A widely-used intervention for victims of domestic violence is safety planning. Murray and Graves (2012) described a safety plan as:

A personalized, detailed document that outlines clear and specific safety strategies that a battering victim can use to promote his/her safety across a wide range of situations. Fundamental to the creation of an appropriate safety plan is a collaborative process to develop it between the client and the professional. (p. 95)

Although safety planning is widespread, research is limited. In particular, a need remains for research that identifies best practices and informs evaluation studies. In this article, we present a review of literature describing safety planning practices and research. Next, we describe the methodology, findings, and implications of a focus group study that identified domestic violence service providers’ perspectives toward safety planning. The purposes of this study were to learn about common approaches to safety planning, identify strengths and limitations to these approaches, and determine future directions for improving research and practice.

Literature Review

According to Campbell (2002), safety planning presents “an opportunity for the abused woman to gain information in order to strategize her responses” (p. 129). The process of safety planning typically includes defining what safety means to the client, viewing the client as the expert in his or her safety, and holding a dialogue through which the client can discuss his or her safety concerns with a trained professional (Melbin 2010). The result of this process should be a detailed, personalized plan that provides specific strategies to help the client make decisions that will promote his or her safety in the face of an abusive situation (Kress et al. 2008).

Recommended Practices for Safety Planning

While safety planning is one of the most widespread interventions with victims of domestic violence, common practices vary widely from site to site, ranging from the simple provision of informational materials or a basic escape plan to a more comprehensive approach that addresses a broad range of the victim’s short- and long-term needs (Waugh and Bonner 2002). This latter format, a more interactive and comprehensive dialogue, is the most recommended in the existing literature. As part of this process, safety plans should be developed following a thorough assessment (Hardesty and Campbell 2004). It is important to respect victims’ wishes regarding the extent to which they want to discuss safety planning (Curry et al. 2006). However, a more proactive approach may be warranted if a professional believes a victim is in imminent danger (Curry et al. 2006). Safety planning should be done through an interactive process that engages the professional and the client in identifying general and specific strategies to promote the client’s safety (Campbell 2002; Melbin 2010; Murray and Graves 2012), and this process should involve a conversation that empowers and promotes the client’s autonomy (Campbell 2002).

Safety planning often takes the form of providing informational documents to victims (Bledsoe et al. 2004; Curry et al. 2006). These documents typically include contact information for local and national resources, such as local domestic violence, legal, law enforcement, and social agencies (Campbell 2002; Curry et al. 2006; Kress et al. 2008, 2012; Lindhorst et al. 2005; Murray and Graves 2012). Professionals can educate clients regarding specific safety strategies and resources. Some tips include the following: (a) have the victim keep cash and copies of key documents readily and securely available (Campbell 2002; Glass et al. 2010; McFarlane et al. 2004); (b) have supplies for children on hand (Campbell 2002); (c) gather contact information for key people (Campbell 2002); (d) make arrangements for pets (Faver and Strand 2003); (e) remove weapons from the home (Glass et al. 2010; McFarlane et al. 2004); (f.) discuss the safety plan with a friend or family member (Glass et al. 2010); (g) increase security in the home, such as by changing locks and installing a security system (Hoyle 2008); (h) acquire a new mobile phone (Hoyle 2008); (i) create an escape plan (Kress et al. 2008); (j) ask neighbors to call police if violence occurs (Kress et al. 2008) or have a code word or sign for neighbors to alert them of the need for help (McFarlane et al. 2004); (k) identify safe rooms in the house (Kress et al. 2008); (l) address transportation and communication needs (Kress et al. 2012); (m) address workplace safety (Smock 2003); and (n) include strategies to increase the safety of any involved children (Kolar and Davey 2007).

The end result of the safety planning process typically is a document with a personalized plan for the unique needs of the victim (Kress et al. 2008; Murray and Graves 2012). The plan should be simple; realistic; and include specific, behavioral strategies (Kolar and Davey 2007; Murray and Graves 2012). A safety plan should not be overly prescriptive, but it should address the client’s unique context and characteristics (Campbell 2002). The safety planning process also may involve the consideration of the safety risks for involved professionals, such as safety when leaving and entering work (Kolar and Davey 2007). Ideally, safety plans address multiple forms of safety (not just physical), identify and strategize how best to address possible barriers, account for the coping strategies the victim uses already, and can be adaptable to new situations (Lindhorst et al. 2005).

Existing Research Evaluating Safety Planning

Research examining the impact of safety planning is limited. Some case studies exist in the literature to illustrate applications of safety planning (e.g., Kress et al. 2012; Lindhorst et al. 2005). Beyond case studies, there have been very few rigorous outcome evaluations of safety planning. However, the existing studies have suggested that these interventions hold promise for promoting victims’ safety. For example, McFarlane et al. (1998) studied a safety planning intervention used with pregnant women who had been abused. Over the 21 months in which the researchers followed participants, the participants demonstrated increased use of nearly all safety behaviors. In a follow-up study, McFarlane et al. (2004) studied a telephone-based safety planning intervention, which involved a series of phone calls to educate the participating women on safety planning skills. In this study’s treatment group, the women who received this intervention were found to utilize a greater number of safety promoting behaviors with greater frequency than women who did not receive the intervention.

More recently, Kendall et al. (2009) studied a brief emergency department-based domestic violence safety planning intervention, which included advocacy counseling and referrals to local resources. Patients who were assessed to be at risk worked with an advocacy counselor to develop a brief safety plan with at least five individualized tips they could use to increase their safety. Following the intervention, nearly all patients (over 96 %) who were able to be surveyed reported feeling safer, and about one-half had implemented some aspects of their safety plans.

Two technology-based safety planning approaches have also been studied. Oschwald et al. (2009) examined a computer-based safety planning intervention, the Safer and Stronger Program (SSP). This program was designed specifically for women with disabilities and hearing impairments, and it provides information about abuse, safety planning, and community resources. A relatively informal program evaluation showed promising results, including client feedback that suggested the program increased participants’ knowledge and met their needs. The women who completed the program demonstrated a generally high level of satisfaction and noted the potential benefit of being able to report abuse via a computer rather than sharing abuse experiences with another person.

Additionally, Glass et al. (2010) developed and tested a computer program, Safety Aid, to help women create a personalized safety plan for domestic violence. Safety Aid asks users to report their background characteristics, current safety resources and behaviors they use, and their access to resources, such as a safe place to escape and emergency money. After completing the program, users receive a printout that evaluates their safety needs, safety priorities, and current level of danger. Glass et al. evaluated the program with a sample of 90 participants, 58 % of whom reported domestic violence in the extreme risk range. These participants demonstrated feeling more supported in their safety decisions and having less conflict about their decisions after they completed the program.

Limitations of Current Safety Planning Practices and Research

Overall, there is evidence that safety planning can promote the increased safety of domestic violence victims. However, more research is needed in order to understand the effectiveness of safety planning across different professional settings (Kress et al. 2012). In addition, there is a need for research that examines the perceptions of the involved professionals and clients and the specific processes within safety planning that promote safety.

The literature reviewed above suggests other limitations to current practices in safety planning. First, safety planning practices may vary widely across settings (Waugh and Bonner 2002), and clients may receive different levels of services depending on the availability of resources in their communities. Second, although many practical safety-related strategies have been discussed, there has been minimal attention on addressing clients’ emotional safety needs. For example, although clients may experience trauma-related symptoms (e.g., flashbacks, severe anxiety) when seeking help and recounting their experiences of abuse, recommended strategies that clients can use to address their emotional safety in these situations are not readily available. Third, safety planning practices typically focus more on immediate, crisis-related safety issues and less on longer-term issues that survivors face. Although the immediate, crisis-related safety risks can be high, the process of leaving an abusive relationship can be extended and unpredictable (Wuest and Merritt-Gray 2001), and different safety needs may arise at different points in this process.

A fourth weakness in current forms of safety planning is a lack of strategies that are appropriate for the most vulnerable populations of abused victims. Indeed, Nurius et al. (2003) found that women who view themselves as more vulnerable, more powerless, and more entrapped may demonstrate a lower ability to develop or carry out a safety plan. A fifth limitation is that although the needs of children should be addressed in safety planning (Kolar and Davey 2007; Waugh and Bonner 2002), typical approaches to safety planning are limited in their focus on children’s safety needs (Waugh and Bonner 2002).

Finally, a limitation that is inherent to the process of safety planning is that it is focused on the victim, but when dealing with a violent perpetrator, the victim is limited in what he or she can do to promote his or her own safety. Ultimately, a safety plan is no guarantee of safety (Campbell 2002), and, therefore, it is important that the limitations of safety plans are communicated to clients (Hardesty and Campbell 2004; Murray and Graves 2012). Because there is minimal evidence as to the overall effectiveness of safety planning and the specific strategies that often are included in safety plans, a need remains for increased attention to developing and testing effective strategies that will promote victim safety.

Methodology

Participants

This study involved a series of nine focus groups with domestic violence service providers (i.e., staff of battered women’s shelters, victim advocates, facilitators of batterer intervention programs, mental health professionals who provide direct services to clients impacted by domestic violence, and other allied professionals; Murray and Welch 2010). Focus groups were conducted on-site at domestic violence agencies across central North Carolina. We invited a diverse group of domestic violence agencies to participate, including those with and without shelters, those representing urban and rural communities, standalone agencies and those connected with other services (e.g., mental health agencies), and agencies with varying amounts of resources. All nine agencies that were invited to participate did so. A lead contact person at each agency enlisted participants using recruitment materials provided by the research team. Most (n = 7) agencies recruited only participants who worked in their agencies. The other two agencies invited local affiliated professionals to participate.

A total of 62 participants were in the nine focus groups. The smallest group had three participants while the largest had 11, with an average size of approximately seven people per group. Regarding gender, 54 participants were female, seven were male, and one did not report his or her gender. The average age of participants was 37.5 years (SD = 11.5). The average number of years in which participants had worked in their current jobs was 4.1 (SD = 4.5), and the average number of years that participants had worked in any job related to domestic violence was 7.4 (SD = 7.4). Table 1 contains a summary of participants’ other key demographic data.

Table 1 Focus group participants’ demographic characteristics
Full size table

Procedures

Each focus group was conducted by a lead facilitator, and a second research team member was an assistant and note-taker. The lead researcher provided a 1-h training session on focus group facilitation to all facilitators; all facilitators had a minimum of a master’s degree in counseling. The focus groups were based on a semi-structured interview guide. The facilitator asked follow-up questions for clarification and to prompt more in-depth discussion as appropriate. Topics covered in the interview guide included safety considerations for victims and children; agency policies, procedures, and forms for safety planning; the extent to which participants believed their agencies’ procedures address safety risks; lethality assessments; the manner in which children’s needs are addressed; differential safety planning for clients in and out of shelter; safety planning procedures at different stages of leaving an abusive relationship; staff training; participants’ views of the strengths and limitations of their agencies’ approach to safety planning; client characteristics that may not be addressed in safety planning; desired changes to agency approaches; and an open-ended question asking participants for any additional information they wished to provide. All focus groups were audio-recorded and later transcribed. Each focus group lasted one to two hours, and participants received refreshments as a token of thanks for their participation in the group.

Coding and Analyses

The content analysis procedures outlined by Stemler (2001) were used to analyze the transcripts. The coding unit was defined as a unique complete statement by each participant within the transcripts. We used an emergent coding strategy that built on some preconceived categories reflecting the topics in the interview guide. Beginning with the initial category list, each transcript was reviewed by two people, and each researcher made a list of ideas within the categories that were addressed within participants’ statements. At first, efforts were made to be exhaustive in identifying ideas. However, after the researchers developed an initial list of ideas, they reviewed their lists and combined duplicate/synonymous ideas into a single idea. Once all research team members completed their transcript reviews, the lead author consolidated the lists into a draft of the final master coding scheme, which was then finalized by the full research team. Next, the members completed a practice test of the coding process, which led to some additional revisions to the master coding scheme for clarification and to achieve adequate interrater reliability.

The final coding scheme included five categories: (a) safety risks for victims; (b) safety planning in a community context; (c) agency policies, procedures, and forms for safety planning, including service providers’ perceptions of them; (d) safety considerations and planning for children; and (e) a no code option for statements for which no other codes were applicable. Within the first four categories, secondary codes addressed sub-themes within the larger categories. Therefore, coders were instructed to give each statement a primary code and a secondary code. In cases in which the coders found that multiple codes applied to a single statement, the coder applied only one primary and one secondary code to best fit the main idea of that statement. In the current article, we describe only the results of the first three categories; the codes relating to safety planning for children are presented elsewhere (Horton et al. 2014), and we chose not to address statements identified as no code.

Each transcript was coded by three researchers, and a final consensus code was then identified. When all three coders agreed, the consensus code was the agreed-upon code. When two coders agreed, the consensus code was the code rated by the two coders. For statements on which all three coders listed different codes, the item was designated into the no code category and excluded from further data analyses, based on the assumption that the coding scheme could not be validly applied to these statements. All statements, including those designated in the no code category, were included in the calculation of the interrater reliability. Overall, 1863 statements were coded, for a total of 5589 codes across the three raters. Using Fleiss’ kappa, we calculated interrater reliability, which was found to be 0.099, for the primary codes only. The overall percentage of agreement was 82.91 %.

Results

Safety Risks for Victims

Table 2 provides the counts for the final consensus codes, both primary and secondary, for the 1863 statements that were coded. Participants differed on which safety risks they thought were most significant and raised particular concern about safety risks that may not be commonly known or considered. This section summarizes the themes identified through the data analysis process.

Table 2 Counts of statements falling within each primary and secondary code
Full size table

Lack of Personal Resources

A common theme related to lack of personal resources addressed the barriers victims face when planning for safety or planning to leave abusive relationships. These included finances, job skills, housing support, transportation, and lack of childcare. Personal resources were described as essential in starting a new life free from violence. The resources provided by agencies or shelters were not considered to be significant enough for long-term sustainment, thereby increasing safety risks for victims who find themselves without financial and other resources once agency resources are exhausted.

Perpetrators’ Use of Technology-Related Violence

Participants noted the growing increase in technology-related violence and stated that they are learning more about incorporating technology into safety planning. Perpetrators’ use of technology to perpetrate abuse included using tracking devices or GPS systems on victims’ cars and cell phones, using hidden cameras in victims’ homes, using computer software to monitor victims’ Internet browsing, and reviewing phone records. Service providers noted that perpetrators’ use of technology may impact victims’ help-seeking behaviors for fear of being tracked to local agencies. Victims also may fear making phone calls or browsing the Internet for resources if they believe they are being monitored by their perpetrator.

Risks Related to Leaving the Abusive Relationship

Participants discussed how basic needs for safety and housing are often at the forefront of victims’ minds as they leave an abusive relationship. They noted that the decision to leave an abusive relationship can escalate physical violence, as well as emotional and mental abuse. The participants talked about how vital it is for victims and service providers to act quickly when a victim decides to leave in order to determine the level of danger in the home, as well as the potential for stalking behavior by the perpetrator. In order to enable quick and effective action, participants emphasized the importance of having a plan for leaving in place, even as a contingency. In cases in which the victim is not ready to leave, they suggested talking with the victim about a safety plan for leaving so they are prepared if and when that decision is made. For example, one participant said:

When victims try to leave, that’s when things escalate. Especially when there’s not a lot of physical violence, I always kind of tell people, you never know what’s going to make that person snap, so just really having a plan and doing that and not just walking out [of the home or relationship] abruptly.

Mental Health and Substance Use

Participants raised concerns about the comorbidity of domestic violence, substance abuse, and mental health issues in a growing number of clientele. They noted that mental health and substance abuse resources were not readily available for victims, and they also expressed frustration about the lack of available training on how to serve this population. Participants identified comorbidity as an obstacle in receiving help and safety planning, noting that developmental disabilities, mental health issues, and substance use may lower victims’ abilities to be safe and to evaluate high-risk situations. As one participant said:

You’re meeting with these folks from this population, and you’re wearing the domestic violence hat when it’s very, very obvious that they need to be speaking with somebody who needs to first address their substance abuse and/or mental health issues... because those issues are direct factors in their vulnerability to be victimized.

Stalking

Participants noted that stalking is difficult to prove, and that petitioning for and receiving a restraining order can be complicated and confusing for victims. They pointed out that perpetrators may use children or other informants to find out where shelters are and continue the stalking, which not only compromises the safety of everyone at the shelter, but also impacts the victim’s ability to engage in treatment. Participants emphasized the importance of documenting stalking behavior in order to improve a victim’s chances of garnering legal protection.

Unique Cultural/Contextual Safety Risks

Participants noted several cultural and contextual safety risks for victims, particularly the difficulties facing undocumented immigrant victims who lack social security numbers and thus cannot apply for housing, benefits, child support, or secure employment. Participants reported that many undocumented victims stay in abusive relationships because of these factors and fear of deportation, making this population particularly vulnerable. Further, participants mentioned the safety concerns of individuals with various other issues, including developmental disabilities, deafness, and pregnancy. Such clients require additional resources, which can be expensive and require agency funds that are already lacking. Obstacles facing gay and lesbian individuals, particularly men, also were discussed, including a lack of shelter services for men and the threat of being “outed” by abusive partners.

Perpetrator Characteristics

Participants were particularly concerned with perpetrator manipulation and noted that perpetrators often have access to the very same resources as the victims, which may enable them to be aware of indicators that the victim is seeking help or planning to leave. They discussed circumstances in which the perpetrator has accompanied the victim to service provider agencies and the unique challenge that situation presents for providers. One participant mentioned a perpetrator who called the shelter daily and kept the victim on the phone, interfering with treatment on a regular basis. In addition, mental health and substance abuse issues affecting the perpetrator may contribute to a higher level of safety risk to the victim and his or her family.

Victims’ Perceptions

Service providers noted that certain perceptions and belief systems among victims may lead them to minimize their risk. Specifically, they discussed concerns about victims who lacked awareness of the existence or severity of abuse in their relationships. They described a pattern of desensitization in which victims have become so accustomed to the abusive dynamics of their relationships or from observing abuse in their families since an early age, that they are no longer able to recognize how toxic and dangerous those dynamics may be. Participants mentioned a sense of denial among victims, who fear retaliation if they publicly acknowledge the abuse or seek help; victims also may fear judgment, worrying that others will believe them to be crazy or think they are overreacting. Participants discussed a lack of self-esteem among victims for whom an environment of abuse is so familiar that it may be difficult to recognize, and some may even come to believe that they deserve the abuse and have no right to seek help. A resounding theme was that victims did not see their safety as a significant concern and that empowering these individuals to understand the dangers of being in an abusive relationship is one of the most instrumental roles professionals can play.

Dynamics of Abusive Relationships

Service providers reported that victims frequently indicate that they do not want to press charges against their perpetrators and often return to abusive situations due to their abusers’ controlling nature. Another related risk included the perpetrators’ use of control in the courtroom by being friendly with judges or police officers; and some perpetrators even have friends in the police department. The following participant quote illustrated these dynamics:

When the client come in [to the court room] and is sitting there and the offender comes in and sees an officer he knows, he goes over there and shakes hands with the officers, one of his drinking buddies, that’s really hard for [the victim].

Safety Planning in a Community Context

Some participants voiced that they received a great deal of support from their communities and had strong connections with community partners. Others, however, expressed concerns about community misperceptions and inadequate community resources. The themes discussed in this section address issues related to how the community context impacts the safety planning process.

Lack of Community Resources

Many participants expressed frustration over the lack of resources in their communities, such as legal aid or attorneys willing to work pro bono in order to assist victims. A number of participants linked recent economic decline to an increased need for community resources, and many reported that finding resources has become more difficult over the past few years. Concerns regarding a lack of community resources, especially transportation, were especially prevalent with participants representing agencies in rural areas.

Perceptions of Domestic Violence in the Community

Participants discussed how community perceptions of domestic violence impact victims. For example, several participants mentioned that local churches were encouraging couples experiencing domestic violence to reunify. Other participants observed that the community may underestimate the severity of the problem or the difficulties involved in leaving a relationship; many reported that community members blamed victims for continuing to stay in violent relationships or believed that a restraining order would be an easy fix to the problem. Some participants noted that there were fewer opportunities for collaboration due to decreased funding to state and county agencies. Other participants, however, described strong collaborations with community partners, which strengthened their abilities to assist victims. As one participant said, “When you can rely on other folks in the community, that really helps. …It’s so heartwarming that we know that.. …the community cares about the victims of domestic and sexual violence here.”

Barriers in the Community

One concern voiced by participants was that some professionals in their communities do not screen adequately or appropriately for domestic violence. A number of participants also mentioned barriers within the court system that prevent victims from being safe. For example, several participants discussed how defense attorneys use confusing tactics to pressure victims to drop charges or to coerce victims or their witnesses to leave court before a hearing. Many explained that going to court for a restraining order was an event that re-traumatized victims and that even once they received the restraining order, victims frequently found that it was not as powerful as they hoped it would be. Other participants stated that judges often appeared to be unsympathetic or irritated by victims.

Agency Policies, Procedures, and Forms for Safety Planning

Participants discussed different categories of safety planning policies, procedures, and forms, which varied depending on the type of agency and the client’s stage of leaving.

Safety a Primary Focus

Some participants remarked that the overarching goal of their agency was safety; thus, safety planning permeated every deliberation and interaction. As one provider stated, “whatever it takes to be safe” was a primary agency focus.

Safety Planning as an Individualized Conversation

While many participants reported using standard procedures when safety planning, the majority also spoke about individualized conversations with their clients in order to meet specific needs. They further noted that safety planning is a process (e.g., “It’s never completed, really. If you think about it, it’s a continuum, each conversation, usually there is something discussed”).

Empowerment Approach to Safety Planning

Many comments demonstrated the agencies’ empowerment-based, client-centered approach. As a participant said, “We try and make things more their choice. Because coming from the relationships they’ve come from, they haven’t always been given a choice. So we try to…work that way toward them most of the time.” This approach honors clients’ self-determination and right to make decisions.

Specific Tips

Many providers shared specific tips, strategies, and approaches they use during safety planning with victims, such as teaching victims to use insect (e.g., wasp) spray as a self-defense tool and providing victims with cell phones that call 911. The participants especially encouraged helping victims to trust their instincts. One provider said, “I reinforce and validate trust your instincts.”

Education and Resources

Providers described specific resources that they offer victims during the safety planning process, for example, the State’s Address Confidentiality program and the Survivor to Survivor videos (see http://www.survivortosurvivor.org). They also commented on the importance of educating victims about the red flags of abusive relationships. Participants noted the success of victim support groups that allow victims to hear each other’s stories.

Timing and Personnel Involved in Safety Planning

Participants commented on the timing of safety planning, commonly stating it should be done “all the time.” Providers also were clear that safety planning is generally conducted by all staff and occurs anytime they have contact with a victim, including before and after court and/or before leaving the agency. One participant said, “Sometimes safety planning gets initiated on the back end, like I see someone in court on criminal charge and... I’ll approach them in court, I tell them who I am, where I work and what services we provide.” Participants also noted that it is sometimes best to do safety planning when the victim is not in a crisis.

Emerging Safety Risks

Participants commented that part of the difficulty of safety planning is keeping up with constant changes in safety risks. As one said, “Every five seconds something’s changing…It’s hard to keep up with it.” Further, participants’ goals of staying a few steps ahead of the perpetrators were reflected in the following quote: “I think it’s always important to always kind of give them a little additional information, you know, what if, just being prepared, or two steps ahead of the perpetrator.”

Staffing Issues

Three themes emerged regarding the within-agency coordination of safety planning. First, participants emphasized the importance of frequent, collaborative communication. In fact, being able to work as a team was what made their agencies successful. Second, participants said that having interns from local universities brought energy and new information to the agency. Last, participants felt it was important for staff members to consider their own personal safety. (i.e., the safety of the staff members).

Training and Adherence and Safety Planning Documents

Most participants said that staff received significant training about safety planning. While each agency had slightly different training processes, nearly all said that staff were adherent to safety planning policies. Additionally, many participants discussed using specific documents and forms that helped with safety planning while also maintaining the need for individuality in each client’s plan. For example, some agencies provided standardized informational documents to clients during safety planning, and others completed interactive forms that could be tailored to the unique needs of each client.

Stages of Leaving and Lethality

Participants spoke about how the safety planning process differs for a client who does not recognize that he or she is experiencing abuse as compared to a client who is planning to leave the abuser. As one participant said, “Different phases bring about different lethality so [safety planning] has to change.” In addition, participants spoke about the importance of completing lethality assessments to identify risk factors for potential homicide as part of the safety planning process for all clients facing domestic violence, although the process for doing so varied across agencies.

Shelter

Many participants discussed the decreased risks when clients are in shelters compared to being in the home with the perpetrator. Participants suggested that this additional safety allowed clients to gain a clearer perspective of their situation. They reported upon extra safety measures that many shelters put in place, including secret locations, security staff on location, coded key-pads, bulletproof glass, and security cameras, and noted shelter-specific safety strategies that are reviewed with clients, such as not sharing the location of the shelter or the identity of other residents.

A number of participants discussed the process of addressing clients’ anxiety upon entering the shelter. As this is a significant time of transition for clients, meeting their basic needs was described as crucial to their sense of security and their ability to move forward in planning for their futures. Despite the safety the shelter provides, it was noted that there remains a risk for clients when leaving the premises (e.g., to work or school). To address this, one strategy identified by participants is to ensure that clients are dropped off and picked up at alternative locations (i.e., not the shelter) when arranging transportation with friends and family members. In addition to planning for safety when in the shelter, there was also a great deal of discussion around helping clients plan for safety upon leaving the shelter.

Cultural Considerations

Participants commented on several cultural issues that may need to be addressed during safety planning. For example, although some participants were glad to have bilingual professionals in their agencies, they also noted the need for interpreters for languages other than Spanish. Further, providers felt that the specific safety planning needs of diverse populations, such as men; lesbian, gay, bisexual, and transgender (LGBT) populations, the elderly, and undocumented victims, needed further attention, and these unique needs should be addressed in clients’ personalized safety plans.

Technology-Related Strategies

Many participants mentioned the use of technology in safety planning. While some participants detailed the helpfulness of technology (e.g., free mobile phones, software applications that address domestic violence, and online safety planning forms), others noted that new technological advances have the potential to be used against victims. Participants also described the challenges of keeping up with ever-advancing technologies, and some expressed that they did not feel prepared to adequately address technology-related risks in their safety planning processes.

Requirements for Safety Planning

In some agencies, safety planning was a requirement from funding agencies, such as grant funders. According to these funders’ guidelines, some agencies are required to track the number of times they provide safety planning. Several participants also commented that safety planning is one of their agencies’ primary goals and, therefore, is written into the agencies’ required policies.

Lack of Agency Resources

Participants reported that their agencies could use more financial and staff resources to increase their ability to engage in safety planning. Several participants stated that their agencies had waiting lists for victims to be able to receive services. Participants also commented that safety planning could be enhanced if they had access to discretionary funds that could help victims take some initial safety planning steps, such as purchasing a bus ticket, reserving a hotel room, or paying the first month’s rent for an apartment.

Effectiveness

Participants described their perceptions of the effectiveness of the safety planning approaches they use. Several agencies reported that they have conducted exit interviews with clients, and the clients have expressed that their safety planning needs were met; however, many participants wondered if they could more effectively be engaging in safety planning. It was noted that participants realized they were not always able to conduct exit interviews with every client. One participant explained,

Sometimes I think that’s just the nature of domestic violence, you sort of do the best you can while the client is with you and try to meet them where they are and help get them information and options and choices and what-not, but so often we don’t know what happens.

Further concerns regarding the effectiveness of agency safety planning included victims not following the safety plans, victims interpreting the safety planning questions differently, systematically checking to ensure safety planning has occurred, and having enough time to thoroughly complete a safety plan. Even with these concerns, participants said that any safety planning efforts, even those that are seemingly minor, may eventually be effective. As one participant reported:

I did have one lady that told me—I met with her in a hospital, went over safety planning. She agreed to come into the shelter. And, then, she changed her mind. And, then, went back home with the abuser. The next altercation that happened, when I got involved again she said, “I know I went back, but I remembered what you said about safety planning, and I was able to get out of the home and call for help.”

Discussion

In our study, we aimed to identify strategies to improve upon current practices in order to develop more comprehensive procedures that address the complex safety needs of domestic violence victims. As noted, we address the needs of victims’ children in a separate article (Horton et al. 2014). In this section, we review the limitations and major findings of this study, as well as implications for future research and practice.

Limitations

The findings of this study must be considered within the context of its limitations. First, participants came from one state, and participants’ experiences may be influenced by unique regulations and service delivery systems in that state. Second, although efforts were made to include participants representing diverse agencies (e.g., rural and urban, standalone, and integrated agencies), other forms of geographic and demographic diversity were more limited. Further, as a focus group study, our research was subject to the standard limitations of this methodology. These included the interdependence of participants’ responses, facilitator bias through in-session behaviors and communication strategies, and the limited ability to examine the validity of the study (Piercy and Nickerson 1996).

In an attempt to address the third limitation, we included a within-study validity check by having three coders per transcript, in part to address these issues. The interrater agreement could be considered to be somewhat low, in that the interpretation of the Fleiss’ kappa coefficient indicates slight agreement (Landis and Koch 1977). In part, this could have been a result of our decision to define the coding unit as a complete statement made by participants, especially because some statements contained more than one distinct idea, and the coders had to identify the single code that they felt best reflected those statements. As such, this introduced variability into the coding process, and different coders viewing the statement could have perceived a different main idea. Although this limitation could have been minimized by using smaller coding units (e.g., sentence-by-sentence coding), because of the bulk of data collected through the nine focus groups, we chose to use the full statements as the coding units in order to keep the coding process manageable. In addition, by using the process of having three coders per each statement, we built in a system for increasing the validity of the coding system. Finally, the exclusion of the statements that were not coded into the main identified categories could mean that some important aspects of safety planning were not addressed in the final interpretation of the results, especially issues that are not commonly encountered, but still may be important to understand.

Summary of Major Findings

This study adds further credence to the notion that domestic violence victims face a plethora of potential safety risks. Professionals who conduct safety planning with domestic violence victims must strive to obtain a comprehensive understanding of the unique risks each client faces in order to inform the safety planning process. The participants offered several insights regarding the safety risks that victims may face. These included (a) lacking personal resources (e.g., finances, job skills, and transportation); (b) perpetrators using technology to exacerbate their abuse and tracking of their partners; (c) an increased risk that may occur when the victim leaves the abusive relationship; (d) victims’ vulnerabilities due to mental health and substance abuse issues; (e) stalking; (f.) unique cultural and contextual risks for clients from marginalized populations; (g) perpetrator characteristics, such as their own mental health and substance abuse issues; (h) perceptions victims may hold that might make it challenging for them to accurately assess the level of risk they face; and (i) the power and control dynamics associated with abuse. Each of these areas is important to address in the safety planning process.

We intentionally recruited participants from different types of communities because we were interested in the impact of the community context of the safety planning process. As this was a qualitative study, we cannot draw strong conclusions about the types of communities (e.g., rural vs. urban, communities with varying degrees of cultural diversity, and communities with many and few community-based resources) that are more or less likely to enhance the safety planning process. However, the findings offer insight into community characteristics that might contribute to more effective safety planning. First, because a lack of community resources was noted as a barrier to effective safety planning, we can infer that more effective safety planning is facilitated by communities with a greater availability of supportive resources. The types of resources that are important go beyond domestic violence-specific resources and include supports for legal assistance, shelter, and transportation. Second, because community collaborations were emphasized by participants, the availability of resources should be combined with effective coordination of these services within the community. Third, these findings suggest that safety planning is enhanced when community members take the issue of domestic violence seriously and have adequate knowledge of what domestic violence is and the resources that are needed to address it. Finally, to the extent possible, the removal of barriers in the community is likely to increase the effectiveness of safety planning in that it will allow for more effectively linking clients with needed resources and more responsive community supports for victims.

The procedures used in safety planning can vary widely across different organizations (Waugh and Bonner 2002). This was evident in our study, as the participating agencies used a broad range of approaches, forms, and procedures when conducting safety planning with clients. Each agency faced unique requirements to their safety planning processes (e.g., requirements from funding agencies), and they also used their own in-house procedures for training staff and ensuring that staff were adhering to relevant policies. However, the findings of this study suggest that, despite some variations across agencies, several common themes can be found in the approaches that service providers use when safety planning with clients.

First, agencies typically view safety to be one of, if not the, major priorities when working with domestic violence victims. In fact, a general impression among the research team was that it was initially challenging for participants to talk about safety planning because it was such a natural, inherent part of their work, which made it difficult to articulate. Second, recognizing the complexity of safety in relation to the dynamics of domestic violence, safety planning is usually done through an individualized conversation that aims to empower victims to make decisions for themselves that are in their best interest. Third, as part of the safety planning process, professionals traditionally share very concrete, specific, and behavioral strategies with their clients, along with detailed information about how to access community resources. Thus, safety planning is not an abstract conversation, but rather it should result in tangible strategies that victims can use to address their unique safety risks.

Fourth, safety planning is not a one-time event. Rather, it is an ongoing process that must address emerging risks as they arise. Likewise, safety planning is not relegated to one certain category of professional helpers, but it is a skill that crosses professional roles. Fifth, professionals must work together, within and across different agencies, to ensure that safety planning is coordinated. Sixth, when safety planning with clients, professionals typically address unique client characteristics that may give rise to specific safety risks, including whether the client intends to leave the abusive relationship, whether the client is currently living in a domestic violence shelter, unique cultural background characteristics, and the extent to which the client faces technology-related risks. Finally, service providers typically view their safety planning procedures as helpful to clients, although the effectiveness is difficult to determine because of the long-term needs of victims and variations in how safety planning may occur from one client to another.

Implications for Future Research and Practice

The most significant research question related to safety planning that remains unanswered is whether and how safety planning actually increases the safety of domestic violence victims. Although a small but growing body of literature suggests that safety planning promotes victim safety (e.g., McFarlane et al. 1998, 2004), a need remains for additional information on the specific strategies that are most likely to keep domestic violence victims safe in the face of the diverse safety risks they encounter. Given the diversity of safety planning practices across agencies (Waugh and Bonner 2002), evaluating practices across sites would be very complex; however, more naturalistic studies of current practices are valuable. Additional relevant research questions are as follows: (a) Does the intervention site (e.g., police department, court, domestic violence agency) impact the effectiveness of safety planning? (b) What are characteristics of effective professionals who are most successful at safety planning? (c) What safety planning strategies are most effective during different stages of clients’ decisions to remain in or leave abusive relationships? and (d) How do clients define success regarding the outcomes of domestic violence safety planning?

As a qualitative study, this study offered some in-depth insights into service providers’ perspectives regarding safety planning, and these insights can inform future qualitative and quantitative research. One area for future research is the impact of technology on safety risks and safety planning, and the participants in this study discussed the relatively newer influence of technology on abuse dynamics. As so many technological advances have occurred in recent years, researchers should strive to develop a greater understanding of the specific technologies that pose threats and effective strategies for managing these threats.

Another important area for future research is to examine the unique cultural and other contextual variables that impact the safety planning process. This study did not delve deeply into the safety risks and associated safety planning strategies of marginalized populations. However, as some safety planning practices are used widely in the field, it is important to study the applicability of these approaches with groups who have unique needs and vulnerabilities. This is especially relevant when traditional resources may be less readily accessible to members of the population (e.g., safety planning documents that are not available in a client’s primary language).

This study also has several implications for practice. One practice-related challenge is how to consider increasing the standardization of safety planning in the field while also maintaining the ability for professionals to attend to each client’s unique needs in the safety planning process. For safety planning to become a much more standardized practice, agencies would lose some flexibility in their ability to use approaches that reflect their communities, agency cultures, and client population needs. However, because practices can vary so widely across organizational settings, it is likely that the services that clients receive also vary widely in their effectiveness. Therefore, we recommend that researchers and practitioners work to ensure that more effective practice strategies are available to clients in all intervention settings.

Finally, just as safety planning with individual clients was viewed by this study’s participants as an ongoing process that must address new and emerging safety risks, we urge professionals from all disciplines who work with clients impacted by domestic violence to view safety planning as a practice that requires ongoing modifications and advancements. Focus group participants identified several newer safety risks that simply were not present in years past, such as perpetrators tracking victims using GPS devices and through social networking websites. Similarly, several participants noted how the current economic decline in the U.S. has impacted the availability of resources for victims in the community. These technology- and economy-related issues reflect a broader need for safety planning practices to continue to advance as societal shifts occur that impact the safety of domestic violence victims. Promoting the safety of domestic violence victims and their children must remain one of the highest priorities among professionals who work with these populations, and developing and advancing more effective safety planning strategies will likewise remain an important step toward achieving that goal.