Improving Domestic Violence Screening and Follow-Up in a Community Health Clinic

Abstract

Purpose:

To implement an improved DV screening and follow-up protocol at a suburban community health clinic that is consistently used by staff with appropriate female patients as well as to assess provider/staff barriers to this.

Methods:

The project was completed at a community health clinic and included 2 presentations: the first to outline current practices as well as the intervention plan at the pre-intervention time point and then to review results of the intervention and elicit feedback post intervention. Provider/staff barrier surveys were completed at both time points. The intervention consisted of completion of the HITS DV screening tool on all appropriate female patients and a system-level algorithm-based follow-up care plan for positive cases.

Results:

Eligible patient screening increased by 3.1% and patients screening positive increased from 0 to 1 from baseline to intervention. Average scores on the barrier surveys improved for all questions and improved significantly for 3 of the questions. The algorithm implemented was utilized for the positive screening.

Conclusions:

Use of a DV screening tool, that has shown good concurrent and construct validity in the literature, a system-level algorithm for follow-up care and addressing provider/staff barriers to screening and follow-up are essential components of a successful DV screening and follow-up program.

Introduction

Background

Domestic Violence (DV)/Intimate Partner Violence (IPV) are used interchangeably and refer to acts of physical violence, sexual violence, stalking, and psychological aggression upon persons with whom the perpetrator has a relationship with. According to the Centers for Disease Control and Prevention ¹ approximately 1 in 4 women and 1 in 10 men experience IPV during their lifetime. Of note, IPV associated health-care costs are greater than $5.8 billion each year. ² Appropriate screening and follow-up of Domestic Violence/Intimate Partner Violence in women in the primary care setting is recommended by the USPSTF along with the American Congress of Obstetricians and Gynecologists (ACOG), the United States Department of Health and Human Services (HHS), and the Institute of Medicine (IOM). ³

Screening rates in primary care are low ranging from only 1.5 to 12%. ³ Black/African Americans, Hispanic/Latina, Native American/Alaska Native, and Asian American women, as well as women from impoverished backgrounds experience disproportionately higher rates of intimate partner violence according to an article by Stockman et al. ⁴ These statistics shows the urgent need for IPV prevention among women and the demand for a comprehensive follow-up system.

Problem

The clinic used in this study is a federally qualified health center (FQHC) that provides primary healthcare for patients from diverse ethnic backgrounds. A large majority of this patient population is Hispanic/Latina and/or impoverished. The current protocol at this clinic for screening for IPV/DV is 1 verbal question completed by the medical assistant and/or healthcare provider assessing whether the patient feels safe at home. This question is not consistently asked of each patient and is typically up to the discretion of the provider and MA. Other barriers to screening at this clinic include lack of provider education on follow-up for patients who screen positive for domestic violence. Also, of note, no specific training is provided to staff or providers on what resources are in the area for patients who are experiencing domestic violence. Given this lack of consistent protocol and barriers to screening for DV/IPV, it is likely many cases go unreported. This clinic is located in DuPage County, IL and in FY2017 there were 7 deaths associated with domestic violence in DuPage County, IL. ⁵ For comparison, there were 84 domestic violence related deaths in the entire state of Illinois between 2013 and 2014. ⁶

Available Knowledge

A literature review was completed to identify the best methods to improve DV/IPV screening in a community health clinic including type of screening tool used, how to address and overcome provider barriers to screening, and to examine for any adverse events related to screening. Current research suggests that screening all women of childbearing age for DV/IPV using a reliable and valid screening method is best practice and that addressing provider barriers to screening and utilizing a systems-based approach to screening and follow-up are the most effective methods to meet the above goals.

A thorough review of the literature was completed using CINAHL, Embase, and PubMed databases. Citations from the reference list of systematic reviews were also analyzed. The search terms used include the following: “domestic violence” OR “intimate partner violence” OR “partner abuse” OR “spousal abuse” OR “screen*” OR “loss to follow up AND adherence to follow up. “community health clinic” Articles were limited to English and limited to the past 13 years, more specifically, articles spanned from 2007 to 2016.

During the review of the literature, several important points emerged to achieve improved domestic/intimate partner violence screening. The main themes included that a chart prompt or educational intervention will increase DV screening, screening does increase positive reports of domestic/intimate partner violence, paper/computer tools, such as the Hurt, Insult, Threaten, and Scream (HITS) tool are as effective as face-to-face screening methods between patient and healthcare provider, there are either none or minimal adverse effects from screening, and that a systematic approach including addressing provider barriers, providing staff education and having appropriate referral systems in place for positive screenings is essential for any DV/IPV intervention.

Multiple studies showed that specifically screening for DV/IPV in women of child-bearing age does increase number of positive reports of this or that specific interventions do increase screening amount.^{7 –10} In the study by Ambuel et al., ⁷ the provider reported positive screenings, completed by chart auditing, increased significantly from baseline to post-implementation of a screening intervention (P < .02). The systematic review completed by Sprague et al. ¹⁰ reported that 27 of the 35 articles reviewed showed increased IPV disclosure after screening interventions were implemented.

The modality of screening for IPV/DV was considered as well during this literature review. According to O’Doherty et al. ⁹ the ability to detect for IPV/DV did not differ based on whether the screening was completed on paper/computer or face-to-face with a provider. This systematic review identified 4 research studies that reviewed face-to-face versus computer or paper-based screening methods and found that there was no difference in detection of IPV/DV between the different modalities in 3 of the 4 studies (OR 1.12, 95% CI 0.53 to 2.36). One of the 4 studies did favor the face-to-face modality but was subsequently removed from analysis due to high risk of bias.

Several specific DV screening tools were referred to in the literature and authors noted that in terms of a specific tool to use for screening, the Hurt, Insult, Threaten, and Scream (HITS) screening instrument is both sensitive and specific to screening for domestic/intimate partner violence. ¹¹ In fact, the study completed by Nelson et al. ¹¹ showed an 85% sensitivity and specificity for domestic/intimate partner violence in both the English and Spanish versions of the tool.

Multiple studies reviewed concluded that a systematic approach to screening and follow-up were essential aspects of a successful domestic violence screening/intervention.^{7,10,12 –15} The great majority of these articles were literature reviews that, in total, appraised hundreds of DV studies to determine this. Provider education on resources in the area as well as a DV advocate in clinic or administrative support improved outcomes by increasing screening amounts as well as increased amount of positive reports of DV. Increasing provider knowledge of DV resources when identifying a victim were essential components of these programs and addressed, as well as improved, the barrier of lack of provider knowledge of these resources.

There were 2 studies that reviewed adverse effects associated with screening and both found that no or minimal adverse effects, including discomfort with screening, loss of privacy, emotional distress, and concerns about future abuse, occurred from screening for DV/IPV.^9,11

Ultimately, screening for DV/IPV does not show major adverse outcomes, can be effectively completed by use of paper/computer tools including the HITS instrument, and requires a systematic approach with the support of staff and administration as well as knowledge of appropriate resources/referrals in order to be successful.

For the literature review completed on follow-up for DV, most studies were done in primary care/outpatient clinics (Alvarez et al., 2018; Clark et al., 2017)^9,16,21,22 and emergency departments (Wolff et al., 2017). ²³ In general, studies were focused on domestic violence screening measures and follow-up interventions for positive IPV cases.

This follow-up specific literature review highlights important ways to increase follow-up and appropriate referrals for women who disclose DV in diverse healthcare settings. Some articles examined the providers’ views on feasibility and acceptability of integrating a safety decision aid and the provider-specific level of preparedness, to prevent long-term health and social consequences among their patients (Alvarez et al., 2018; Clark et al., 2019; Williams et al., 2016).^21,16,24 Several other studies reviewed policies and procedures across health care facilities that yield empirical information on the impact of comprehensive DV response programs (Clark et al., 2019; Williams et al., 2016).^16,24 In conclusion, a systems-level response based on healthcare utilization through the use of flowsheets or algorithms is more effective than referral service alone for patients who screen positive for DV.

Lastly, a review of the existing DV literature on survivor specific preferences showed several overarching themes including that persons who disclose DV frequently need or request assistance in understanding the community resources available for support including housing, legal aid, and child-care options and that ideally the clinic and providers seen will have some understanding of and connection with resources in the area. ¹⁷ Additionally, the study by Dichter et al. ¹⁷ found that survivor preferences for support from providers included trust, privacy, self-determination and individual empowerment. In another study by Ferranti et al. ¹⁸ the health education needs identified as highest priority by female DV survivors were depression and self-esteem. Primary care providers and clinics are in an ideal position to provide health education and interventions on depression as well as to be familiar with resources to aid in improving self-esteem regardless of whether the patient chooses to disclose IPV or not and thus this particular intervention is not reliant on the need for positive DV screening or disclosure. Although not directly addressed in the study, these preferences present a future opportunity for further evaluation and considerations for further provider education that integrates survivor preferences into interventions.

Project Purpose

The purpose of this project was to implement a screening tool with high sensitivity and specificity, more specifically the Hurt, Insult, Threaten, and Scream (HITS) DV screening tool, with all appropriate patients who self-identify as female at a suburban community health clinic as well as to develop, and utilize a system-level algorithm for DV follow-up and to provide a local resource list for patient referral.

Methods

The project was initiated at a suburban community federally qualified health center and included 2 presentations to staff. The first presentation was to outline the current practices, need for improved DV screening and follow-up of positive DV cases, and outline the plan for the intervention. This was completed at the pre-intervention time point. The second presentation to staff was to review results of the intervention and allow for feedback and was completed at the post-intervention time point. A provider/staff barrier survey was completed at both the pre and post-intervention time points and stakeholder feedback was elicited. Four providers at this clinic as well as all medical assistant staff volunteered to be a part of this study.

The intervention itself consisted of completion of a paper version of the Hurt, Insult, Threaten, and Scream (HITS) DV screening tool on all appropriate unaccompanied female patients ages 18 to 60 years from August to October 2020. Appropriate follow-up was completed for all patients that screened positive for DV/IPV using a system-level algorithm. A flowsheet was built into the EPIC EMR system for input of scores on the HITS tool and several short cuts for documentation were created for use by providers when charting the details associated with the domestic violence and of the follow-up plan. When completing the flowsheet, if the HITS score input was >10, the staff entering the score would receive an automatic notification that the patient screened positive for domestic violence and needed to make a follow-up appointment within 2 weeks of the initial visit. The clinic staff was also educated that if any patient screened positively on the HITS tool, they would need to set up a follow-up visit with the provider within 2 weeks.

For those patients that did screen positive, a community resource list was provided in the form of a chap-stick so that this would not be readily identified by a perpetrator. An optional system-level algorithm was also given to providers to follow-up for those patients that screened positive.

Baseline domestic violence screening data was collected specifically for the 4 providers who were a part of this study from the corresponding time points of August to October 2019 and compared to the data from the intervention time of August to October 2020.

This project was classified as Quality Improvement and approved as such by UIC IRB.

Instruments

The Hurt, Insult, Threaten, and Scream (HITS) tool was used to screen for DV/IPV in this study. This DV/IPV screening instrument was developed by Sherin et al. ¹⁹ Author approval was given for use of this tool in this study. The HITS tool, including both the English and Spanish versions, has shown 85% sensitivity and specificity for screening for domestic/intimate partner violence. ¹¹ This tool has also previously shown internal reliability and concurrent validity in a study by Rabin et al. ²⁰ and high construct validity according to Sherin et al. ¹⁹ This instrument consists of 4 questions including how often does your partner physically hurt you, insult you or talk down to you, threaten you with harm, and scream or curse at you? This tool was chosen for the high level of sensitivity and specificity in both the English and the Spanish version as a large majority of the population served at this clinic are primarily Spanish speaking. Any score great than or equal to ten was considered a positive screening of DV. When a patient screened positive on this tool, a system-level algorithm, adapted and modified from Clark et al. ¹⁶ and Womenshealth.gov, was available to complete appropriate follow-up. This algorithm is listed in Appendix B.

A survey was given to 3 of the 4 providers and to the majority of medical assistants to assess barriers to DV/IPV screening and follow-up. This survey is listed below. This author was a provider in the study and was excluded from the barrier survey to prevent bias. This survey was completed both at baseline and post-intervention time points. This survey is listed in Appendix A and includes a numerical rating scale for staff to complete on a 5-point Likert scale for 12 of the 15 questions. The survey was distributed in paper format, was completed independently, and staff had the option of remaining anonymous.

Average scores were calculated on the pre-intervention and post-intervention barrier survey scores and compared via an unpaired 2 tailed t test using Microsoft Excel. All numerical questions improved from pre to post intervention time points (question #9 showed improvement by reduction in the numerical value) and 3 of the questions scores improved statistically significantly.

Patient chart audits were completed to collect DV screening data from August to October of 2019 (baseline data) and from August to October of 2020 (intervention data). This data was collected from the EPIC EMR charting system and the analysis included all eligible patients who self-identify as female between the ages of 18 to 60 years of age. Total eligible patients, total patients who screened positive for DV and total patients who screened negative were calculated for each time point. A patient who reported not feeling safe at home was considered positive for domestic violence for the baseline data and a patient who scored greater than or equal to ten on the HITS screening tool was considered positive for the intervention data. Data from each time point was then compared.

Results

As noted above, the average scores on the barrier survey numerical questions improved from baseline to intervention and the improvement was statistically significant for 3 of the questions (α = .05). These questions included “I am aware of the current screening protocol for IPV in my clinic (P = .04), I am comfortable with responding to a positive screening for IPV (P = .05), and there is a protocol/algorithm in place at my clinic to respond to positive IPV screenings (P = .04).” Please see Figure 1 for graphic of scores.

Figure 1.

Average provider/staff barrier scores pre and post-intervention.

Domestic violence screening data also showed improvement with both an increase in positive screenings and in total percentage of patients screened. The total amount of patients who screened positive for DV increased from 0 at baseline to 1 at intervention. Also, the percentage of patients who were eligible for screening but were not screened decreased from 73.1% of total eligible patients to 70% from baseline to intervention time points. Data is included in Tables 1 and 2 below.

Table 1.

Baseline Domestic Violence Screening Data.

Baseline screening	Sample (N)	Comments
Screened negative 2019	275	26.9% pts screened
Screened positive 2019	0
Not screened 2019	749	73.1% of total patients were not screened
Total	1,024

Table 2.

Intervention Domestic Violence Screening Data.

Intervention data	Sample (N)	Comments
Screened negative 2020	158	29.4% pts screened
Screened positive 2020	1
Not screened 2020	377	70% of total patients were not screened
Total	536

Of note, the number of eligible patients seen in clinic did decrease largely due to the global Covid pandemic. This clinic implemented telehealth visits shortly before the intervention began and these patients seen via telehealth were not included in the study. Also, 3 of the 4 study providers did have at least 1 week during the intervention period where they were working in a Covid testing site and thus were not seeing patients in clinic that week. This reduced the overall number of eligible patients.

Consistent positive verbal feedback was given by providers and staff on the availability and use of the system-level follow-up algorithm and provided resource list but level of satisfaction was not explicity measured. The improvements in the barrier surveys correspond with this provided information.

Discussion

The key findings of this study demonstrate that a paper based empirically validated (construct and concurrent validity) tool such as the HITS is easily implemented in the primary care setting and may result in higher levels of identification for patients experiencing domestic violence than prior utilized methods of screening. For this particular study, it is likely that the increase in positive screenings resulted from poor screening at baseline, to an established screening tool for the intervention. No significant challenges or discomfort from this tool was reported by patients or staff.

The improvement in barrier survey scores from pre to post-intervention time points suggest that a presentation to clinic staff to discuss the importance of domestic violence screening, proposed interventions including the provided algorithm and local resource list, and allowance for questions is an essential component to a successful DV screening and follow-up protocol.

Limitations

The major limitations to the study were due to the COVID pandemic. Three of the 4 providers that were part of the study were also out of the clinic for at least 1 week during the intervention time period and the number of in clinic patients was reduced from 2019 due to increased telehealth visits.

Additionally, the small sample size of the clinic and the difference in sample size from the pre-intervention to post-intervention barrier surveys are also limitations. Also, satisfaction of provider use of the HITS tool and TLC algorithm should have been measured as this could help predict likelihood of future use of these tools. Satisfaction of the patients with these tools would also be helpful to assess in future research.

Conclusions

Domestic violence screening is recommended by a myriad of healthcare related entities and yet still remains dismally low due to a variety of factors. Clearly there is still more work to be done to improve this in the primary care setting. The findings of this study support that the addressing of provider and staff barriers in primary care are essential components of a successful domestic violence screening and follow-up program. Further evaluation of satisfaction of the tools utilized would be important to evaluate in future research as well as integration of essential patient voices and preferences for screening and following up on positive screenings. Although the positive screenings did increase from 0 to 1, it is clear that an educational program that seeks to integrate vital patient preferences with DV screening and follow-up as well as provider preferences would be essential for a successful DV interventional program. It is imperative that any education of DV screening offered to providers also integrate how to appropriately follow-up on positive reports and that trainings are offered regularly to provide for updates as appropriate. Additionally, it is important that all medical clinics, especially primary care clinics, have DV resources available for all patients at all times as there are clear barriers to patients disclosing domestic violence such as personal safety, loss of a close relationship, child welfare involvement, and fear of judgment to name a few. How to best provide these resources for patients unwilling to disclose DV is another topic of necessary future research but as stated in the available knowledge section, there are certain considerations and interventions that providers can offer without the need for a positive DV disclosure such as treatment for depression and support for building self-esteem.

Additional future research on this topic should include multiple primary care clinics as well as all providers and medical assistants at these clinics. There is also need for further study with a larger and more diverse patient and staff sample to promote diversity and external validity as well as to assess for differences related to cultural priorities when screening for domestic violence.