Developing the Emergency Department Elder Mistreatment Assessment Tool for Social Workers Using a Modified Delphi Technique

Abstract

Elder mistreatment is common and has serious consequences. The emergency department (ED) may provide a unique opportunity to detect this mistreatment, with social workers often asked to take the lead in assessment and intervention. Despite this, social workers may feel ill-equipped to conduct assessments for potential mistreatment, due in part to a lack of education and training. As a result, the authors created the Emergency Department Elder Mistreatment Assessment Tool for Social Workers (ED-EMATS) using a multiphase, modified Delphi technique with a national group of experts. This tool consists of both an initial and comprehensive component, with 11 and 17 items, respectively. To our knowledge, this represents the first elder abuse assessment tool for social workers designed specifically for use in the ED. The hope is that the ED-EMATS will increase the confidence of ED social workers in assessing for elder mistreatment and help ensure standardization between professionals.

assessment, elder abuse, emergency department, hospitals

Elder mistreatment is an important but underrecognized social issue. Elder mistreatment is often referred to as elder abuse, which is defined as “a single or repeated act, or lack of appropriate actions, which causes harm, risk of harm, or distress to [an older adult]” (New York City Elder Abuse Center, n.d., para. 1). Elder mistreatment is perpetrated by a person in a position of trust or by someone who targets an older adult based on age or disability (National Research Council [NRC], 2003). It includes physical abuse, sexual abuse, psychological abuse, financial exploitation, and neglect, with many victims experiencing multiple types of mistreatment concurrently (NRC, 2003; U.S. Department of Justice and Department of Health and Human Services, 2014).

Elder mistreatment is common, affecting approximately 10 percent of community-dwelling older adults each year (Acierno et al., 2010; Lachs & Pillemer, 2004; NRC, 2003; Pillemer, Burnes, Riffin, & Lachs, 2016). This mistreatment often has serious consequences, including increased mortality (Lachs & Pillemer, 2015; Lachs, Williams, O'Brien, Pillemer, & Charlson, 1998), hospitalization (Dong & Simon, 2013b), and nursing home placement (Dong & Simon, 2013a; Lachs, Williams, O'Brien, & Pillemer, 2002). Elder mistreatment is also associated with increased risk of depression and exacerbation of chronic illness (Dyer, Pavlik, Murphy, & Hyman, 2000). Despite the high prevalence and severity, only one in 24 cases are recognized and reported to authorities (Lifespan of Greater Rochester, Weill Cornell Medical Center of Cornell University, and New York City Department for the Aging, 2012). The vast majority of these victims remain unidentified (Burnes, Acierno, & Hernandez-Tejada, 2018).

A hospital emergency department (ED) visit provides an important opportunity to improve elder mistreatment detection (Fulmer, Paveza, Abraham, & Fairchild, 2000) and to initiate intervention. An ED presentation for acute injury or illness may represent the only time a mistreated older adult ever leaves the home (Bond & Butler, 2013). Although elder mistreatment identification in the ED is currently infrequent (Rosen, Hargarten, Flomenbaum, & Platts-Mills, 2016), the ED already plays an essential role in identifying and intervening in cases of child abuse (Hochstadt & Harwicke, 1985; Kistin, Tien, Bauchner, Parker, & Leventhal, 2010) and intimate partner violence (IPV) (Choo et al., 2015; Rhodes et al., 2015).

In EDs that have a social worker as part of the clinical team, this social worker may be expected to take the lead in assessing for potential elder mistreatment and initiating interventions for victims. The value of social workers as part of the ED multidisciplinary care team, particularly for older adult patients, is increasingly recognized in this setting (Auerbach & Mason, 2010; Hamilton et al., 2015). Despite this, many ED and hospital-based social workers lack training for or experience with comprehensive assessment for elder mistreatment and are unclear about their responsibilities (Felton & Polowy, 2015). In contrast to child abuse and IPV, hospitals do not routinely offer elder mistreatment training to social workers. Although the Council on Social Work Education (CSWE) has developed elder justice curriculum modules for social work schools, preprofessional social workers have little knowledge of elder abuse (CSWE, 2015; Policastro & Payne, 2014). Information about state-mandated training for social workers is not readily available, but anecdotal evidence suggests that this training is uncommon, even while social workers are mandated reporters in most states. Therefore, many ED social workers are unsure how to approach an assessment for elder mistreatment and have little guidance to assist them.

An understanding of the need to increase identification of elder mistreatment in the ED has led to a new focus on improving recognition by medical professionals in this setting. This includes additional training on both warning signs and risk factors, and consideration of implementation of a brief universal or targeted ED screening (Rosen, Stern, Elman, & Mulcare, 2018). As a result, it is likely that many more cases of suspected elder mistreatment will be referred to ED social workers for assessment.

To our knowledge, no tools to assist ED-based social workers with elder mistreatment assessment have been developed. Therefore our goal was to create an elder mistreatment assessment tool for social workers in the ED.

METHOD

To design our ED social work assessment for elder abuse, we used a modified Delphi technique (Custer, Scarcella, & Stewart, 1999). This multistep process included an examination of existing assessment and screening tools, design of a preliminary version of the assessment tool, initial review by an expert panel with individual feedback via survey, expert panel focus group discussion, integration of panel recommendations, and final review of the revised version by the expert panel. The steps of this process are shown in Figure 1.

Figure 1

Summary of Multistep Process Using Modified Delphi Technique to Develop Emergency Department Elder Mistreatment Assessment Tool for Social Workers

Examination of Existing Assessment and Screening Tools

We conducted a comprehensive examination of existing elder abuse assessment and screening tools for professionals in different disciplines. We identified existing tools using several recent systematic and comprehensive reviews (Beach, Carpenter, Rosen, Sharps, & Gelles, 2016; Fulmer, Guadagno, Bitondo Dyer, & Connolly, 2004; Gallione et al., 2017; National Center on Elder Abuse [NCEA], 2017; Phelan, 2012; Worrilow & Barraco, 2015). The reviews are shown in more detail in Table 1. We also collaborated with NCEA (2017), who had prepared a Research to Practice report on the topic. We evaluated four tools for elder abuse assessment and 14 published tools for elder abuse screening. For tools that were not readily available, we contacted the authors to receive a copy. In addition, with the permission of the creators, we examined a screening tool designed for ED use that had not yet been published but has been subsequently (Platts-Mills et al., 2018). Table 2 describes the assessments and tools we examined.

Table 1

Published Reviews of Elder Abuse Screening Tools and Assessments

	Systematic	Databases Used	Number of	Time Period
Authors, Year	Review	in Review	Results	Covered	Summary of Results
Fulmer, Guadagno, Bitondo Dyer, & Connolly, 2004	No	Not disclosed	11	1981–2000	• Consensus must be achieved to determine what constitutes an appropriate/acceptable screen/assessment• Focus should be on developing instruments for rapid assessment and longer diagnostic assessments• Much need for additional research
Gallione et al., 2017	Yes	MEDLINE, Cochrane, EMBASE, Scopus	11	1980–2015	• No tool identified as a gold standard• Many tools require validity and reliability testing
National Center on Elder Abuse, 2017	No	Not disclosed	17	1980–2016	• There is no gold standard for elder abuse screening• Many screening tools exist, designed primarily for health care providers• Not all tools have psychometrics
Worrilow & Barraco, 2015	No	MEDLINE	9	1991–2015	• Most tools were not designed for the emergency department• More research should be done to standardize criteria for evaluation and diagnosis of elder abuse

Table 2

Published Elder Abuse Screening Tools and Assessments

Screening Tools and			Domains	Social Worker
Comprehensive Assessments	Citation	Setting	Addressed	as Authora
Comprehensive Assessments
Case Detection of Abused Elderly Patients	(Rathbone-McCuan & Voyles, 1982)	Clinical setting	All	Yes
Assessment of Elder Mistreatment: Issues and Considerations. Elder Mistreatment Guidelines for Health Care Profession: Detection, Assessment and Intervention	(Ansell & Breckman, 1988)	Clinical setting	All	Yes
Written protocol for identification and assessment in Strategies for Helping Victims of Elder Mistreatment	(Breckman & Adelman, 1988)	Any, but has medical component for clinical setting	All	Yes
Harborview Medical Center elder abuse diagnostic and intervention protocol in Elder Abuse and Neglect: Causes, Diagnosis, and Intervention Strategies	(Quinn & Tomita, 1997)	Any, but has medical component for clinical setting	All	Yes
Screening Tools
Screening Protocol for Identification of Abuse and Neglect of the Elderly	(Johnson, 1981)	Not specified	Physical, verbal/psychological, financial, neglect	No
Health, Attitudes Toward Aging, Living Arrangements, and Finances (HALF) Assessment	(Ferguson & Beck, 1983)	Health services setting	All, but is observational	No
Hwalek-Sengstock Elder Abuse Screening Test (H-S/EAST)	(Neale, Hwalek, Scott, Sengstock, & Stahl, 1991)	Health and social services agencies	All	Yes
Brief Abuse Screen for the Elderly (BASE)	(Reis, Nahmiash, & Shrier, 1993)	Clinical setting	All, but is observational	Yes
Caregiver Abuse Screen (CASE)	(Reis & Nahmiash, 1995)	Community-based health and social services agencies	Physical, verbal/psychological, neglect	Yes
Indicators of Abuse (IOA)	(Cohen, Halevi-Levin, Gagin, & Friedman, 2006; Reis & Nahmiash, 1998)	By a professional, following a home visit	All, but is observational	Yes
Screening Tools and Referral Protocol for Stopping Abuse Against Older Ohioans	(Bass, Anetzberger, Ejaz, & Nagpaul, 2001)	Not specified	Physical, verbal/psychological, financial, neglect	Yes
Vulnerability to Abuse Screening Scale (VASS)	(Schofield, Reynolds, Mishra, Powers, & Dobson, 2002)	Mailed to clients	All	No
Elder Assessment Instrument (EAI)	(Fulmer, 2003)	Clinical setting	Physical, financial, neglect	No
Elder Abuse Suspicion Index (EASI)	(Yaffe, Wolfson, Lithwick, & Weiss, 2008)	Family practice ambulatory care	All	Yes
Elder Abuse and Neglect Risk Assessment Tool	(Cohen, Halevi-Levin, Perlozki, Gagaeien, & Freedman, 2010)	Not specified	Physical, financial, neglect	Yes
Older Adult Financial Exploitation Measure (OAFEM)	(Conrad, Iris, Ridings, Langley, & Wilber, 2010)	Social services agencies	Financial	Yes
Geriatrics Mistreatment Scale (GMS)	(Giraldo-Rodriguez & Rosas-Carrasco, 2013)	Clinical setting	All	No
ED Senior AID (Emergency Department Senior Abuse Identification) Tool	(Platts-Mills et al., 2018)	Emergency setting	All	No

a

Design team included author with social work background.

Design of Preliminary Version

Two authors (Alyssa Elman [AE] and Tony Rosen [TR]) initially comprehensively examined each of the existing published tools. We assessed the level of evidence supporting each tool, including whether it had been validated, and, if so, how applicable the setting in which the validation occurred was to ED social work. We used in-person meetings to qualitatively analyze and categorize the individual items from all tools to identify relevant domains for inclusion. We then built a preliminary version of our ED social work assessment tool based on the themes that emerged. We preferentially included items from validated scales when possible while attempting to be mindful of the unique challenges associated with conducting a social work assessment in an ED setting. This was developed into a preliminary version of the tool through several team meetings, where it was revised iteratively through a consensus process. Team members included three social workers with expertise in acute care, geriatrics, family violence, and elder abuse; three emergency physicians with additional training in geriatrics and a focus in elder mistreatment; a geriatrician with extensive elder mistreatment research and leadership experience; and an epidemiologist focusing on acute care and public health and sociomedical issues in older adults. Attempting to balance the need to ensure assessment for multiple types of mistreatment with a sensitivity to the omnipresent time pressure in an ED, we divided the preliminary version of the tool into a brief initial assessment and then a comprehensive assessment to be performed if concern existed after completing the former. In the preliminary version, the initial part of the tool had 14 items and the comprehensive had 24 items.

Modified Delphi: Overview

We used a modified Delphi approach to incorporate the input of a panel of experts to refine the tool. The Delphi method is well established and has been used extensively in health care and other sectors to reliably develop consensus. This method, initially described by the RAND Corporation (Dalkey & Helmer, 1963), is an iterative process that uses a systematic progression of voting by an expert panel to build consensus around a research question about which little is known (Eubank et al., 2016; Hsu & Sandford, 2007). The original Delphi method developed at RAND did not include a face-to-face meeting of experts; however, the Delphi method has been “modified” to include the opportunity for interaction between experts (Gustafson, Shukla, Delbecq, & Walster, 1973). Previous studies have shown that this modified Delphi approach can be better than the traditional Delphi method (Eubank et al., 2016; Graefe & Armstrong, 2011; Gustafson et al., 1973). In our study, we used this modified Delphi approach, which has also been successfully used for elder abuse risk indicators and screening questions (Erlingsson, Carlson, & Saveman, 2003), roles and responsibility in a multidisciplinary elder abuse intervention (Du Mont, Kosa, Macdonald, Elliot, & Yaffe, 2015), skills-based competencies for forensic nurse examiners providing elder abuse care (Du Mont, Kosa, Macdonald, Elliot, & Yaffe, 2016), and ED palliative care screening (George et al., 2015).

Modified Delphi: Expert Panel Recruitment

We identified experts to participate in this panel based on our experience in the field and on recommendations from colleagues. Specifically, we described our project to the National Collaboratory to Address Elder Mistreatment and asked fellow members to suggest social work experts. The panelists included professional social workers with experience in elder abuse and sociomedical issues among older adults. Participants had expertise in family violence, ED social work, outpatient geriatrics, and palliative care. Several participants have worked closely with law enforcement and criminal justice, and many have participated in multidisciplinary community-based teams that address challenging elder abuse cases. Details about the credentials and relevant experience of the expert panel can be found in Table 3.

Table 3

Expert Panel Characteristics

		Experience
							Years
			Total Years	Years of Elder	Years of	Years of Domestic	Emergency
Participant	Reason for	Credentials/	of Clinical	Abuse–Focused	Geriatric	Violence Social	Department
#	Selection	Licensure	Social Work	Work	Social Work	Work	Social Work
1	Elder abuse expertise	LCSW, MSW	38	20
2	Elder abuse expertise	LCSW, MSW, PhD	36	24	36	12	2.5
3	Elder abuse expertise	LCSW, MSW	12				6
4	Geriatric expertise	LCSW, MSW	25		23
5	Geriatric expertise	LCSW, MSW	16		16
6	Geriatric expertise	LCSW, MSW	10	10	10
7	Geriatric expertise	LMSW, MSW	4		2
8	Domestic violence expertise	LCSW, MSW	19		3	5
9	Domestic violence expertise	LCSW, MSW	15			7
10	Emergency department expertise	LCSW, MSW	26				20

Modified Delphi: Initial Independent Review and Survey

Panel members were sent the preliminary version of the assessment tool and the 18 existing tools on which it was based. In this first phase, each member of the panel reviewed the preliminary tool individually. They were asked to make recommendations for modifications using an online survey.

The survey asked participants to indicate how critical each question was (on a scale of 0 to 8) and whether each question belonged in the initial or comprehensive part of the assessment or should not be included at all. We also asked if any items should be reworded for clarity. Panelists were sent a link to the survey Web page within the e-mail and used an online interface to enter their answers. The survey was designed and the results stored using Typeform (Barcelona, Spain). The entire survey is available on request from the first author.

Results were analyzed to identify the level of consensus for each item and to determine which items needed additional discussion. Based on survey results, consensus did not exist among experts for 13 items from the preliminary version of the tool. Panelists did not agree about whether six of the 13 items should be included in the tool and whether an additional seven belonged in the brief initial section or the comprehensive section.

Modified Delphi: Focus Group with Expert Panel

A focus group was conducted in April 2018 with all expert panelists participating in person or via telephone to discuss the tool and build consensus. Qualitative approaches are often helpful to explore topics that lack significant previous research and may not lend themselves to quantitative analysis (Green & Britten, 1998; Krumholz, Bradley, & Curry, 2013; Pope & Mays, 1995). Author TR has formal training in qualitative research and has conducted successful studies with such approaches in the past. This focus group was moderated by several of the authors (AE, Sarah Rosselli [SR], Sunday Clark, Risa Breckman, and TR). A summary of the survey results was sent to all panelists in advance of the focus group. The focus group guide incorporated these results and included questions such as “Should any of the questions be removed entirely?” and “Would you reword any of the questions for clarity, completeness, or acceptability?” (The entire focus group guide is available on request from the first author.) We used the focus group to allow the expert panel to discuss, in detail, all questions for which initial written responses differed substantially. Through this focus group, consensus was arrived at for all questions. The focus group was audio-recorded and professionally transcribed. Transcripts were then reviewed in detail and discussed in depth by authors AE, SR, and TR to identify themes and review items for consensus.

Figure 2

Emergency Department Elder Mistreatment Assessment Tool for Social Workers (ED-EMATS)—Initial Assessment

Figure 3

Emergency Department Elder Mistreatment Assessment Tool for Social Workers (ED-EMATS)—Comprehensive Evaluation

Modified Delphi: Final Revisions and Panel Rereview

Based on feedback from the focus group, we made revisions to the tool. Per suggestions from the expert panel, two toolboxes were incorporated into the assessment: (1) a rapport-building toolbox for the initial assessment and (2) a sexual assault toolbox for the comprehensive section. Items for the toolboxes were developed during the focus group.

Seven items were removed entirely, primarily because they would likely be asked in a standard social work assessment. One question was added to the initial tool and one question was moved from the initial section to the comprehensive section. Four questions were moved from the comprehensive section to the rapport-building toolbox. Many questions were rephrased for clarity and the order of questions was revised.

Before finalization, the revised version was sent to all expert panelists via e-mail to ensure that the changes reflected what had been discussed and to solicit any final comments.

NAME OF TOOL

The final versions of the tool, which we have named Emergency Department Elder Mistreatment Assessment Tool for Social Workers (ED-EMATS), are shown in Figure 2 and Figure 3. The initial screening section (Figure 2) has 11 items and the comprehensive assessment section (Figure 3) has 17 items. We included rapport-building and sexual assault toolboxes. In our experience using the tool, the initial screening section takes five to 15 minutes to complete and the comprehensive follow-up takes 15 to 25 additional minutes to complete. Patient interactions vary in length based on their responses and the need to ask follow-up questions.

DISCUSSION

We have attempted to design an assessment tool specifically for use by social workers in the ED setting. Although brief screening instruments exist for varied disciplines, including one currently intended for use by ED medical providers and nurses (Platts-Mills et al., 2018) and one for health care professionals in the community (Bomba, 2006), which social workers may use, previously published tools we reviewed were not developed expressly for social workers and case managers. ED social workers have previously indicated that they believe many cases of elder abuse are likely missed, in part because of a lack assessment tools that are available for use in the ED (Rosen, Stern, Mulcare, et al., 2018). Our goal is to provide an additional framework that may be of assistance to those already seasoned in practice and likewise develop a standardized approach for inexperienced social workers to feel more confident in conducting this assessment, especially in the absence of formal training. We believe that the ED-EMATS will increase the confidence of ED social workers in assessing for elder mistreatment and help ensure standardization between professionals.

The ED-EMATS may be added to electronic medical record to further assist ED social workers in appropriately assessing these vulnerable patients and accurately documenting the encounter. It may also be incorporated into regular ED and hospital social work and case manager trainings. In addition, it may serve as an element of an elder mistreatment curriculum that should be implemented in professional social work schools. We plan to test the efficacy of this tool qualitatively and quantitatively, including examining validity, reliability, and ease of use in a clinical setting. Doing so may allow us to compare it to existing tools. Our goal with the ED-EMATS tool, though, is not only to assist ED social workers in accurately identifying elder abuse, but also to give a structure for a comprehensive assessment to help determine appropriate next steps. We hope to test its ability to effectively fulfill this function as well.

We recognize that responses to many items in this tool may require additional probing. We have not provided guidance or advice for this, as we believe that this additional follow-up exploration is a core part of social work training. Also, we have not provided suggestions for interventions for patients identified as elder mistreatment victims or at risk. Available resources and appropriate next steps may differ between communities and based on the patients’ circumstances.

Limitations

This study has several limitations. We used a comprehensive qualitative review of existing tools to categorize items and identify domains to develop a preliminary version of our ED social work assessment tool. Although this is an established research technique, domains and their categorization may have been affected by the subjective way in which the tools were assessed by the research team. In addition, although we attempted to include items from validated tools when possible, given that most tools had not been previously validated, many items in the tool we developed had not been formally validated. Another limitation is that we chose expert panelists based on our experience in the field and on recommendations from colleagues rather than using objective criteria. We attempted to mitigate this by inviting social workers with various experiences and from different geographical regions, nearly all of whom agreed to participate. Also, our expert panel was relatively small. The small size of the panel is partially because limited expertise about elder mistreatment currently exists among social work professionals. Notably, it is unclear how many experts are needed for the Delphi process (Smith, Clark, Kapoor, & Handler, 2012), but it is possible that our panel does not represent broad enough perspectives on the topic (Lerner et al., 2015). The goal of consensus may also have weakened the final tool, as panelists may have conformed to the majority opinion rather than disagreeing (Mukherjee et al., 2015). While moderating the focus group, we attempted to minimize this limitation by providing opportunities for everyone to share their perspective and explore concerns or differences of opinion in detail.

Conclusion

Through a rigorous modified Delphi process, we have developed the ED-EMATS, an ED social work assessment tool. We anticipate that it will standardize and improve the care social workers provide older adults in the ED. We believe it will become a useful part of an increased professional focus on assessing and initiating intervention for vulnerable victims of elder mistreatment in the ED and other settings.

Alyssa Elman, LMSW, is supervising social worker of the Vulnerable Elder Protection Team (VEPT), Department of Emergency Medicine, New York–Presbyterian/Weill Cornell Medical Center. Sarah Rosselli, LCSW, is Victim Intervention Program social worker, Department of Emergency Medicine, New York–Presbyterian/Weill Cornell Medical Center. David Burnes, PhD, is associate professor and associate dean, University of Toronto, Factor-Inwentash Faculty of Social Work. Sunday Clark, ScD, is associate professor of epidemiology research in emergency medicine and director of research, Department of Emergency Medicine, New York–Presbyterian/Weill Cornell Medical Center; Michael E. Stern, MD, is associate professor of clinical emergency medicine and chief, geriatric emergency medicine; Veronica M. LoFaso, MD, is associate professor of medicine; and Mary R. Mulcare, MD, is assistant professor of clinical emergency medicine and director, undergraduate medical education for the Department of Emergency Medicine, New York–Presbyterian Hospital/Weill Cornell Medical Center. Risa Breckman, LCSW, is director, New York City Elder Abuse Center, Weill Cornell Medical College. Tony Rosen, MD, is assistant professor of emergency medicine and program director of VEPT, Department of Emergency Medicine, New York–Presbyterian/Weill Cornell Medical Center. Address correspondence to Alyssa Elman, Department of Emergency Medicine, New York–Presbyterian Hospital/Weill Cornell Medical Center, 525 E. 68th Street, New York, NY 10065; e-mail: [email protected].