1/19/2016 Jacq Jacquely lyn C. C. Cam Campbell ll, , PhD PhD, - - PDF document

1/19/2016 1

Jacq Jacquely lyn C.

• C. Cam

Campbell ll, , PhD PhD, RN, RN, FAA FAAN Jane net Su Sull lliv ivan W Wilson ilson, , PhD, R D, RN Jill T Ther eresa M esa Messing, M ng, MSW, W, P PhD Sh Sher eryl yll Brow l Brown, MPH H Beve Beverly Pat Patchell, PhD PhD, RN RN

Research Funded by NIJ #2008-WG-BX-0002

Research Team

Janet Wilson, PhD, RN, OU College of Nursing Dave Sargent, Maryland Coalition Against Domestic Violence Jacquelyn C. Campbell, PhD, RN, FAAN, Johns Hopkins School of Nursing Jill Theresa Messing, MSW, PhD, Arizona State University Sheryll Brown, MPH, Injury Prevention Service, Oklahoma State Department of Health Beverly Patchell, PhD, APRN, PMHCNS-BC, University of Utah College of Nursing  Approximately one-third of U.S. women have

experienced IPV in their lifetime (Black et al., 2011).

 IPV has significant consequences for victims,

including poor health and mental health outcomes, particularly for those who experience severe IPV (e.g., being “beaten up,” assaulted with a weapon) (Campbell, 2002).

 Of all violent crimes committed against women in

2010, 22% were perpetrated by a current of former intimate partner (Truman, 2011)

 Multiracial and American Indian/Alaskan Native

women have the highest lifetime IPV prevalence rates (54% and 46%, respectively) (NISVS 2010)

1/19/2016 2

 Femicide = the killing of women  40%-55% of adult femicide victims are killed

by a husband, boyfriend, partner or ex- partner (Campbell et al., 2003; Violence Policy Center (VPC), 2015).

 Physical IPV was reported to have preceded

homicide in 65-80% of intimate partner (IP) femicide cases (Campbell et al., 2003; Moracco, Runyon & Butts, 1998; Pataki, 1997).

 47% of women killed by partner in health care

system year before killed (Campbell 2002)

 Nearly half (49%) of women in Oklahoma (OK) have

experienced rape, physical violence, and/or stalking by an intimate partner in their lifetime, a higher proportion than in most states (Black et al., 2011).

 In Oklahoma, the past-year physical IPV prevalence

among multiracial and Native American women was roughly twice that of white women (2.3 and 1.8, respectively) (OWHS 2002)

 In 2011, Oklahoma had the third highest rate of

female homicides in the nation (VPC, 2013)

 Developed by MNADV (http://mnadv.org/lethality/)  Collaboration between the police department and

advocacy organization

 Brief intervention takes place at scene of a

domestic violence incident

 Police complete their usual intervention  Police officer asks victim 11 questions on the

Lethality Screen (based on Danger Assessment) to determine if victim is at high risk for homicide

 If at high risk, police officer tells victim that is

high risk, calls local DV advocacy organization &

• ffers phone to victim for safety planning – so

voice to voice contact can be made if she wants

1/19/2016 3

 There are two short video clips in the

presentation that were filmed in OK after the research was completed when the LAP was being implemented statewide

 They were filmed by an ABC affiliate there

during a “ride along” – all participants provided written agreement to the station to be filmed and the clips were aired and archived by the station on its website and could be viewed by the public there

 There are two different victims in the video –

there is one clip here and toward the end of the presentation

Lethality Screen

1/19/2016 4

 If the victim is at high risk, a collaborating

domestic violence service provider is called

 Victim chooses whether she wants to speak on

the phone with DV advocate

 Advocate/victim discuss safety plan; police

assists with any immediate safety actions (e.g., transport to shelter)

 When the victim comes in for services, the

advocate does the full Danger Assessment (with weighted scoring) and uses this information for more in depth safety planning www.dangerassessment.org

 Field trial (real world) funded by the National

Institute of Justice (#2008-WG-BX-0002)

 7 sites in Oklahoma, includes police

departments + collaborating domestic violence service providers

 Hypotheses:

 The LAP increases safety behaviors (e.g., safety planning, shelter, etc.)  The LAP decreases the severity and / or frequency of violence

Cimarron Texas Beaver Harper Ellis Roger Mills Beckham Woodward Woods Alfalfa Major Dewey Custer Washita Greer Kiowa Blaine Caddo Grant Garfield Kingfisher Kay Noble Payne Logan Canadian Oklahoma Cleveland Grady Pawnee Osage Creek Lincoln McClain Jackson Tillman Comanche Cotton Stephens Jefferson Garvin Carter Murray Love Pontotoc Johnston Coal Atoka Bryan Hughes Okfuskee Pittsburg McIntosh Pushmataha Choctaw

McCurtain

Le Flore Latimer Haskell Sequoyah Muskogee Tulsa Wagoner Cherokee Adair Ottawa Washington Rogers Nowata Craig Mayes Marshall Harmon Delaware Okmulgee Seminole Potta- watomie

More than 15 Deaths 7-15 Deaths 1-6 Deaths No Deaths *Includes all victims (males, females, and bystanders) killed in intimate partner homicide incidents from 1999-2007

Lo Location o

• f OK-

OK-LA Stu Study Sit Sites an and Conce ncentra tratio ion o n of Intimate P mate Partner rtner Homi mici cides des from from 1999- 1999-2008 2008

1/19/2016 5

Phase 1 – Comparison Group

• Police respond to domestic violence

incident as usual.

• Recruit participants to the study.

Phase 2 – Intervention

• Police complete normal intervention
• Use Lethality Screen
• Phone local domestic violence provider

if victim screens as high violence

• Recruit participants to the study.

The OK-L The OK-LA Study: Study:

Qua Quasi-Exp i-Experimental NOT Ra rimental NOT Randomiz ndomized ed

Two structured telephone interviews conducted approximately 7 months apart. Participants were asked questions about:

 Demographic and relationship information  The violence (CTS) that they had experienced

(prior to interview #1, between interviews #1 & #2)

 Risk of homicide on the Danger Assessment  Protective actions taken (prior to interview

#1, immediately after the intervention, between interviews #1 & #2)

1,137 women referred Comparison Group: Jan 2009 – Oct 2010 604 eligible referrals contacted 486 (42%) unable to be contacted 47 ineligible 440 participated in baseline interview 164 (27%) refused to participate 433 women in the comparison group 7 duplicates removed 342 screened in as high risk 91 screened out (low risk – 79% high risk) 212 participated in follow-up interview 130 unable to re-contact (62% Retention) 2,022 women referred Intervention Group: Oct 2010 – Feb 2013 938 eligible referrals contacted 657 participated in baseline interview 648 women in the intervention group 538 screened in as high risk 347 spoke with a hotline counselor 1041 (51%) unable to be contacted 43 ineligible 281 (30%) refused to participate 9 duplicates removed 110 screened out (low risk

• 83% high risk)

145 unable to re-contact (59% retention) 202 participated in follow-up interview 191 (35.5%) did not speak with a hotline counselor

Recruitment & Retention

1/19/2016 6

Study Samp Study Sample le: Demograp : Demographi hics cs

Var Variable le Val Values Compa mparis ison

• n Gr

Group N (%)/M N (%)/Mean an (SD) (SD) In Intervention G Grou

• up

N (%)/M N (%)/Mean an (SD) (SD) Age Years 32.78 (9.758) 32.26 (10.130) Race/Ethnicity White African American Native American Latina Multiracial Other 141 (41.35%) 107 (24.83%) 36 (10.56%) 22 (6.45%) 29 (8.50%) 6 (1.76%) 147 (44.28%) 91 (26.61%) 31 (9.34%) 31 (9.34%) 23 (6.93%) 9 (2.71%) Born Outside the U.S.* Yes 8 (2.35%) 19 (5.58%) Education No HS degree HS degree/higher 73 (21.35%) 269 (78.65%) 95 (27.38%) 252 (72.62%) Employment Full/Part Time Neither 146 (42.69%) 196 (57.31%) 133 (38.33%) 214 (61.67%) Pregnant Yes 26 (7.93%) 23(7.06%) *Significant differences between groups

Study Samp Study Sample le: Relat : Relationshi

• nship Charact

Characteri risti stics

Var Variable le Val Values Compa mparis ison

• n Gr

Group N (%)/M N (%)/Mean an (SD) (SD) In Intervention G Grou

• up

N (%)/M N (%)/Mean an (SD) (SD) Currently living with partner 58 (16.96%) 58 (16.71%) Marital Status* Single Married Separated Divorced 197 (58.28%) 77 (22.78%) 17 (5.03%) 47 (13.91%) 221 (64.62%) 83 (24.27%) 17 (4.97%) 21 (6.14%) Children in household Yes 232 (67.84%) 219 (63.11%) Children with partner Yes 156 (45.61%) 159 (45.82%) *Significant differences between groups

Nearl arly 90% 90% of the

• f the sam

sample rep reporte rted severe or severe or nea near-lethal vio

• lethal violen

ence ce

• Used a knife or gun on you/threatened you with a

weapon

• Punched you/hit you with something that could

hurt

• Strangled/tried to strangle you
• Beat you up
• Burned or scalded you on purpose
• Kicked you
• Did anything that might have killed you/nearly

killed you

• Tried to kill you

1/19/2016 7

Study Samp Study Sample le: Injury from IPV : Injury from IPV

Var Variable le Yes, Ev Yes, Ever er Ti Time 1, me 1, %(n) Physical pain that hurt the next day 84.3% (581) Sprain, bruise or cut 83.9% (578) Blacke ked o

• ut f

t from

• m being

g hi hit o t on the the hea head 20.3% (140) 20.3% (140) Broken bone 12.8% (88) Permanent impairment of disability 11.2% (77)) Internal injuries to vital organs 5.7% (39) Lo Lost st co cons nsci ciou

• usness due to

due to str strangulation 23.8% (164) 23.8% (164) Lo Lost st co cons nsci ciou

• usness due to

due to head head inju juries 3.8% 3.8% (26) (26) Lost enough blood to need a transfusion 1.0% (7) Needed surgery 4.2% (29) Hospitalized or in rehab > 4 days 4.2% (29)

Ma Main Findi in Findings gs – Violence iolence

Depende ndent V t Variabl able: : Weight hted F ed Freque uency b ncy by S Severi rity C ty CTS-2 S 2 Score Inde depe penden ent Var Variable les Indica ndicator

• r

Coef efficien ent (95% (95% CI) p-v

• valu

alue In Intervention G Grou

• up

Yes

• 14.71
• 14.71 (-28.60 to -0.81)

Significantly lower level of violence at follow up if got LAP (15 points = beating up X2 + slapping X2 + punching X1) p=.038 Dang Danger A Asse ssessme ment Cat Category Ordinal (0-3)

• 23.10
• 23.10 (-29.78 to -16.43)

Significantly lower level of violence at follow up if higher danger p=.000 Mar Marita tal S l Statu atus, Immig Immigration st stat atus, t , time me bt btw basel w baseline & & foll follow- up al up all no not t si signi gnificant Demographics made no difference Fi Fit S t Statis atistic tics: : F(7,3 F(7,397)= = 8.29, p<.001 8.29, p<.001, Pseu Pseudo R R2=.1121 121 – – (Fit f fairly g good)

0.00% 5.00% 10.00% 15.00% 20.00% 25.00% 30.00% 35.00% Received Services related to DV (n=681) Removed or hid partner's weapons (n = 689) Intervention Group (n=347) Comparison Group (n=342)

N=75

Conditional OR = 1.79* (CI = 1.25-2.56)

N=27

N=13

*p<.05 **p<.01

Women in intervention group 1.8 times more likely to immediately receive DV service And about 2 ½ times more likely to have removed or hidden partner’s weapons, controlling for demographic differences N=106

Conditional OR = 2.48** (CI = 1.14-5.37)

1/19/2016 8

0% 10% 20% 30% 40% 50% 60% Established code w/family friends Applied for PO Obtained something to protect self Engaged in other protective actions Intervention Group (n=202) Comparison Group (n=212) Conditional OR =1.63* Conditional OR = 1.64* Conditional OR = 2.17**

Women in intervention group 1.6 times more likely to establish code with family, friends, & apply for protective order, 2.2 times more likely to obtain something to protect self, 1.5 times more likely to be engaged in other protective actions than those in comparison group

*p<.05

Conditional OR = 1.54* 0.00% 5.00% 10.00% 15.00% 20.00% 25.00% 30.00% 35.00% 40.00% 45.00% 50.00% Received OP Go somewhere partner can't find Received medical care due to violence Partner went somewhere he could not see you Intervention Group (n=202) Comparison Group (n=212) Conditional OR = 1.88* Conditional OR = 2.53** *p<.05; **p<.001)

Women in intervention group about 1.6 times more likely to obtain protective order & go somewhere partner could not find her, 1.9 times more likely to obtain medical care due to violence, & partner 2.5 times more likely to go somewhere he could not see her

Conditional OR =1.59* Conditional OR = 1.61*

• The Lethality Assessment Program
• Decreased women’s violent victimization
• Increased immediate protective actions
• Increased protective actions at 7 months follow-up

 There do not appear to be differences in

effectiveness across ethnicity (sample sizes per ethnic group too small to test)

 The intervention is “supported” on the

effectiveness dimension and shows “promising direction” on the external validity dimension of the Continuum of Evidence Effectiveness (Puddy & Wilkins, 2011).

1/19/2016 9

 Generalizability: Oklahoma may not be

representative

 Selection bias: Police officers did not refer

everyone, women who chose to participate may not be representative

 Attrition: nearly 40% did not participate in the

follow-up interview, women who were employed or with higher levels of education were more likely to complete follow-up

 Historical comparison group: Something may

have happened between comparison and intervention phases

 49 women went into shelter  178 women accessed protective order

assistance/advocacy/safety planning

 5 women accessed legal services  58 women accessed counseling/crisis walk in  59 partners accessed Batterers Intervention

Program

 The LAP is an immediate and brief intervention

intended to educate women about their risk and encourage further, more specialized services

 Victims who call the police experienced

severe/near lethal violence across race & ethnicity

 Victims who call police are in a unique position to

seek assistance

 Collaboration plays a critical role in intervention

success – LAP training initiated with police AND DV Advocacy Organization

 The Lethality Screen in sensitive, not specific

– it casts a wide net – overestimates risk

1/19/2016 10

 85.3% of the sample had been strangled at

least once – 23% to unconsciousness

 High levels of violence and injury in the

sample as a whole

 After the intervention:

• 16.3% of the intervention group sought

healthcare services

• 10.4% of the comparison group sought healthcare

services

• Those that sought healthcare were actually

experiencing less violence (i.e., reaching out)

 Healthcare as an intervention opportunity  Ongoing assessment for danger & homicide

is important among women who call police

 A trained advocate / healthcare provider

should conduct the full Danger Assessment with women who screen in and seek services since more accurate than LAP screen – levels

• f danger

 Utilize a protocol similar to the LAP (or a

short form of the DA) in the Emergency Department

 Self determination is key – Victims have the right

to refuse the telephone call

 Training is important – Officers play a critical role

in making the intervention a success, advocates have critical information to give in a limited time frame

 Educational component is important – Women

have a right to know about the risk factors for homicide and their own risk

 Every intervention counts – Even when the victim

is not ready for the advocacy intervention, she knows that it is available if she needs it

1/19/2016 11

 In U.S., multiracial and American Indian/Alaskan Native

women have the highest lifetime IPV prevalence rates (54% and 46%, respectively) (NISVS 2010)

 Oklahoma’s female population has a higher proportion

• f American Indian females than the U.S. (9% vs. 1%)

 IPV & IP homicide rates are disproportionately high

among U.S. Native American females

 Few studies in DV field include Native American

women

 In Oklahoma, the past-year physical IPV prevalence

among multiracial and Native American women was roughly twice that of white women (2.3 and 1.8, respectively) (OWHS 2002)

 67 Native American women interviewed in

comparison / intervention groups – primarily Cherokee but other tribes also (e.g. Comanche, Apache)

 42 retained at follow-up (62.7% retention)

• 24 Native American women -intervention group
• 18 Native American women - comparison group

 In spite of efforts and collaboration with Tribal

DV Advocates, tribal police department recruited very few women into study & LAP implementation limited

 Due to low numbers, only some descriptive

results possible

1/19/2016 12

 No significant differences in demographics

with rest of sample

 No significant differences in severity and

frequency of violence or violent acts with rest

• f sample EXCEPT at T1
• Native American women significantly less likely to

report strangulation than larger group but still 74%! (vs. 84%)

• Native American women significantly higher mean

score on Danger Assessment than African American women (other groups in between – all high – in severe or extreme danger range)

• Native American women higher scores on PTSD

 74% of high danger (on LAP) Native American

women agreed to speak with counselor – not significantly different from rest

 The decrease in severity/frequency of

violence at T2 slightly less but not significantly different from rest

*There are no significant differences between groups but on adjusted logistic regression Native American women were LESS likely to remove a partner’s weapon (Cond OR=0.46) and MORE likely to seek DV services (Cond. OR=1.41) than rest (controlling for demographics)

10.4 31.2 10.7 39.1

5 10 15 20 25 30 35 40 45 Removed or hidden a partner's weapon Received domestic violence services

Native American Non-Native

1/19/2016 13

29.2 22.9 20.8 20.8 12.5 25 20.8 27.1 27.8 21.4 24.4 19.7 9 26.1 24.1 26.4

5 10 15 20 25 30 35 Established a code with family and friends Obtained something to protect yourself Applied for an

• rder of

protection Received an

• rder of

protection Seen a doctor

• r a nurse due

to injuries from violence Taken any

• ther

protective actions Went somewhere your partner couldn't find you Partner went somewhere Native American Non-Native

*There are no significant differences between Native American women and rest but on adjusted logistic regression, notable differences were that Native American women were less likely to go somewhere he couldn’t find her (or he go) & less likely to get something to protect themselves but more likely to do other protective actions  Tailor research recruitment strategies for Native

American women in collaboration with others

 Important to examine implementation and

results in collaboration with Native American groups even if only descriptively so they can use – being done but challenging

 Modifications for different tribes – need to tailor

LAP protocol and training for tribal police

 Dissemination of information – how to do for

tribal communities

 Authored by Representative Floyd and

Senator Holt

 Passed in the Oklahoma House/Senate May,

2014,

 Amended a Victim’s rights bill for police

• fficers to ask the 11 items in the lethality

assessment and, depending on answers, call collaborating advocacy agencies

 Effective November 1,2014  How is it being implemented - appropriately

for tribal communities?

1/19/2016 14



Messing, J. T., Campbell, J. C., Webster, D. W., Wilson, J., Brown, S. Patchell, B. (2015). The Oklahoma Lethality Assessment Study: A quasi-experimental evaluation of the Lethality Assessment Program. Social Service Review, 89(3), 499–530. http://doi.org/10.1086/683194



Messing, J.T., Campbell, J.C., Brown, S., Patchell, B., Androff, D.K., Wilson, J.S. (2014). The association between protective actions and homicide risk: Findings from the Oklahoma Lethality Assessment Study. Violence & Victims, 29(4), 543- 563.



Wilson, J.S., Messing, J, West, J., Brown, S., Patchell, B., & Campbell, J. (2011), Factors related to posttraumatic stress symptoms in women experiencing police-involved intimate partner violence, Advances in Nursing Science, 34(3), pp14- 28.



Messing, J, Cimino, A., Campbell, J., Brown, S., Patchell, B., & Wilson, JS (2011) Collaborating with Police Departments: Recruitment in the Oklahoma Lethality Assessment (OK-LA) Study, Violence Against Women, January, 17(2), pp 163- 176.

Oklahoma Lethality Assessment Study