Factors associated with domestic violence prior to and during pregnancy in HIV infected and HIV at risk women

1
Factors associated with domestic violence prior
to and during pregnancy in HIV infected and HIV
at risk women

• Rodney Wright, MD
• Mentor Mayris Webber, DrPH

2
Background

• Verbal and physical abuse of women is a
  significant public health problem
• Approximately 1.5 million women are raped and/or
  physically assaulted by an intimate partner each
  year in the US
• Over 1200 women were killed by an intimate
  partner in the year 2000
• Domestic violence has been estimated to occur in
  4-8 of pregnancies

CDC Intimate Partner Violence Fact Sheet,
2006 Tjaden, et al Department of Justice Report
No. NCJ 181867
3
Background

• Domestic violence during pregnancy associated
  with
• Low birth weight
• Neonatal death
• Increased NICU admissions
• Domestic violence in the home has been associated
  with increased aggressive behavior in childhood

Yost et al, Obstetrics and Gynecology 2005
Jul106(1)61-5 Silverman et al, Am J Obset Gyn,
2006195(1)140-8 Whitaker, et al. Archives of
general psychiatry 2006 May63(5)551-60
4
Background

• HIV infection is likewise a major health concern
  for women
• AIDS is the leading cause of death in African
  American women aged 25 34
• Second leading cause of death in Hispanic women
  of the same age group

CDC, HIV/AIDS Among US Women, July 2003
5
Regional HIV statistics and features for women
2003 and 2005
Number or women (15-49) living with HIV Percent of adults (15-49) with HIV who are women ()
Sub-Saharan Africa 2003 2005 13.1 million 13.5 million 57 57
South and SE Asia 2003 2005 1.6 million 1.9 million 25 26
Caribbean 2003 2005 140 000 140 000 50 50
Western and Central Europe 2003 2005 180 000 190 000 27 27
North America 2003 2005 270 000 300 000 25 25
Total (worldwide) 2003 2005 16.5 million 17.5 million 47 46
Source UNAIDS / WHO AIDS Epidemic Update
December 2005
6
(No Transcript)
7
Background

• Many risk factors associated with domestic
  violence also associated with increase risk of
  contracting HIV
• Annual income lt 10,000
• Trading sex for drugs or money
• Having sex with men who use drugs
• Drug abuse
• Alcohol abuse

Koenig et al, Mat Child Health J, 2000
Jun4(2)103-9
8
Objectives

• Primary To determine the factors associated with
  domestic violence during and prior to pregnancy
  in HIV positive and HIV at risk women
  participating in the MIRIAD study
• Secondary To determine if domestic violence is
  associated with adverse birth outcomes in this
  cohort

9
MIRIAD

• Mother Infant Rapid Intervention at Delivery
• Evaluated the feasibility of performing rapid HIV
  testing on the labor floor
• Cohort consisted of poor, inner city, primarily
  minority women, some with limited access to
  medical care

10
Hypothesis

• HIV positive status is associated with domestic
  violence in the MIRIAD cohort of women

11
Methods

• MIRIAD conducted at 16 sites in six US cities
• Women offered enrollment into MIRIAD if they
  presented to the labor floor without
  documentation of HIV status
• Data collected from November 2001 - January 2005
• Data analyzed from post partum interviews
  conducted as part of this study
• Initially when a patient tested HIV positive, an
  interview was offered to her and the next three
  HIV negative MIRIAD enrollees at her site
• Procedure changed in December 2002 when 500 HIV
  negative women were interviewed
• Additional peri-partum data were extracted from
  patient charts

12
(No Transcript)
13
Methods

• Bivariate analyses were performed using ?²
  derived by logistic regression
• Odds ratios and 95 confidence intervals were
  derived
• Students t tests were performed to analyze
  demographic data
• Both linear and logistic regression were used to
  evaluate birth outcomes

14
Methods

• Separate multivariate logistic regression models
  were created to describe factors associated with
  ever experiencing domestic violence and for
  experiencing domestic violence during pregnancy
• Age, race and variables with plt0.25 were
  initially included in the violence during
  pregnancy model
• Age, race and variables with plt0.10 were
  initially included in the ever experiencing
  domestic violence model

15
Methods

• Multiple gestations were excluded from birth
  outcome analyses
• Women who were very ill or whose children did not
  survive were not interviewed
• Data analysis was conducted using Stata 9.1

16
Characteristics of population studied

HIV-Negative (n 653) HIV-Negative (n 653) HIV-Positive (n 35) HIV-Positive (n 35)
Characteristic Percent n Percent n p value
Age (years) 25.9 6.6 29.0 6.7 0.01
African American race 64 421 89 31 0.003
Single 51 326 77 27 0.002
History of Drug Use 33 216 29 10 0.58
History of STD 18 115 17 6 0.94
Received psychiatric services in past year 5 326 14 5 0.01
Total monthly household income less than 250 13 87 23 8 0.11
standard deviation mean includes marijuana and street drugs (cocaine, heroin, crack) standard deviation mean includes marijuana and street drugs (cocaine, heroin, crack) standard deviation mean includes marijuana and street drugs (cocaine, heroin, crack) standard deviation mean includes marijuana and street drugs (cocaine, heroin, crack) standard deviation mean includes marijuana and street drugs (cocaine, heroin, crack) standard deviation mean includes marijuana and street drugs (cocaine, heroin, crack) standard deviation mean includes marijuana and street drugs (cocaine, heroin, crack)
17
Results

• Ever experienced domestic violence 22
• HIV positives 24.2
• HIV negatives 21.8
• p0.75
• Experienced domestic violence during pregnancy
  5.8
• HIV positives 3.0
• HIV negatives 5.9
• p0.37

18
Correlates of domestic violence during pregnancy
n 688 Odds Ratio 95 Confidence Interval p
HIV status 0.37 0.45 3.19 0.37
Age lt24 years 2.71 1.27 5.80 0.01
Less than high school education 1.72 0.81 3.66 0.16
Total monthly household income less than 250 2.52 1.00 6.34 0.05
Mexican / Central American born 4.73 1.26 17.78 0.02
Used drugs during pregnancy 3.22 1.39 7.48 0.01
Ever used drugs 2.13 0.87 5.27 0.10
Father of baby uses drugs 4.71 1.69 13.14 0.003
Depression score 1.07 0.99 1.16 0.10
STD during pregnancy 1.22 0.43 3.44 0.71
Single 1.92 0.85 4.34 0.12
Homeless in past year 1.62 0.73 3.63 0.23
Currently lives in shelter 8.92 0.90 88.44 0.06
Moved in past year 1.97 0.93 4.16 0.08
Number of years lived in US 0.95 0.91 0.99 0.01
19
Correlates of ever experiencing domestic violence
n 688 Odds Ratio 95 Confidence Interval p
HIV status 1.14 0.51 2.59 0.75
Age lt 24 years 0.67 0.46 0.96 0.03
US born 2.26 1.47 3.47 lt0.001
Caribbean born 0.52 0.28 0.99 0.05
Less than high school education 1.51 1.05 2.19 0.03
Used drugs during pregnancy 3.36 1.99 5.66 lt0.001
Ever used drugs 3.45 2.34 5.11 lt0.001
Father of baby uses drugs 3.13 1.64 5.99 0.001
Depression score 1.16 1.11 1.22 lt0.001
History of STD 3.96 2.59 6.05 lt0.001
Single 1.89 1.30 2.76 0.001
Homeless in past year 3.35 2.11 5.31 lt0.001
In prison in past year 3.23 1.76 5.93 lt0.001
History of trichomonas 5.28 2.37 11.76 lt0.001
In drug rehab program in past year 7.28 1.32 40.13 0.02
New male partner during this pregnancy 3.55 1.67 7.54 0.001
Has children in foster care 4.13 2.07 8.03 lt0.001
20
Birth Outcomes

• Data available for 301 subjects (269 HIV negative
  and 32 HIV positive)
• No association between domestic violence during
  pregnancy and low birth weight, low Apgar score,
  preterm labor, cesarean delivery, or post partum
  complications
• HIV status was associated with low birth weight
  (OR 2.26, 95CI 1.00-5.10)
• This association was not present after adjusting
  for maternal age, race, and drug use (OR 1.64,
  95CI 0.67-4.03)

21
Birth outcomes
N 301 VARIABLE ODDS RATIO 95 CONFIDENCE INTERVAL p
Violence During Pregnancy
Cesarean delivery 0.26 0.05 1.28 0.10
Preterm labor 1.63 0.46 5.76 0.45
Positive urine toxicology screen 5.38 1.59 18.12 0.01
Low birth weight 2.40 0.69 8.29 0.17
Apgar score 0.13 -0.26 0.52 0.50
Post partum complications 0.98 0.17 5.51 0.98
Ever Experienced Violence
Cesarean delivery 1.01 0.53 1.92 0.97
Preterm labor 1.15 0.48 2.62 0.80
Positive urine toxicology screen 2.11 1.04 4.27 0.04
Low birth weight 1.03 0.51 2.05 0.94
Apgar score 0.03 -0.20 0.26 0.78
Post partum complications 2.37 0.86 6.47 0.09
Linear regression performed. Values reported are ß coefficients Linear regression performed. Values reported are ß coefficients Linear regression performed. Values reported are ß coefficients Linear regression performed. Values reported are ß coefficients
22
Multivariate logistic regression models
VARIABLE ODDS RATIO 95 CI p
Ever Experienced Domestic Violence
Age 1.04 1.01 1.07 0.02
Ever used drugs 2.33 1.46 3.71 lt0.001
Ever had STD 3.08 1.86 5.09 lt0.001
Number of children in the home 1.15 1.03 1.28 0.01
Depression score 1.10 1.04 1.16 lt0.001
Homeless in past year 2.67 1.51 4.72 0.001
Domestic Violence During Pregnancy Domestic Violence During Pregnancy Domestic Violence During Pregnancy Domestic Violence During Pregnancy
Age 0.94 0.88 1.01 0.07
Father of baby uses drugs 3.45 1.12 10.59 0.03
Lives in shelter 9.38 0.78 112.84 0.08
Used drugs during pregnancy 4.02 1.57 10.24 0.004
Variables considered in the prediction model HIV status, ever used drugs, new partner during pregnancy, partner support score, depression (CESD) score, ever had STD, homeless in past year, race, marital status, in prison in past year, number of children in the home, insurance status, place of birth and age Variables considered in the prediction model HIV status, educational level, lives in shelter, used drugs during pregnancy, depression (CESD) score, marital status, age, total monthly household income, place of birth, and father of baby currently uses drugs Variables considered in the prediction model HIV status, ever used drugs, new partner during pregnancy, partner support score, depression (CESD) score, ever had STD, homeless in past year, race, marital status, in prison in past year, number of children in the home, insurance status, place of birth and age Variables considered in the prediction model HIV status, educational level, lives in shelter, used drugs during pregnancy, depression (CESD) score, marital status, age, total monthly household income, place of birth, and father of baby currently uses drugs Variables considered in the prediction model HIV status, ever used drugs, new partner during pregnancy, partner support score, depression (CESD) score, ever had STD, homeless in past year, race, marital status, in prison in past year, number of children in the home, insurance status, place of birth and age Variables considered in the prediction model HIV status, educational level, lives in shelter, used drugs during pregnancy, depression (CESD) score, marital status, age, total monthly household income, place of birth, and father of baby currently uses drugs Variables considered in the prediction model HIV status, ever used drugs, new partner during pregnancy, partner support score, depression (CESD) score, ever had STD, homeless in past year, race, marital status, in prison in past year, number of children in the home, insurance status, place of birth and age Variables considered in the prediction model HIV status, educational level, lives in shelter, used drugs during pregnancy, depression (CESD) score, marital status, age, total monthly household income, place of birth, and father of baby currently uses drugs
23
Conclusions

• Many factors were found to be associated with
  domestic violence
• HIV infection was not associated with domestic
  violence
• Possibly due to small number of women testing
  positive
• Risk profiles of HIV positives and HIV negatives
  were similar in this population
• Similar number of sexual partners, prevalence of
  homelessness, history of STD, substance abuse
• Place of birth can be a significant predictor of
  risk of domestic violence

24
Limitations

• Data collected for reasons other than the
  evaluation of domestic violence
• Some questions related to domestic violence were
  not asked
• History of childhood sexual abuse
• How recently did the domestic violence that was
  reported to occur actually occur
• No scale of partner aggression or relationship
  power included
• Possible social desirability bias
• Direct questioning done on the labor floor may
  have limited responses to questions
• Possibly underestimates rates of domestic
  violence and risk factors

25
Limitations

• Birth outcome data limited by the exclusion of
  women who were very ill themselves or whose
  infants died
• Part of the original design of the MIRIAD study
• Data only available for 301 women
• Due to IRB issues
• Some records incomplete

26
Conclusions

• Even though domestic violence was not associated
  with HIV status, many important social and
  demographic factors are.
• Additional targeted interventions are needed to
  identify and protect at risk women and to lessen
  the potentially devastating long term effects of
  domestic violence.

27
Acknowledgments

• Mayris Webber DrPH
• Peter Bernstein MD MPH
• Irwin Merkatz MD
• MIRIAD staff and team
• Especially Kelly Monaghan
• MIRIAD study participants

• MIRIAD funded by CDC
• This analysis funded through Bronx CREED