any of the other 7 pattern groups of abused women.
This is the first statewide study, to our knowledge, to examine women's
postpartum physical abuse experiences in addition to abuse before and during
pregnancy. Abuse prevalence was relatively low (3.2%) during the mean 3.6-month
postpartum period studied compared with prevalence of abuse during the 12
months before pregnancy (6.9%) and the approximate 9 months of pregnancy (6.1%).
It is noteworthy that the highest prevalence estimate resulted from examination
of the longest period, whereas the lowest prevalence estimate resulted from
examination of the shortest period. It is important to note that even this
relatively low prevalence translates into the abuse of more than 3000 new
mothers annually in North Carolina. Previous abuse was a strong risk factor
for subsequent abuse (including postpartum), a finding consistent with other
research documenting the often long-term nature of violence.2
Although women abused after pregnancy often were injured (with few receiving
medical care for their injuries), these women managed to bring their infants
to well-baby care visits as frequently as nonabused women, and this care was
provided most often by private physicians.
These findings should be interpreted in light of the study's methodological
constraints. For example, 25% of women invited to participate in NC PRAMS
did not complete the survey, with nonrespondents more likely than respondents
to be young, unmarried, black, and of low education levels. Furthermore, survey
responses concerning sensitive topics such as abuse are prone to response
bias, which may lead to underestimation of the true extent of abuse. Moreover,
women's ability to recall abusive events may vary as a function of the period
asked about, with less recall of events that occurred in the more distant
past. In addition, the survey did not ask about the composition of women's
households or whether they changed intimate partners during the 3 periods
examined; thus, we are unsure of whether the initiation or discontinuation
of violence as time progressed was associated with these types of alterations.
Similarly, information was not available concerning types of abuse other than
physical abuse (eg, psychological). Finally, since NC PRAMS includes only
women whose pregnancies resulted in live births, these findings may not be
generalizable to women with other types of pregnancy outcomes.
Despite these study limitations, these findings should alert health
care practitioners that women who are physically abused before and/or during
pregnancy often continue to experience abuse after infant delivery, placing
the health of both mother and child in jeopardy. Furthermore, these abused
mothers do take their infants to well-baby care visits. Thus, repeated clinical
screening of women for violence within various health care settings, including
pediatric practices, appears warranted. Given the current relatively low rate
of violence screening by pediatric practitioners,22- 25
enhanced education and training of pediatricians concerning the often long-term
nature of violence as well as appropriate abuse screening protocols and referral/intervention
procedures are needed. These are important steps toward ensuring that women
who experience physical abuse are provided with optimal care for this important