Intimate Partner Violence Health Care Utilization and Insurance

Intimate Partner Violence, Health Care Utilization and Insurance Status: Results from a Large Population-Based survey Kenneth J. Steinman, Ph. D, MPH Amy E. Bonomi, Ph. D, MPH Academy on Violence & Abuse April 25, 2009 Supported by a grant from the Ohio Family Health Survey

Background • IPV (intimate partner violence) is associated with HCU (health care utilization) (Coker et al. , 2004; Helweg-Larsen, 2003; Rivara et al, 2007) • Few population-based studies • Little consideration of insurance status

Goals • Assess IPV-HCU association in a population-based sample • Does association vary by insurance status?

Ohio Family Health Survey • • • September 2008 -January 2009 n=50, 944 (23, 083 women <65) Computer-Assisted Telephone Interviews Random Digit Dialing Over-sampling of ethnic minorities, select counties • Representative of Ohio adults & households • Response rate: 47%

Dependent Variables • During the past 12 months, how many times were you a patient in… – an urgent care center? – a hospital emergency room? – a hospital overnight?

IPV measure • During the past 12 months, how many times, if any, has anyone hit, slapped, pushed, kicked or physically hurt you? • Think about the time of the most recent incident involving a person or persons who hit, slapped, pushed, kicked or physically hurt you. What was that person’s relationship to you? (open-ended)

classifying a case of intimate partner violence How many times…? 0 times Not a case 1+ times What was that person’s relationship to you? 04 Male/Female first date 05 Someone you were dating 06 Former boyfriend/girlfriend 07 Current boyfriend/girlfriend or fiancé 08 Spouse or live-in partner 09 Former spouse or live-in partner intimate partner violence 01 Stranger 02 Coworker 03 Professional caretaker 10 S/he is my Child 11 S/he is my Stepchild 12 Another family member 13 Acquaintance/friend 97 OTHER 98 DK 99 REFUSED Other violence

Other Variables • Age, ethnicity, region • Socioeconomic status – Income, education, home ownership • Insurance status – Uninsured; Medicaid; Employer-based; Other

Data Analysis • Bivariate association – Covariates with IPV – IPV with dependent variables • Generalized Linear Model – Poisson distribution, log link – Controlling for age, ethnicity, SES • Weighted data; complex survey design

IPV prevalence n Weighted % 95%CI 351 1. 9 [1. 6 -2. 2] 18 -24 75 5. 3 [4. 0 -6. 9] 25 -34 93 2. 1 [1. 5 -2. 8] 35 -44 85 1. 8 [1. 3 -2. 5] 45 -54 72 1. 1 [0. 8 -1. 5] 55 -64 26 0. 4 [0. 2 -0. 8] white 259 1. 7 [1. 4 -2. 1] black 70 3. 3 [2. 4 -4. 5] Hispanic 22 2. 7 [1. 5 -4. 7] Asian 0 --* Total Age Race/Ethnicity

IPV prevalence (2) n Weighted % 95%CI <100% 145 4. 3 [3. 4 -5. 5] 100 -300% 133 2. 1 [1. 7 -2. 7] >300% 73 0. 8 [0. 5 -1. 1] owns home 173 1. 3 [1. 6 -2. 2] does not own home 176 3. 5 [2. 8 -4. 3] Income as % of federal poverty level Home ownership

IPV prevalence (3) n Weighted % 95%CI < High school 44 3. 8 [3. 4 -5. 5] High school graduate 259 1. 3 [1. 7 -2. 7] College Graduate 48 0. 2 [0. 5 -1. 1] Education

IPV prevalence (4) n Weighted % 95%CI Insurance type Medicaid 127 5. 2 [4. 1 -6. 7] employer-based 105 0. 7 [0. 5 -0. 9] uninsured 83 4. 5 [3. 4 -6. 0] Other a 36 1. 8 [1. 1 -2. 9] a Other insurance types include: Medicare only, directly purchased, unknown

Association of IPV with Health Care Utilization: a Prevalence Ratio (unadjusted) 95% CI Prevalence Ratio (adjusted)a 95% CI Urgent care 2. 1 [1. 7 -2. 7] 1. 6 [1. 3 -2. 1] Emergency room 2. 2 [1. 9 -2. 6] 1. 5 [1. 3 -1. 7] Hospital Admission 1. 7 [1. 3 -2. 2] 1. 3 [1. 0 -1. 7] Adjusted for age, ethnicity, income, education, home ownership, insurance status

Association of IPV with Hospital Admission: Variation by Insurance Status Uninsured Medicare Employer. Based (n=2, 924) (n=2, 998) (n=13, 763) Other (n=2, 395) PRR 95% CI Urgent care 2. 3 [1. 5 -3. 5] 1. 4 [1. 0 -1. 9] 1. 3 [0. 8 -2. 4] 1. 9 [1. 0 -3. 7] Emergency room 1. 7 [1. 3 -2. 3] 1. 4 [1. 1 -1. 7] 1. 4 [0. 9 -2. 2] 1. 9 [1. 1 -3. 2] Hospital admission 1. 2 [0. 6 -2. 4] 1. 1 [0. 8 -1. 6] 1. 1 [0. 6 -2. 0] 3. 0 [1. 7 -5. 2] PRR=Prevalence Ratio (adjusted for age, ethnicity, income, education, home ownership)

Discussion • IPV – HCU association holds in general population • IPV prevalence varies by insurance status • IPV – HCU association may vary by insurance status – Smaller effects for among women with employer-based insurance (EBI)?

Implications • Urgent care may be important source of care • EBI samples may underestimate IPV-HCU association • Include insurance as a component of SES • Useful alternative measurement approach

Thank you