An Evaluation of AoA's Program to Prevent Elder Abuse: Final Report. V. FINDINGS

08/01/2016

In this section, we present key findings related to three domains examined by the evaluation. We begin by presenting the characteristics of elders served by the prevention interventions, including: demographic information and household composition (where known); the type of abuse experienced; psychological and physical health indicators; and social well-being. Next, we present findings related to service referrals to address elders' identified needs and service utilization. Finally, as applicable, we address common and intervention-specific outcomes.

A. Risk Factor Profiles

1. What are the Demographic and Household Characteristics of Victims, At-risk Elders and Care Recipients?

The demographic and household characteristics of the elders served by the five prevention interventions are presented in Table 3. They are presented according to their status as substantiated victims (UTHSC), elders at-risk or known victims (AK DSDS, NYSOFA, and TX/WellMed), and care recipients (USC).

TABLE 3. Demographic and Household Characteristics of Participants
  Victims At-risk Elders/Victims Care Recipients
UTHSC
(N=34)
AK DSDS
(N=87)
NYSOFA
(N=221)
TX/WellMed
(N=310)
USC
(N=76)
N Mean N Mean N Mean N Mean N Mean
Age 34 74.4 55 70.3 151 80.9 310 76.3 76 80.8
  N % N % N % N % N %
Gender 34   87   206   310   76  
   Male   38.2   39.1   33.9   36.1   55.3
   Female   61.8   60.9   66.0   63.9   44.7
Race 34   79   157   294   67  
   African American/Black   61.8   13.9   15.9   7.8   1.5
   American Indian/Native Alaskan   0   24.1   0   0.3   0
   Asian/Pacific Islander   0   2.5   0   1.0   7.5
   Caucasian/White   20.6   55.7   84.1   90.1   91.0
   Other*   17.6   3.8   0   0.7   0
Ethnicity 34   84   131   301   76  
   Hispanic   17.6   3.6   5.3   53.5   13.2
   Not Hispanic   82.3   96.4   94.7   46.5   86.8
Education 33   75   29   ---   67  
   Less than high school   42.4   14.7   3.4   ---   7.5
   High school   36.4   41.3   34.5   ---   15.0
   Some college   21.2   25.3   6.9   ---   9.0
   College graduate   0   14.7   27.6   ---   34.3
   Some graduate work   0   4.0   27.6   ---   34.3
Income 24   36   122   ---   29  
   Less than $15,000   70.8   47.2   22.1   ---   10.3
   $15,000-$25,000   20.8   41.6   31.1   ---   17.2
   $25,001-$35,000   4.2   5.6   23.0   ---   3.4
   $35,001-$50,000   4.2   5.6   10.7   ---   27.6
   $50,001-$75,000   0   0   9.8   ---   20.7
   $75,001-$100,000   0   0   0.8   ---   13.8
   Greater than $100,000   0   0   2.5   ---   6.9
Marital Status 34   82   166   274   73  
   Married   11.8   14.6   15.1   25.5   71.2
   Single/Never married   2.9   20.7   14.5   14.2   4.1
   Divorced/Separated   32.4   41.4   10.8   6.9   1.4
   Widowed   52.9   23.2   54.2   21.2   23.3
   Other/Unknown   0   0   5.4   32.2   0
Primary Language 34   72   169   309   76  
   English   82.4   97.2   95.3   86.1   88.2
   Spanish   14.7   1.4   2.4   13.3   1.3
   Other   2.9   1.4   2.4   0.6   10.5
Place of Residence 34   76   170   308   ---  
   Private home**   91.2   88.1   72.4   77.9   ---
   Multi-family home   ---   0   7.1   0   ---
   Assisted living/Nursing home   ---   6.6   14.7   5.8   ---
   Group home/Unlicensed personal care homes   ---   1.3   2.4   1.3   ---
   Other   8.8   3.9   2.9   14.9   ---
   Unclear   ---   0   0.6   ---   ---
Living Situation 34   64   170   ---   ---  
   Alone   52.9   78.1   72.9   ---   ---
   With spouse   2.9   21.9   7.1   ---   ---
   With family   35.3   0   14.7   ---   ---
   With provider   0   0   2.4   ---   ---
   Other   8.8   0   2.9   ---   ---
* For UTHSC, "Other" refers to Hispanic/Latino.
** This includes apartments and rented rooms.

On average, elders served by the prevention interventions ranged from 75 years to 81 years of age. For those served through UTHSC's medication adherence intervention, the mean age was 74.4. For the at-risk elders served by AK DSDS's CTI, the mean age was 70.3. The mean age for the elders at-risk of financial exploitation served by NYSOFA was 80.9, and those served by TX/WellMed was 76.3 years. For the care recipients participating in USC's intervention, the mean age was 80.8.

The majority of victims participating in the interventions were female. For four of the five prevention intervention, about two-thirds of the elders were females and one-third were male (UTHSC, AK DSDS, NYSOFA, and TX/WellMed). By contrast, USC's prevention intervention did not focus on victims but on care recipients. More than one-half of USC's elders were males.

The race and ethnicity of the elders served by the prevention interventions varied by locale. Nearly 62 percent of the elders taking part in UTHSC medication adherence intervention were African American, 21 percent Caucasian, and 18 percent identified as Hispanic/Latino. AK DSDS served a diverse population: nearly 56 percent were Caucasian, 24 percent were American Indian or Alaska Native, and 14 percent were African American. Four-fifths of the elders served by NYSOFA's E-MDTs were Caucasian (84 percent) and less than one-fifth were African American (15.9 percent). This was similar to the racial and ethnic background of the elders served by TX/WellMed, as 90 percent of elders were Caucasian and only 8 percent were African American. More than one-half (53.5 percent) of the elders served by TX/WellMed identified as Hispanic/Latino. Of the care recipients served by USC, 91 percent were Caucasian, 7.5 percent were Asian/Pacific Islander, and 1.5 percent were African American.

Across the five prevention interventions, more than 80 percent of elders spoke English as their primary language. Spanish was the primary language for more than 10 percent of the elders served by UTHSC and TX/WellMed (14.7 percent and 13.3 percent, respectively).

The education levels of the elders served varied considerably across the prevention interventions.8 Most elders served by UTHSC had less than a high school education (42.4 percent) or were high school graduates (36.4 percent); about 21 percent had some college education but were not college graduates. The diversity of the AK DSDS population was further demonstrated by their education levels. About 15 percent had less than a high school education, 40 percent were high school graduates, 25 percent had some college education, and about 15 percent were college graduates, with 4 percent having some graduate school education. Elders served through NYSOFA tended to be high school (34.5 percent) and college graduates (27.6 percent) or completed graduate work (27.6 percent); only a very small percentage had less than a high school education (3.4 percent). More than one-third of the elders served by USC were college graduates and another one-third had graduate school education. About 15 percent were high school graduates and 7.5 percent had less than a high school education.

Income levels of elders across three of the prevention interventions tended to be low.9 Of all the data elements, information on elders' income levels was among the most challenging for grantees to obtain. Response rates for the four interventions that collected this information ranged from 38.2 percent to 70.6 percent. Of those participants who provided information on income, nearly 92 percent of the elders served by UTHSC had an annual income of less than $25,000, as did 89 percent of the elders served by AK DSDS. Similarly, about 53 percent of the elders served by NYSOFA had annual incomes less than $25,000. About 44 percent of the elders served by NYSOFA had incomes between $25,000 and $75,000, but very few had incomes exceeding $75,000 (3 percent). In contrast, some of the elders served by USC appeared to be affluent, with 14 percent having incomes greater than $75,000. Most elders (52 percent) had incomes between $35,000 and $75,000 yet about 28 percent had incomes less than $25,000.

Marital status also varied across the elders served by the prevention interventions. More than half of the elders served by UTHSC and NYSOFA were widowed (52.9 percent and 54.2 percent, respectively). About 40 percent of the elders served by AK DSDS were divorced or separated. Most elders served by USC were married (71 percent). The marital status of the elders served by TX/WellMed varied across all categories.

Most elders lived in a private home, which included apartments and rented rooms, although a few resided in assisted living or nursing facilities. The place of residence for elders was consistent across four of the prevention interventions.10 Most elders lived in a private home, which included apartments and rented rooms. About 15 percent of elders served by NYSOFA lived in assisted living or nursing home facilities; less than 2 percent lived in a group home or an unlicensed personal care homes.

Many elders lived alone. There was limited information available about the living situation of the elders. This information was available for three of the prevention interventions. Almost 80 percent of the elders served by AK DSDS lived alone, as did 73 percent of the elders served by NYSOFA, and 53 percent of the elders served by UTHSC. Twenty-one percent of the AK DSDS elders lived with a spouse. Fifteen percent of the NYSOFA elders and 35 percent of the UTHSC elders lived with family.

2. What are the Psychological, Physical Health and Social Conditions of Victims, At-risk Elders and Care Recipients?

The psychological, physical health and social conditions of victims, at-risk elders and care recipients were assessed using multiple measures. Together these measures present a sense of the vulnerability of elders served by the prevention interventions. Findings on physical health, psychological health, and social conditions are presented in Table 4 and discussed below.

TABLE 4. Physical Health, Psychological and Social Characteristics of Participants
  Victims At-risk Elders/Victims Care Recipients
UTHSC
(N=34)
AK DSDS
(N=87)
NYSOFA
(N=221)
TX/WellMed
(N=296)
USC
(N=76)
N Mean N Mean N Mean N Mean N Mean
Physical Function
ADL (0-35) --- --- 82 7.9 --- --- --- --- --- ---
ADL (0-6) --- --- --- --- --- --- --- --- 75 3.5
ADL (0-6) --- --- --- --- 91 1.0 --- --- --- ---
IADL (0-45) --- --- 84 16.6 --- --- --- --- --- ---
IADL (0-16) --- --- --- --- 106 2.8 --- --- --- ---
SF-36 (10-30) 30 17.1 --- --- --- --- --- --- --- ---
IADL (8-31)                 74  
   Independent (8-10)                   1.3
   Moderate (11-16)                   6.8
   Dependent (17-31)                   91.9
Mobility Impaired --- --- --- --- --- --- --- 24.0 --- ---
Physically Disabled --- --- --- --- --- --- --- 26.0 --- ---
Cognitive Function
Cognitively Impaired --- --- --- --- --- --- --- 9.3 --- ---
Dementia--Mini-Cog (0-5) --- --- --- --- --- --- --- --- 76 ---
   Positive screen (0-2) --- --- --- --- --- --- --- --- --- 85.1
   Negative screen (3-4) --- --- --- --- --- --- --- --- --- 14.9
MMSE (0-30) 31   ---   ---   ---   ---  
   Cognitively intact (25-30)   67.7   ---   ---   ---   ---
   Mild cognitive impairment (21-24)   19.3   ---   ---   ---   ---
   Moderate cognitive impairment (10-20)   12.9   ---   ---   ---   ---
   Severe cognitive impairment (<10)   0 --- --- --- --- --- --- --- ---
Depression
PHQ-9 (0-27) ---   51   ---   ---   67  
   No depression (0-4)   ---   58.8   ---   ---   78.9
   Minimal symptoms (5-9)   ---   17.7   ---   ---   14.7
   Minor symptoms (10-14)   ---   13.7   ---   ---   4.9
   Major depression, moderate (15-19)   ---   5.9   ---   ---   1.6
   Major depression, severe (>=20)   ---   3.9   ---   ---   0
GDS (0-15) 27 --- --- --- --- --- --- --- --- ---
   Absent (1-4)   74.1                
   Present (>4)   25.9                
Depression ---   ---   32   ---   ---  
   No   ---   ---   40.6   ---   ---
   Mild   ---   ---   25.0   ---   ---
   Moderate   ---   ---   21.9   ---   ---
   Severe   ---   ---   12.5   ---   ---
  N Mean N Mean N Mean N Mean N Mean
Anxiety
GAD-7 (0-21) --- --- 53 4.3 --- --- --- --- 63 1.5
  N % N % N % N % N %
Anxiety ---   ---   27   ---   ---  
   No anxiety   ---   ---   29.6   ---   ---
   Mild anxiety   ---   ---   25.9   ---   ---
   Moderate anxiety   ---   ---   37.0   ---   ---
   Severe anxiety   ---   ---   7.4   ---   ---
Stress
PSS (0-40)     72 18.8 --- --- --- --- --- ---
  N % N % N % N % N %
Stress ---   ---   ---   ---   ---  
   No stress   ---   ---   11.1   ---   ---
   Mild stress   ---   ---   18.5   ---   ---
   Moderate stress   ---   ---   51.8   ---   ---
   Severe stress   ---   ---   18.5   ---   ---
  N Mean N Mean N Mean N Mean N Mean
Social Support
LSNS-6 (0-30) --- --- --- --- --- --- --- --- 62 14.3
LSNS-R (0-60) --- --- 50 18.9 --- --- --- --- --- ---
DSSI (11-33) 31 24.3 --- --- --- --- --- --- --- ---
  N % N % N % N % N %
Social Support (Yes=1)
   Socially isolated --- --- --- --- 88 0.59 --- --- --- ---
   Leave the house --- --- --- --- 70 0.34 --- --- --- ---
   Friends or family --- --- --- --- 68 0.54 --- --- --- ---
   Emotional support --- --- --- --- 78 0.70 --- --- --- ---
   Perpetrator part of social support --- --- --- --- 94 0.80 --- --- --- ---

Physical Health

The physical functioning of elders served by four interventions was fairly low. For the elders served by UTHSC's medication adherence intervention, physical function was measured using the Short Form 36 (SF-36) Health Survey, ten item questions of which assess physical functioning. The mean score for elders served was 17.1, indicating a moderate level of physical functioning. AK DSDS, NYSOFA, and USC assessed physical functioning (i.e., the degree of an elder's dependency on others) by measuring participants' difficulty with performing ADLs and Instrumental Activities of Daily Living (IADLs). Using various ADL and IADL scales, elders served by AK DSDS's CTI were found to have low levels of difficulty (with a mean ADL score of 7.9 and a mean IADL score of 16.6) whereas elders served by NYSOFA's E-MDTs were assessed as having a high level of dependence (mean score was 1.0 on their ADL). On the IADL (0-8), the mean score for these elders was 2.8. For the care recipients assessed by USC, the mean score on the ADLs was 3.5 at baseline. On the IADL, the majority of care recipients were identified as dependent, with scores ranging 17-31. Among TX/WellMed participants, 24 percent were mobility impaired and 26 percent were physically disabled.

Cognitive Function

Levels of cognitive impairment varied. Self-neglecting elders were cognitively intact but most of the elder care recipients were cognitively impaired. UTHSC measured the cognitive function of self-neglecting elders using the Mini Mental State Exam (MMSE), which assesses orientation to time and place, tracking a sequence, attention in a mathematical context, short-term memory, several forms of language challenge, and ability to follow instructions involving visual-spatial manipulations. About 68 percent of the elders were cognitively intact (67.7 percent), with about 20 percent demonstrating mild cognitive impairment (19.3 percent), and about 13 percent with moderate cognitive impairment (12.9 percent). No elders demonstrated severe cognitive impairment. Nine percent of elders served by TX/WellMed were identified as cognitively impaired (9.3 percent). Using the Mini-Cog test to screen for cognitive impairment, USC found that 85 percent of the care recipients were cognitively impaired and 15 percent were not (85.1 percent and 14.9 percent, respectively).

Depression

There was limited evidence of depression among the elders served by four of the preventions interventions. Using the Geriatric Depression Scale (GDS), depression was found to be absent for 74 percent (74.1 percent) of the elders served by UTHSC, but present for nearly 26 percent (25.9 percent). About 60 percent (58.8 percent) of elders served by AK DSDS did not exhibit signs of depression, as measured by the Patient Health Questionnaire-9 (PHQ-9), although 10 percent suffered from moderate to severe major depression. Using the PHQ-9, no depression was found for nearly 80 percent of the care recipients served by USC, however, about 20 percent had minimal to minor symptoms (14.7 percent and 4.9 percent respectively). Major depression was limited to less than 2 percent of the care recipients (1.6 percent).

In a similar vein, some elders demonstrated low levels of anxiety. Those assessed by AK DSDS using the General Anxiety Disorder (GAD) scale had a mean score of 4.3, indicating low anxiety. About 30 percent of the elders served by NYSOFA showed no anxiety while the remaining elders (70 percent) exhibited mild to moderate to severe anxiety (25.9 percent, 37 percent, and 7.4 percent, respectively). Having a geropsychiatrist as part of the prevention interventions (NYSOFA, USC) helps allow for appropriate mental health interventions to be put in place.

AK DSDS measured elders' stress using the Perceived Stress Scale (PSS), finding that at-risk elders had a mean score of 18.8 or a moderate level of stress. Conversely, elders served by NYSOFA exhibited higher levels of stress, with almost 20 percent with severe stress (18.5 percent) and more than 50 percent with moderate stress (51.8 percent). About one-fifth exhibited mild stress (18.5 percent) and about one-tenth had no stress (11.1 percent).

Social Support

Elders served by three prevention interventions had low-to-moderate levels of social support but elders experiencing financial exploitation tended to be socially isolated. Victims of self-neglect served by UTHSC had a mean score of 24.3 on the Duke Social Support Index (DSSI), which assesses the social network of the elderly and the support provided by that network. With a mean score of 24.3, these elders indicated a more than moderate level of support. As subset of elders served by AK DSDS had fairly low levels of social support (mean score of 18.9), as measured by the Lubben Social Network Scale (LSNS-R). Care recipients served by USC reported moderate levels of social support (mean score of 14.3) using the Lubben Social Network Scale 6-item (LSNS-6). The NYSOFA teams assessed the level of social support for elders at-risk of financial exploitation to identify sources of support and whether elders felt socially isolated, left the house, had friends or family, had emotional support and if the perpetrator was part of the social support network. While NYSOFA was only able to obtain data on less than half of participants on this measure, of the elders who responded, more than half were socially isolated (59 percent) and the vast majority of perpetrators were part of their social network (80 percent).

3. What are Participants' Risk of Abuse?

Risk of abuse for at-risk elders, victims, and care recipients was measured by four of the prevention interventions using different screening tools or methods, as shown below in Table 5. AK DSDS and USC assessed risk using the VASS, TX/WellMed did so using the EASI screening tool, and NYSOFA assessed the level of risk for financial exploitation based on information provided by collateral contacts such as APS or law enforcement. Elders served by UTHSC had been substantiated for self-neglect by APS.

TABLE 5. Risk of Abuse
  At-risk Elders/Victims Care Recipients
AK DSDS NYSOFA TX/WellMed USC
N % N % N % N %
VASS screening tool 80   ---   ---   76  
Vulnerability                
   (1) Are you afraid of anyone in your family?   8.8   ---   ---   8.8
   (2) Has anyone close to you tried to hurt you or harm you recently?   23.8   ---   ---   1.5
   (3) Has anyone close to you called you names or put you down or made you feel bad recently?   28.8   ---   ---   11.8
Dependence                
   (4) Do you have enough privacy at home?   76.3   ---   ---   94.3
   (5) Do you trust most of the people in your family?   83.4   ---   ---   91.5
   (6) Can you take your own medication and get around by yourself?   80.8   ---   ---   83.8
Dejection                
   (7) Are you sad or lonely often?   31.7   ---   ---   9.9
   (8) Do you feel that nobody wants you around?   15.0   ---   ---   2.9
   (9) Do you feel uncomfortable with anyone in your family?   23.1   ---   ---   26.1
Coercion                
   (10) Does someone in your family make you stay in bed or tell you you're sick when you know you're not?   1.3   ---   ---   4.5
   (11) Has anyone forced you to do things you didn't want to do?   16.3   ---   ---   11.8
   (12) Has anyone taken things that belong to you without your OK?   31.7   ---   ---   11.8
EASI screening tool ---   ---   11,426   ---  
   (1) Have you relied on people for any of the following: bathing, dressing, shopping, banking or meals?   ---       9.6   ---
   (2) Has anyone prevented you from getting food, clothes, medication, glasses, hearing aids or medical care, or from being with people you wanted to be with?   ---       0.3   ---
   (3) Have you been upset because someone talked to you in a way that made you feel shamed or threatened?   ---       1.1   ---
   (4) Has anyone tried to force you to sign papers or to use your money against your will?   ---       0.2   ---
   (5) Has anyone made you afraid, touched you in ways that you did not want, or hurt you physically?   ---       0.3   ---
   (6) Doctor: Elder abuse may be associated with findings such as: poor eye contact, withdrawn nature, malnourishment, hygiene issues, cuts, bruises, inappropriate clothing, or medication compliance issues. Did you notice any of these today or in the last 12 months?   ---       0.3   ---
Stoplight rating                
   Green   ---   ---   98.6   ---
   Yellow   ---   ---   1.1   ---
   Red   ---   ---   0.3   ---
APS history ---   ---   573   ---  
   Yes   ---   ---   52.9   ---
   No   ---   ---   47.1   ---
Risk of Financial Exploitation ---   117   ---   ---  
   No risk   ---   1.7   ---   ---
   Low   ---   1.7   ---   ---
   Medium   ---   16.2   ---   ---
   High   ---   80.3   ---   ---

Risk of Abuse for At-risk Elders and Victims

Elders served through by AK DSDS reported feeling vulnerable and subject to coercion. Using the VASS screening tool, AK DSDS assessed the vulnerability, dependence, sense of dejection, and degree of coercion for the elders participating in the CTI. About one-third (28.8 percent) of elders reported that someone close to them had not treated them well. Nearly one-quarter (23.8 percent) indicated that someone close to them had tried to hurt or harm them. Less than 10 percent (8.8 percent) reported being afraid of anyone in their family. These elders felt they had enough privacy in their home (76.3 percent), could trust most of the people in their family (83.4 percent), and could take their own medication and get around (80.8 percent). Yet, about one-third indicated that they often felt sad or lonely (31.7 percent), that nobody wanted them around (15 percent), and they were uncomfortable with a family member (23.1 percent). About one-third reported that someone had taken their belongings (31.7 percent). Sixteen percent indicated that they had been forced to do something they did not want to do.

The majority of patients screened using the EASI tool in TX/WellMed's intervention were at low risk of abuse. Over half of patients brought to the attention of the APS Specialists, however, had a prior APS case. As part of the primary prevention component of TX/WellMed, over 11,000 patients at 73 of WellMed primary care clinics were screened using the EASI tool to determine their risk of abuse. The vast majority of elders screened fell into the low-risk range (98.6 percent). Thirty-four elders were identified as high-risk (0.3 percent) and referred to APS. About 10 percent of these elders relied on others to help them bathe, dress, shop, bank or prepare meals. About 1 percent reported that someone had shamed or threatened them. In addition to the EASI screening, at-risk elders could be identified through services provided by the two APS Specialists who were embedded at WellMed clinics. The APS Specialists provided consultation to clinical staff through individual inquiries or PCC meetings where patients identified as high-risk (for being hospitalized, discharged home or some other issue, but not necessarily elder abuse) are discussed among a team of WellMed staff.

While the APS Specialists' participation in the PCC meetings was not originally a feature of the intervention, the number of referrals to APS that were generated through the PCC was higher than through the EASI screening tool. Based on data collected by the APS Specialists, 82 patients were identified through the PCC meetings (compared to 35 through the EASI screening tool). An additional 434 patients were referred to the APS Specialists by clinical staff outside of those meetings. WellMed staff included social workers, health coaches, nurses, doctors, and case managers among others. Among the 588 patients brought to the attention of the APS Specialists, 6.1 percent patients had been administered the EASI tool. Of the 36 who were administered the tool, 14.2 percent (or five patients) scored as high-risk. Importantly, of the 573 patients for whom information on APS history was available, a little over half (52.9 percent) had a prior history with APS.

The majority of elders served by the E-MDTs were at high-risk of abuse. NYSOFA assessed the risk of financial exploitation at case intake. Eighty percent were identified as at high-risk, whereas 16 percent demonstrated a medium level of risk. Less than 2 percent were determined to be low-risk.

Most care recipients served by the USC prevention intervention were not dependent, but some reported feeling uncomfortable with family members. The care recipients served by USC's Take AIM project were also assessed using the VASS screening tool. (Not all care recipients responded to each measure though.) At baseline, more than 10 percent (11.8 percent) of elders reported that someone close to them had not treated them well (i.e., put them down, called them names, made to feel bad). Less than 10 percent (8.8 percent) indicated that they were afraid of someone in their family and very few indicated that someone close to them had tried to hurt or harm them (1.5 percent). The majority of care recipients felt they had enough privacy in their home (94.3 percent), could trust their family members (91.5 percent), and were capable of taking their medications and getting around on their own (83.8 percent). However, about 26 percent reported being uncomfortable with a family member, and about 10 percent felt sad or lonely. Very few care recipients had the impression that nobody wanted them around (2.9 percent). In terms of coercion, about 12 percent (11.8 percent) indicated that they had been forced to do something they did not want to do and a similar proportion reported that someone had taken their belongings (11.8 percent). A few elders (4.5 percent) reported being made to stay in bed or told that they were sick (when they knew they were not).

4. What are the Demographic and Household Characteristics of Perpetrators and Caregivers?

The demographic and household characteristics of suspected perpetrators of elder abuse and caregivers are presented in Table 6. Two of the prevention interventions collected information on perpetrators: NYSOFA's E-MDTs and TX/WellMed.

TABLE 6. Demographic and Household Characteristics of Perpetrators and Caregivers
  Perpetrator Care Recipients
NYSOFA TX/WellMed USC
N Mean N Mean N Mean
Age 76 44.5 414 67.2 76 68.9
Gender 170   398   76  
   Male   52.3   41.1   26.3
   Female   47.6   58.9   73.7
Race 119   414   71  
   African American/Black   12.6   8.0   1.4
   American Indian/Native Alaskan   0   0.5   0
   Asian/Pacific Islander   0.8   1.5   9.9
   Caucasian/White   75.6   90.0   88.7
   Other   10.9   0   0
Ethnicity 98   398   76  
   Hispanic   4.1   55.0   6.6
   Not Hispanic   95.9   45.0   93.4
Education 12   ---   76  
   Less than high school   16.7   ---   1.3
   High school   25.0   ---   9.2
   Some college   0   ---   31.6
   College graduate   25.0   ---   23.7
   Some graduate work   33.3   ---   34.2
Income ---   ---   76  
   Less than $15,000   ---   ---   25.0
   $15,000-$25,000   ---   ---   6.6
   $25,001-$35,000   ---   ---   5.3
   $35,001-$50,000   ---   ---   9.2
   $50,001-$75,000   ---   ---   19.7
   $75,001-$100,000   ---   ---   15.8
   Greater than $100,000   ---   ---   18.4
Marital Status 114   ---   76  
   Married   32.4   ---   81.6
   Single   29.8   ---   10.5
   Divorced/Separated   11.4   ---   6.6
   Widowed   2.6   ---   1.3
   Other/Unknown   23.7   ---   0
Primary Language 132   415   76  
   English   96.2   90.1   92.1
   Spanish   2.3   9.6   0
   Other   1.5   0.2   7.9
Place of Residence ---   394   ---  
   Private home*   ---   76.1   ---
   Multi-family home   ---   ---   ---
   Assisted living/Nursing home   ---   4.3   ---
   Group home/Unlicensed personal care home   ---   1.3   ---
   Other   ---   18.2   ---
   Unclear   ---   ---   ---
Relationship to Victim/Care Recipient 150   415   76  
   Self   ---   66.7   ---
   Spouse   6.0   4.1   64.5
   Parent   0   0.5   ---
   Child   37.3   14.9   26.3
   Sibling   0.7   0.2   1.3
   Other Relative   20.6   4.8   5.3
   Other Nonrelative   34.0   8.0   2.6
   Unknown   1.3   0.7   ---
* This includes apartments and rented rooms.

Perpetrators

Alleged perpetrators tended to be middle-aged or elderly, and included both males and females. The mean age of suspected perpetrators identified by the prevention interventions was 44.5 for NYSOFA and 67.2 for TX/WellMed. The mean age of caregivers was 68.9 for USC. The mean age for the elders at-risk of financial exploitation was 80.9, and those served by TX/WellMed was 76.3 years. For perpetrators identified by NYSOFA, 52 percent were male and 47 percent were female. This gender ratio was inverted for the TX/WellMed, as 41 percent of the perpetrators were male and 59 percent were female.

Race and ethnicity was known for only a subset of perpetrators, but they were predominantly Caucasian. For NYSOFA, 76 percent were Caucasian, 13 percent were African American, and 11 percent were noted as "Other." Only 4 percent were identified as Hispanic. Among the perpetrators identified by TX/WellMed, 90 percent were Caucasian and 8 percent were African American. Fifty-five percent of perpetrators were identified as Hispanic. The primary language for perpetrators was English (96 percent for NYSOFA and 90 percent for TX/WellMed). Almost 10 percent of the perpetrators identified by TX/WellMed spoke Spanish; only 2 percent did with NYSOFA.

The educational background of the alleged perpetrators of financial exploitation ranged from those with limited education to the highly-educated. The education levels of suspected perpetrators were identified by NYSOFA's E-MDTs. Seventeen percent had less than a high school education and 25 percent were high school graduates; 25 percent college graduates and 33 percent had some graduate education. Information about the perpetrators' income levels was unknown.

Most alleged perpetrators were family members or relatives. Information was provided about the relationship of the perpetrator to the victim. For suspected perpetrators identified by NYSOFA's E-MDTs, 37 percent were the child of the victim, 34 percent were a relative (such as a grandchild, niece or nephew, or in-law) and 6 percent were a spouse. Less than 1 percent were siblings of the victims. Thirty-four percent were classified as "other nonrelative" which included a girlfriend or boyfriend, friend, paid caregiver, roommate, tenant, or legal guardian. The majority of perpetrators identified by TX/WellMed were the victims themselves. These reflect cases of self-neglect. Nearly 15 percent of the perpetrators were the child of the victim (14.9 percent). Eight percent of perpetrators were nonrelatives, which included a paramour, friend, paid caregiver, service provider, and unrelated home member. Nearly 5 percent were relatives (4.8 percent).11

Caregivers

Caregivers were mostly female, Caucasian, married to the care recipient, college-educated, and had fairly high incomes. Many caregivers were adult children. As shown in Table 6, for the caregivers participating in USC's Take AIM intervention, the mean age was 68.9. About 74 percent of the caregivers were females and 26 percent were male. In terms of race and ethnicity, the majority of caregivers were Caucasian (89 percent) and nearly 10 percent were Asian/Pacific Islanders (9.9 percent). About 7 percent were identified as Hispanic. Caregivers tended to have a college education: 32 percent had some college, 24 percent were college graduates, and 34 percent had graduate school education. Less than 10 percent had only a high school education.

Twenty-five percent of the caregivers had incomes less than $25,000 and nearly 41 percent had incomes between $25,000 and $75,000. Many of the caregivers served by USC were middle income or affluent, with 34 percent having incomes greater than $75,000. Caregivers tended to be married (81.6 percent); some were single (10.5 percent) or divorced/separated (6.6 percent). More than 90 percent spoke English and nearly 8 percent spoke Spanish. There was no information available about the residence of the caregivers, although the relationship of the caregiver to the care recipient was known. Two-thirds of the caregivers were the spouse of the care recipient (64.5 percent) and over one-quarter were the adult child (26.3 percent). About 5 percent were other relatives, such as a grandson or granddaughter.

5. What are the Psychological, Physical Health and Social Conditions of Perpetrators and Caregivers?

The psychological, physical health and social conditions of perpetrators was identified by NYSOFA and TX/WellMed, as well as the caregivers participating in USC's intervention, and the results are shown in Table 7. Although sparse information is available it sheds light on elders' risk for abuse.

TABLE 7. Physical Health, Psychological and Social Characteristics of Perpetrators and Caregivers
  Perpetrator Caregiver
NYSOFA
(N=221)
TX/WellMed
(N=134)
USC
(N=76)
N % N % N %
Mobility Impaired   ---   3.0   ---
Physical Disability   ---   6.0   ---
Depression
   PHQ-9 ---   ---   75  
      No depression (04)   ---   ---   40.0
      Minimal symptoms (5-9)   ---   ---   37.3
      Minor symptoms (10-14)   ---   ---   16.0
      Major depression, minor (15-19)   ---   ---   5.3
      Major depression, severe (GE 20)   ---   ---   1.3
   Depression 5   ---   ---  
      No   20.0   ---   ---
      Yes   80.0   ---   ---
      Moderate   0   ---   ---
      Severe   0   ---   ---
  N Mean N Mean N Mean
Anxiety
   GAD --- --- --- --- 76 5.5
  N % N % N %
Social Support
   LSNS-6 --- --- --- --- 75 17.2
  N Mean N Mean N Mean
Social Support
   Socially isolated 29 0.1 --- --- --- ---
   Leave the house 19 1.0 --- --- --- ---
   Friends or family 17 1.0 --- --- --- ---
   Emotional support 12 1.0 --- --- --- ---
Caregiver Burden
   Zarit Burden Interview --- --- --- --- 76 8.2
Potential Substance Dependency--CAGE --- --- --- --- 11 1.0
  N % N % N %
History of substance abuse 29 82   ---   2.6
History of alcohol abuse 4 25   ---   12.0
Alcohol abuse --- ---   1.0   ---
History of violence 19 63   ---   13.3
History of abuse 9 55   ---   55.3
APS history --- ---   ---   32.9

Alleged perpetrators tended to have issues with substance abuse, regardless of health status. In terms of physical health, perpetrators assessed by TX/WellMed were able-bodied, finding that only 3 percent were mobility impaired and only 6 percent were physically disabled. They also were prone to alcohol abuse. For the perpetrators identified by NYSOFA, 80 percent experienced depression, yet they were not socially isolated and appeared to have had social support. Perpetrators tended to have a history of substance abuse, coupled with a history of violence and abuse.

Although exhibiting low levels of anxiety and burden, and with moderate levels of support, caregivers showed signs of depression. USC assessed caregivers on multiple dimensions, using a battery of measures to gauge depression, anxiety, and burden. Using the PHQ-9, no depression was found for 40 percent of the caregivers, however, about 53 percent had minimal to minor symptoms (37.37 percent and 16 percent respectively). Major depression was found for almost 7 percent of the caregivers (6.6 percent). Using the GAD scale, caregivers had a mean score of 5.5, indicating low anxiety. Caregivers served by USC reported moderate levels of social support (mean score of 17.2) using the LSNS-6. Based on the scores assessed using the Zarit Burden Interview, caregivers reported little burden with regard to the impact of the dementia patient's disabilities on the caregiver's life (mean score of 8.2). They also showed little propensity for developing a potential dependency on alcohol, given the mean CAGE substance abuse screening tool score of 1.0.

6. What is the Frequency of Abuse Types, by Site?12

Self-neglect was the most common type of abuse and co-occurred with all forms of abuse. The frequency of abuse types for each prevention intervention is presented in Table 8. Three of the four prevention interventions for which data were applicable reported high rates of elder self-neglect.13 For the UTHSC prevention intervention, 86 percent (85.7 percent) of elders were substantiated by APS for self-neglect and about 4 percent were substantiated for medical self-neglect. Fifty-five percent of elders served by TX/WellMed were also identified for self-neglect. Forty-six percent of the cases served by AK DSDS involved self-neglecting elders. Across all four prevention interventions, a substantial proportion of cases involved co-occurring forms of abuse. Self-neglect co-occurred with all forms of abuse.

TABLE 8. Distribution of Abuse Types/Allegations, by Site
Type of Abuse UTHSC
(N=28)
AK DSDS
(N=63)
NYSOFA
(N=182)
TX/WellMed
(N-310)
N % N % N % N %
Physical 0 0 1 1.6 2 1.1 5 1.6
Psychological 0 0 2 3.2 1 0.5 5 1.6
Sexual 0 0 1 1.6 0 0 1 0.3
Financial exploitation 0 0 2 3.2 99 54.4 2 0.6
Neglect 0 0 2 3.2 0 0 5 1.6
Self-neglect 24 85.7 29 46.0 1 0.5 169 54.5
   Medical self-neglect 1 3.6 0 0 0 0 0 0
Co-occurring 3 10.7 26 41.3 79 43.4 123 30.9

Financial exploitation co-occurred with other forms of abuse. More than half of the elders served through NYSOFA's E-MDTs had experienced financial exploitation; this was the specific focus of their intervention. However, financial exploitation was identified as a concern with other interventions and populations. A very small percentage of elders served for the AK DSDS and TX/WellMed prevention interventions experienced financial exploitation (3.2 percent and 0.6 percent, respectively).

7. What is the Frequency of Single, versus Multiple Types of Abuse, by Site?14

Many elders served by the preventions interventions experienced more than one type of abuse. As shown in Table 9, across the four prevention interventions that were serving elders who were substantiated victims or at-risk of elder abuse, more than half of them had experienced one form of abuse. This percentage ranged from 89 percent for UTHSC to 57 percent for NYSOFA. Yet many elders experienced more than one type of abuse. Twenty-five percent of the elders served by AK DSDS and TX/WellMed experienced two forms of abuse (25.4 percent and 25.8 percent, respectively), as did 15 percent of elders served by NYSOFA. More than 25 percent of the elders served by NYSOFA experienced 3-5 types of abuse, as did 15 percent of elders served by AK DSDS and almost 14 percent by TX/WellMed.

TABLE 9. Number of Abuse Types/Allegations, by Site
Number of Abuse Types UTHSC
(N=28)
AK DSDS
(N=63)
NYSOFA
(N=182)
TX/WellMed
(N-310)
N % N % N % N %
1 25 89.3 37 58.7 103 56.6 187 60.3
2 3 10.7 16 25.4 28 15.4 80 25.8
3 0 0 4 6.3 18 9.9 27 8.7
4 0 0 4 6.3 4 2.2 9 2.9
5 0 0 2 3.2 24 13.2 6 1.9
6 0 0 0 0 5 2.7 0 0
7 0 0 0 0 0 0 0 0
8 0 0 0 0 0 0 1 0.3

B. Service Referral and Utilization

In addition to data elements used to create the risk factor profiles, grantees collected information on the referral source and service utilization for elder victims and care recipients. The nature of risk for abuse, elder characteristics, and the types of services referred and received may play important roles in achieving outcomes for each of the respective interventions.

As part of their prevention interventions, grantees tracked and documented the service referrals and in some cases, services provided. Using data from the grantees, we describe the types of services for which elders were referred and their average duration. Based on this information, we present the type and mix of services and the length of the intervention in a series of tables below. We point out, however, that some grantees played direct roles in service delivery (UTHSC, AK DSDS, USC) while others acted as coordinators for treatment or services (TX/WellMed, NYSOFA).

1. What are the Types and Frequencies of Services that are Referred to/Received by Participants?

Below we describe the types of service referrals made on behalf of elder served by the prevention interventions. They are presented according to their status as substantiated abuse victims (UTHSC), at-risk elders or know abuse (NYSOFA, TX/WellMed) and care recipient/caregiver dyads (USC).

Many self-neglecting elders served by UTHSC were in need of social work services. Through UTHSC's medication adherence intervention, all elders with APS substantiated self-neglect received social support, education and medication management during monthly, one-hour home visits. These check-ins also revealed areas where additional supports were needed and where UTHSC staff made referrals for services. As shown in Table 10a below, 60 percent of the elders were referred to social work services. About 25 percent were referred to provider services or skilled nursing and 14 percent were in need of medication. Referrals were made for utility payments, rental assistance, and home repairs.

TABLE 10a. Service Referral and Utilization, UTHSC
Type of Service Received Victims (N=28)
N
(times services were delivered)
%
Home visits 580 100
Phone calls 2,353 100
Pill counts 302 100
Education 139 100
Type of Service Referred N %
Social work services 17 60.7
Medication 4 14.3
Utilities 5 17.9
Rent 1 3.6
Provider services/Skilled nursing 7 25.0
Home repair 1 3.6
NOTE: Only individuals who completed the intervention were included here. Individuals can fit into more than one category.

Elders participating in the AK DSDS's prevention intervention had multiple service needs. All participating elders received frequent home visits and phone calls during the first three months of the intervention. Table 10b below presents common areas of assistance provided to participants. This included education and advocacy relating to housing, income and benefits, systems navigation, transportation and basic needs. Information on the specific number of individuals who received these services, however, is not available.

TABLE 10b. Service Referral and Utilization, AK DSDS
Type of Service Received At-risk Elders/Victims
(N=88)
N %
Home visits   100
Phone calls   100
Education/Assistance/Advocacy   NA
   Housing    
   Income and benefits advocacy and support    
   Systems navigation and service access    
   Transportation    
   Basic needs (food, rent and utility assistance)    
NA refers to not available.

Financial and legal interventions were the most common service needs of the elders served by NYSOFA's E-MDTs. For elders who were the victims of financial exploitation, the most common type of referral was financially-related (30 percent), such as requesting a bank hold or freezing accounts, cancelling credit or debit cards, contacting fraud alert departments, etc. As shown in Table 10c, this was followed by referrals for legal assistance or criminal justice intervention (22 percent), social or protective services (14 percent), and efforts to involve law enforcement (11 percent). Across the cases served, there were more than 200 referrals for professional services, of which 13 percent were for analysis by a forensic accountant and referrals for mental health evaluations (12 percent) or follow-up services with a mental health provider (4 percent).

TABLE 10c. Service Referral and Utilization, NYSOFA
Type of Service Referred At-risk Elders/Victims
(N=224)
N %
Financial 403 29.9
Home safety 17 1.2
Law enforcement 144 10.7
Legal/Criminal justice 291 21.6
Living arrangement/Housing 57 4.2
Medical/Physical 113 8.4
Mental health and cognitive issues 102 7.6
Social services/Protective services 187 13.9
Social support/Integration/Network 15 1.1
Perpetrator interventions 13 1.0
Victim-perpetrator relationship 0 0
Other 3 0.2
Type of Professional Referrals N %
Forensic accountant assistance    
   Analysis 29 12.9
   Obtaining documents 0 0
   Other 1 0.4
Geropsychiatric assistance    
   Refer for mental health evaluation 27 12.1
   Follow-up with mental health professional 8 3.6

Most elders served by TX/WellMed were not at-risk of abuse, but some needed assistance from APS. As part of their prevention intervention, TX/WellMed incorporated use of the EASI screening tool into the electronic health record in primary care settings. Using a "stoplight" approach to identify risk, elders received services according to the level of perceived risk (low/green, medium/yellow, high/red). Table 10d indicates that the vast majority of elders screened fell into the low-risk range (98.6 percent) and received patient education materials only. However, a fraction of all elders screened--less than 1 percent--were identified as high-risk and referred to both the Complex Care worker and to APS. These referrals sparked a number of actions by APS, commonly for counseling and education, referrals to targeted services, and to a lesser extent, mediation or to purchase services on behalf of the client such as Meals on Wheels. In some cases, no action was taken.

TABLE 10d. Service Referral and Utilization, TX/WellMed
Type of Service Received N %
WellMed patients    
Green rating on EASI tool 11,266 98.6
   (1) Provision of patient education materials related to the prevention of elder abuse (green rating on EASI)    
Yellow rating on EASI tool    
   (1) Provision of patient education materials related to the prevention of elder abuse
   (2) Referral to Complex Care (yellow rating on EASI)
125 1.1
Red rating on EASI tool    
   (1) Provision of patient education materials related to the prevention of elder abuse
   (2) Referral to Complex Care
   (3) Referral to APS
35 0.3
APS Referred/Received N
(times services referred/received)
%
Counseling/Education 417  
Referral 182  
Purchase client services 45  
Mediation 69  
County services 1  
No action 58  
State: DADS 6  
Assistance/Documentation assistance 27  
Benefits 2  

The main service needs of care recipient/caregiver dyads were for in-home care as well as financial and legal assistance. USC's prevention intervention focused on 76 dyads in which one member had dementia. Risks assessments prompted linking one or both persons to community-based services to address identified needs. As shown in Table 10e, the need for an in-home caregiver was identified for 27 percent of the care recipients and 47 percent of the caregivers. Referrals for legal advice/assistance (24 percent and 25 percent) and financial planning or assistance were common (19 percent and 29 percent) across care recipients and caregivers, respectively. Additionally, 17 percent of care recipients and 41 percent of the care recipients were referred to a memory loss support group. Referrals for various forms of psychiatric intervention and care were made for both care recipients and caregivers (between 10-12 percent of care recipients and 3-9 percent of caregivers). Twenty percent of caregivers were referred to respite care.

TABLE 10e. Service Referral and Utilization, USC
Type of Service Referred to Dyad Care Recipient
(N=76)
Caregiver
(N=76)
N % N % N %
Memory loss support group 10 17.2 31 41.3    
Memory loss education program 8 13.6 20 26.7    
In-home caregiver 16 27.1 35 46.6    
Individual psychotherapy 7 12.1 5 6.7    
Conjoint psychotherapy 6 10.3 2 2.7    
Psychiatric care (for the caregiver) 6 10.2 7 9.3    
Legal advice/assistance (related to caregiving role) 14 24.1 19 25.3    
Financial planning/assistance (related to caregiving role) 11 19.3 22 29.3    
Respite Care --- --- 15 20.0    
Type of Service     Service Provided Service Completed
N % N %
Caregiver resource center     35 53.8 3 8.6
Savvy caregiver course     25 38.5 0 0
In-home caregiver agency     20 30.8 1 5.0
Support groups     23 35.4 0 0
Legal aid services     57 87.7 17 29.8
Geropsychiatry     14 21.5 1 7.1
Family counseling     23 35.4 1 4.3
Individual counseling     19 29.2 1 5.3
Friendly visitor program     58 89.2 8 13.8
Adult day care     24 36.9 2 8.3

For the caregiver, USC tracked they type of service referral, whether it was actually provided, and if the caregiver completed or followed through with the activity. Nearly 90 percent of caregivers were referred to the Friendly Visitor program and Legal Aid services. Of the caregivers referred to the Friendly Visitor program, about 14 percent availed themselves of this service. Of the caregivers referred to legal aid services, about 30 percent followed through with this activity. While more than 50 percent of care recipients were referred to the Caregiver Resource Center, only 9 percent of the caregivers followed through with the service. Additionally, at least one-third of the caregivers were referred to the Savvy Caregiver course, adult day care, support groups, and family counseling. Yet, there was little uptake for these services on the part of the caregivers (0 percent, 8.3 percent, 0 percent, and 4.3 percent, respectively).

2. What is the Duration of the Intervention? What Percentage of Participants Completed the Intervention Protocol? What Percentage of Participants Partially Completed all the Intervention Components (but did not drop out)? What is the Percentage of Participant Attrition?

The duration of the five prevention interventions depended on the type of victimization or risk addressed (e.g., substantiated, at-risk) and/or the nature of the treatment or intervention protocol.15

Prevention interventions were conducted within the time period designated by the protocol. As presented in Table 11, three of the prevention interventions had a defined period of service delivery associated with an evidence-based (AK DSDS) or theory-informed intervention (UTHSC, USC). Across these prevention interventions, on average, services were completed within the time period designated by the protocol. Based on the case start and closure dates, the mean amount of time for full completion of AK DSDS's CTI was 9.6 months. Among the elders that only partially completed the intervention, the mean was 4.2 months. For the elders taking part in UTHSC's medication adherence intervention, the mean amount of time was 6.4 months to complete the treatment protocol. For all of the care recipients and caregivers that participated in USC's Take AIM intervention, the duration for each dyad was 3.0 months.

TABLE 11. Characteristics of Elder Abuse Prevention Intervention Protocols, by Site
Intervention Characteristics Victims At-risk Elders/Victims Care Recipients Caregivers
UTHSC AK DSDS NYSOFA TX/WellMed USC
N Mean N Mean N Mean N Mean N Mean N Mean
Duration 23   88   103       76   76  
   Duration of treatment (months) for full completion 20 6.4 54 9.6 77 9.1 --- --- 56 3.0 58 3.0
   Duration of treatment (months) for partial completion 3 4.0 33 4.2     --- ---        
  N % N % N % N % N % N %
Case Duration             296          
   0-30 days             53 17.9        
   31-60 days             64 21.6        
   61-90 days             56 18.9        
   91-120 days             45 15.2        
   121-150 days             23 7.8        
   151-180 days             28 9.5        
   Over 180 days             27 9.1        
Protocol completion 34   87   222   ---   76   76  
   Full 20 58.7 54 62.1 103 46.4   ---   73.7   73.7
   Partial/Dropped 14 41.2 33 37.9 119 53.6   ---   26.3   26.3

For prevention interventions where APS was a service partner, case duration ranged from three months to nine months. Two prevention interventions followed APS practices. The same period of service delivery was not predetermined for all participants. Investigations were handled on a case-by-case basis and were open-ended in duration, resulting in case closure once the desired outcome was achieved (NYSOFA, TX/WellMed). The mean amount of time for the NYSOFA E-MDTs to resolve an elder's case of financial exploitation was 9.1 months, from the point of intake to the date that the case was considered inactive and an outcome achieved. There was considerable variability in case duration for elders served by TX/WellMed. Of the 296 WellMed patients who were served by APS for whom data are available, the majority of cases (58.4 percent) were handled within 90 days. Almost 20 percent of cases took over 151 days to address.

Between 50-75 percent of elders completed the intervention protocol or had their case resolved for four of the prevention interventions. Nearly 60 percent of the elders participating in UTHSC's medication adherence intervention, completed the full protocol and 40 percent partially completed it or dropped out. A similar percentage of elders (62 percent) completed the CTI protocol with AK DSDS. However, about 38 percent either partially completed or did not complete the protocol (due to inability to locate, death, moving out of the service area, or that it was unsafe for the case manager to continue the services). About 46 percent of the elders served by NYSOFA's E-MDTs are counted as having completed the protocol, given that their cases were closed. About 54 percent of the financial exploitation cases were still pending (at the time the dataset was obtained). For the care recipient/caregiver dyads served by USC, 74 percent full completed the protocol and 26 percent only partially completed.

C. Outcomes Achieved

We attempted to identify outcomes that were relevant across five grantees, taking into consideration the variety of prevention interventions implemented and the specific nature of elder abuse involved. To this end, we were able to examine intervention-specific outcomes and changes in key characteristics of elders. Findings are presented in Table 12 below. We examined whether cases had been referred to APS once the intervention had been completed (i.e., recidivism for those with prior APS histories) for a subset of elders served by AK DSDS and UTHSC.16 For NYSOFA, we examined outcomes achieved regarding financial exploitation. For TX/WellMed, APS data collected on reasons for case closure are presented. Changes measured in elders' state of vulnerability, characteristics or circumstances varied across the grantees. As the type of change was intervention-specific, the measures and quantity of data available also varied.

1. What were the Outcomes of the Elders' Participation in the Intervention?

About half of the elders served by UTHSC and AK DSDS prevention's intervention did not have a re-referral to APS. Almost one-half of the elders participating in UTHSC's medication adherence intervention did not have a subsequent referral to APS once the treatment protocol was completed, meaning that the elders were not reported for self-neglect or another form of abuse. However, about one-third of the elders did have a subsequent referral to APS and 14 percent had two referrals.17 For the elders served by AK DSDS for whom data are available, 90 percent did not have a subsequent referral to APS. Only 6 percent were referred to APS within three months of the intervention and 3 percent were referred with six months.

TABLE 12. Description of Key Outcomes, by Site
Outcome UTHSC
(N=28)
AK DSDS
(N=63)
NYSOFA
(N=103)
TX/WellMed
(N=296)
N % N % N % N %
APS Referral 28   63   ---   ---  
3 months post-intervention --- ---   6.3   ---   ---
6 months post-intervention --- ---   3.2   ---   ---
1 subsequent referral 9 32.1   ---   ---   ---
2 subsequent referrals 4 14.3   ---   ---   ---
Financial Exploitation ---   ---   103*   ---  
Reduction in exploitation of assets   ---   --- 65 81.5   ---
End to exploitation of assets   ---   --- 58 68.9   ---
Funds spent on appropriate care   ---   --- 49 83.7   ---
Value of assets protected   ---   --- 48     ---
   Value protected   ---   ---   50.0   ---
   Value protected pending   ---   ---   50.0   ---
Restitution of assets   ---   --- 7     ---
   Assets restored   ---   ---   71.4   ---
   Asset restitution pending   ---   ---   28.6   ---
Recovery of assets   ---   --- 9     ---
   Assets recovered   ---   ---   44.4   ---
   Asset recovery pending   ---   ---   55.6   ---
APS Investigation Closure Reason             296  
Valid, continue as APS --- --- --- --- --- --- 141 47.6
Resolved during investigation with service --- --- --- --- --- --- 19 6.4
Progress to ICS --- --- --- --- --- --- 20 6.8
Progress to maintenance --- --- --- --- --- --- 4 1.4
No services required --- --- --- --- --- --- 30 10.1
Services needed, but not available --- --- --- --- --- --- 1 0.3
Client refused services --- --- --- --- --- --- 2 0.7
Client died --- --- --- --- --- --- 11 3.7
Moved/Unable to locate --- --- --- --- --- --- 3 1.0
Unable to determine --- --- --- --- --- --- 9 3.0
Invalid --- --- --- --- --- --- 56 18.9
*N excludes data that were missing, unknown, cases that are not applicable or were classified as other.

Intervention by NYSOFA's E-MDTs stopped financial exploitation of elder assets. Outcomes for a subset of elders served by NYSOFA Lifespan and NYCEAC included a reduction in the exploitation of assets (81.5 percent), an end to the exploitation of assets (68.9 percent), and having funds spent on appropriate care (83.7 percent). A subset of elders had their assets protected (50 percent) or their assets restored (71.4 percent). For others, protection or restitution of assets was pending.

Despite screening and identification of the majority of elders as at a low-risk of abuse, there were cases brought to the attention of APS by TX/WellMed. About 19 percent of cases of suspected elder abuse of WellMed patients that were referred to APS for service were closed for being invalid, meaning that there was no indication that the alleged victim was in a state of abuse, neglect, or financial exploitation following a thorough investigation. Six percent of cases were noted as resolved during an investigation with services provided. Yet almost 50 percent of the cases brought to the attention of APS were identified as valid and progressed to the stage of being an active case with services provided. Another 6 percent of cases were flagged as in need of intensive services, as elders were identified as being at moderate to high-risk of recidivism and in need of services to remedy the root cause of the abuse, neglect, or financial exploitation. Ten percent were closed as no services were required. However, in these cases the alleged victim was experiencing some form of mistreatment, but APS intervention was not necessary to resolve the problem.

2. Did Participants' Level of Social Support Increase Over the Course of the Intervention? Did Participants become more Independent/less Dependent on Others over the Course of the Intervention? Did Participants' Sense of Safety Increase over the Course of the Intervention? Were there Changes in Caregiver Behaviors over the Course of the Intervention?

As shown in Tables 13a-13d, elders' were assessed on a variety of measures to determine changes in well-being, degree of risk and social support, etc. that may be attributed to the prevention intervention.

The moderate level of social support and physical functioning reported for elders served by UTHSC at baseline remained stable following the intervention. For elders served by the UTHSC, the mean baseline score on physical functioning (as measured by the SF-36 Health Survey) was 17.1, suggesting moderate functioning. The score increased slightly by the end of the intervention to 19.3, but regressed at three months and six months post-intervention. Social support was measured using the DSSI and assessed the social network of the elderly and the support provided by that network. Baseline scores indicate a moderate level of support with a mean score of 24.3 with a slight increase by the end of the intervention (25.7). This level of social support was sustained following the intervention (24.0 at three months and 24.5 at six months).

TABLE 13a. Change in Key Characteristics among Participants, UTHSC
  Baseline End of Intervention 3 Months Post-Intervention 6 Months Post-Intervention
N Mean N Mean N Mean N Mean
Physical Function (10-30) 30 17.1 19 19.3 13 18.8 10 16.8
Social Support--DSSI (11-33) 31 24.3 20 25.7 13 24.0 15 24.5

While elders served by AK DSDS reported less vulnerability following the intervention, they may have been susceptible to harm by others. Using the VASS screening tool, AK DSDS assessed the vulnerability, dependence, sense of dejection, and degree of coercion experienced by the elders participating in the CTI at multiple points in time. As shown in Table 13b, at baseline, over one-quarter (28.8 percent) of elders reported that someone close to them had not treated them well (i.e., put them down, called them names, made to feel bad). By the end of the intervention, this had decreased to about 3.5 percent. However, six months post-intervention, for elders contacted, the percentage increased to 9 percent, suggesting the need for follow-up assistance. A similar pattern was found with respect to risk of harm, with nearly one-quarter (23.8 percent) of elders at baseline indicating that someone close to them had tried to hurt or harm them. This decreased to about 7 percent at the end of the intervention, and then increased to about 18 percent post-intervention, indicating that elders may have been at imminent risk.

TABLE 13b. Change in Key Characteristics among Participants, AK DSDS
  Baseline End of Intervention 3 Months Post-Intervention 6 Months Post-Intervention
N % N % N % N %
VASS 80   39   34   12  
Vulnerability                
   (1) Are you afraid of anyone in your family?   8.8   3.5   6.1   9.1
   (2) Has anyone close to you tried to hurt you or harm you recently?   23.8   6.9   6.1   18.2
   (3) Has anyone close to you called you names or put you down or made you feel bad recently?   28.8   3.5   6.1   9.1
Dependence                
   (4) Do you have enough privacy at home?   76.3   79.3   84.9   81.8
   (5) Do you trust most of the people in your family?   83.4   100   93.9   54.6
   (6) Can you take your own medication and get around by yourself?   80.8   75.9   81.8   45.5
Dejection                
   (7) Are you sad or lonely often?   31.7   22.2   19.5   36.4
   (8) Do you feel that nobody wants you around?   15.0   10.3   9.7   9.1
   (9) Do you feel uncomfortable with anyone in your family?   23.1   11.1   6.9   10.0
Coercion                
   (10) Does someone in your family make you stay in bed or tell you you're sick when you know you're not?   1.3   0   3.2   10.0
   (11) Has anyone forced you to do things you didn't want to do?   16.3   10.3   3.1   18.2
   (12) Has anyone taken things that belong to you without your OK?   31.7   10.3   12.5   27.3
Social support                
Professional network ---   37   31   13  
   Yes   ---   70.2   61.3   84.6
   Somewhat   ---   8.1   12.9   7.7
   No   ---   21.6   25.8   7.7
Family/Friend network ---   33   26   11  
   Yes   ---   72.7   69.2   90.9
   Somewhat   ---   12.1   7.7   0
   No   ---   15.1   23.1   9.1

On measures of dependence, elders reported that they had sufficient privacy at home both at baseline and at the end of the intervention. This remained fairly stable at three and six months post-intervention. Elders' sense of trust increased during the intervention from 83 percent to 100 percent, but diminished considerably six months later to 55 percent. They also were less able to take their medication and get around independently six months later. While elders were less likely to feel dejected over time, feelings of sadness of loneliness did not change. Concerning trends were evident regarding measures of coercion over time, as they either increased or stayed the same. Yet, the majority of elders reported having social support increase over time from both professional networks and among their family and friends.

High risk of financial exploitation decreased for a subset of elders served by one of NYSOFA's E-MDTs. The level of risk of financial exploitation at baseline was deemed "high" for a subset of elders (80.3 percent) served by NYSOFA's E-MDT operating in the Finger Lakes Region. As shown in Table 13c, by the end of the intervention, this perception had decreased to 11 percent, and about 72 percent of elder were at low-risk.18 It should be noted, however, that a response of "No Risk" was not an option at the end of the intervention, only during intake.

TABLE 13c. Change in Key Characteristics among Participants, NYSOFA (Lifespan)
  Baseline End of Intervention
N % N %
Risk of Financial Exploitation 117   46  
   No Risk   1.7   NA
   Low   1.7   71.7
   Medium   16.2   17.4
   High   80.3   10.9

Care recipients' sense of vulnerability and coercion fluctuated over the course of the USC intervention. As shown in Table 13d, a subset of care recipients served by USC's Take AIM project were also assessed using the VASS screening tool. Over the course of the intervention, care recipients reported that someone close to them had not treated them well (i.e., put them down, called them names, made to feel bad), changing slightly from 12 percent to 18 percent. Levels of trust with family members remained stable over time (91.5 percent to 91.3 percent). Yet elders reported feelings of dejection increased from baseline to the end of the intervention. While there was little change in the degree of coercion experienced overall, elders reported being forced to do things they did not want to do (increasing from 11.8 percent to 15.6 percent).

Care recipients' level of dependency did not change appreciably, although more needed assistance with taking medication. Dependence remained essentially the same over three months. The physical functioning of the care recipients that participated in USC's Take AIM was assessed through the ADLs measure. ADL functions are essential for meeting basic needs (e.g., dressing and feeding oneself). The care recipient's ADL score was reported by the caregiver. The measure indicates whether the care recipient can perform an activity independent of the caregiver or whether s/he is dependent on the caregiver. At baseline the mean score was 3.5. Although it decreased slightly at the end of the intervention period to a mean of 3.2, the scores reflect a moderate level of functioning with some minimal loss of the ability to care for oneself.

Physical function of the care recipient was also measured by the IADL scale. This was reported by the caregiver. IADL functions are more concerned with independent living skills rather than basic ADLs. This includes the ability to use the telephone, shop, prepare food, do housekeeping, take medications, and handle finances. At baseline and at the end of the intervention, about 92 percent of care recipients were assessed as dependent.

TABLE 13d. Change in Key Characteristics among Participants, USC
  Care Recipients Caregivers
Baseline End of Intervention Baseline End of Intervention
N % N % N % N %
VASS
Vulnerability 76   46   ---      
   (1) Are you afraid of anyone in your family? 6 8.8 4 8.9 --- --- --- ---
   (2) Has anyone close to you tried to hurt you or harm you recently? 1 1.5 1 2.2 --- --- --- ---
   (3) Has anyone close to you called you names or put you down or made you feel bad recently? 8 11.8 8 18.2 --- --- --- ---
Dependence                
   (4) Do you have enough privacy at home? 66 94.3 45 95.7 --- --- --- ---
   (5) Do you trust most of the people in your family? 65 91.5 42 91.3 --- --- --- ---
   (6) Can you take your own medication and get around by yourself? 57 83.8 30 66.7 --- --- --- ---
Dejection                
   (7) Are you sad or lonely often? 7 9.9 7 15.2 --- --- --- ---
   (8) Do you feel that nobody wants you around? 2 2.9 5 11.4 --- --- --- ---
   (9) Do you feel uncomfortable with anyone in your family? 18 26.1 10 22.2 --- --- --- ---
   Coercion                
   (10) Does someone in your family make you stay in bed or tell you you're sick when you know you're not? 3 4.5 2 4.4 --- --- --- ---
   (11) Has anyone forced you to do things you didn't want to do? 8 11.8 7 15.6 --- --- --- ---
   (12) Has anyone taken things that belong to you without your OK? 8 11.8 4 9.1 --- --- --- ---
  N Mean N Mean N Mean N Mean
IADLs 74   55   ---   ---  
   Independent (8-10)   1.3   0        
   Moderate (11-16)   6.8   7.3        
   Dependent (17-31)   91.9   92.7        
ADLs 75 3.5 55 3.2 --- --- --- ---
Social Support (LSNS-6) 62 14.3 40 14.5 75 17.2 55 17.7

The degree of social support remained constant for care recipients and caregivers served by USC. Social support, as measured by the LSNS-R, assesses the frequency of contact and quality of contact that one has with family and friends and the extent to which one could confide in or ask them for assistance. Care recipients served by USC reported a moderate level of social support at baseline, with a mean score of 14.3. By the end of the intervention, this had remained virtually the same, with a mean score of 14.5. A similar shift from baseline to the end of intervention was found for the caregivers, with a slight increase in the mean score from 17.2 to 17.5.

D. Replication and Lessons Learned

Potential for Replication

Overall, key stakeholders found that the prevention interventions they had implemented could be replicated in other locales under similar conditions. Lessons learned also pointed to some modifications that would be helpful for future implementation.

For AK DSDS's implementation of the CTI, key informants indicated that the intervention could be easily replicated with similar populations or in other communities where there is easy access to community resources. No single feature was highlighted as being more amenable to replication than others. As the intervention relies on existing service infrastructure, implementing the program in geographically dispersed communities with few services would be difficult.

Stakeholders noted that with established knowledge of the community resources and their accessibility, USC's Take AIM program could be replicated and adapted to other communities. Most amenable to replication is the targeted, multi-disciplinary approach to identifying risk factors and appropriate resources for the care recipient/caregiver dyad. Team members come from a variety of backgrounds, both medical and social, and contribute to this multi-pronged approach.

UTHSC stakeholders reported that the medication adherence prevention intervention could be replicated with similar populations in other states. Features of the prevention intervention that are well-suited to replication are the educational component to increase health literacy, weekly personal contact and check-ins, and the use of environmental cues. The intervention requires a registered nurse and research staff with pharmacy and geriatrics background, and a geriatrician to conduct the medication reconciliation.

Stakeholders thought that it was very realistic to replicate the E-MDT in other jurisdictions and with similar population of vulnerable elders, although adjustments would be needed depending on the context of implementation. Based on the Lifespan and NYCEAC's experience, stakeholders identified some prerequisites or key ingredients for implementing an E-MDT. The convening organization must have a good relationship with the community in order to build and sustain the team. Key personnel that are necessary for implementation include a forensic accountant and geriatrician for consultation, plus an E-MDT coordinator with knowledge of financial exploitation and ability to work with professionals across systems (i.e., legal, social services, financial services and banking industry). From a legal perspective, there may be some restrictions on confidentiality and the use of power of attorney. The E-MDT might need adjustment to align with the legal framework in a different jurisdiction, as well as the service delivery system in the community.

TX/WellMed stakeholders indicated that the intervention could be replicated, either as a whole or each component separately although certain components may be replicated more easily than others. Many felt the EASI tool was the easiest to replicate given its ready availability. Replicating the role of an APS specialist embedded in other organizational entities, such as a clinical setting or hospital, to facilitate risk assessments, provide educational training, and coordinate care may be challenging, and would require establishing expectations and protocols. It was noted that the intervention is well-suited to implementation in a state-administered APS system and perhaps less conducive to county-administered systems.

Across the prevention interventions, having an adequate array of aging and elder services and relationships with community-based providers to facilitate recruitment and referrals would be necessary contextual components. Having an outreach component with community organizations or service networks is necessary to establish referrals, both for enrollment and services. Resource- or program-rich communities have an array of services and supports to offer at-risk elders and victims, making it easier to make connections to address abuse, neglect and exploitation and to meet co-occurring needs. Jurisdictions that lack resources may need to seek partnerships with the public and private sector in order to bolster its service array.

E. Lessons Learned

1. Implications for the Elder Abuse Field

Lessons learned from the prevention interventions focused mainly on how one approaches working and conducting research with vulnerable elders while addressing risks of abuse, co-occurring needs, and co-morbid conditions. Additional lessons learned stemmed from forming partnerships across systems and working together to meet elders' needs.

Working with Vulnerable Elders

UTHSC noted that developing a helping relationship with vulnerable elders and becoming a needed, dependable presence in their lives was a "huge responsibility." Implementing the prevention intervention required compassion, patience, and professionalism, and constant mindfulness of ethical practice and research. Balancing elder autonomy and safety was an ever-present concern. Rapport building, empathy, commitment, and gentle persistence were essential. AK DSDS found that it is important to adjust to seniors' needs and timeline and support their decisions when they are ready, particularly as habits have been strongly formed over a lifetime.

AK DSDS found that the CTI appeared to benefit some types of elders and forms of abuse more so than others. More progress was seen with cases of self-neglect and those cases that had not risen to the level of needed protective services. The CTI appeared to be less successful for elders experiencing significant mental health issues, dementia, substance abuse issues, ongoing physical abuse, sexual abuse, and domestic violence. Cases with public guardians who had conflicting goals with their wards or where the abusing caregiver remained in the home also tended to be less successful.

TX/WellMed echoed these concerns, noting that the prevention intervention addressed not only elder abuse, neglect, and exploitation but other issues facing elders, including poverty. As they learned, no one agency can address the issue of elder abuse alone: health and human services need to work together to meet the needs of vulnerable elders.

Working Across Organizations

As USC found, pre-existing and positive working relationships with community members provided a foundation to implement the prevention intervention and access available resources. Starting up a new prevention intervention requires understanding of roles and recognizing limits. NYSOFA found that designing and implementing the E-MDTs required a culture change in the way that professionals worked together across systems, requiring some education across these parties. In a similar fashion, TX/WellMed recognized the importance of having the support of all partners involved in the prevention intervention, regular communication, clear expectations and understanding of roles, and check points to ensure expectations are being met.

2. Implications for Research and Evaluation

Lessons learned from conducting the evaluation center on establishing the preconditions that would facilitate external evaluation and reduce burden on grantees. It would be helpful to include all data requirements in the Funding Opportunity Announcement (FOA), to the extent possible. While the FOA for this effort indicated that the collection of a set of core data elements would be required by all grantees, the evaluation team was tasked with identifying the specifics of those data elements during the first year of the grant. Understanding the data requirements prior to application may help ensure that grantees are aware of and can address any capacity issues that the requirements involve. Some grantees lacked the infrastructure to collect and track data, leading to delays in intervention start-up. Others began with a mix of paper-based data collection before transitioning into electronic records or databases. With the completion of this study, a refined set of data elements can be included in any future FOA to facilitate grantee planning. (See Appendix A for the data elements used.)

At the same time, an important task will be to reassess and reduce the number of core data elements for collection. Given the complexity of interventions and the multiple partners involved, data collection becomes a challenging task for not only grantees but participants themselves. Reducing the number of data elements to the most promising risk factors for further study would ease the burden of this component of the intervention. Identification of those essential measures can be facilitated by a close investigation of each individual intervention's outcomes (while each grantee collected data that are relevant to their intervention, only a subset of which is examined here). To determine the feasibility of conducting an external evaluation, future studies should require evaluability assessment as a standard procedure to assess the ability of grantees to provide needed data to address multiple domains of interest (e.g., APS involvement, health, well-being).

While certain aspects of data collection may warrant trimming, others are worth expanding. Data collection on a comparison or control group is critical to assess the associations between putative risk factors and elder abuse. In addition, requiring a common measure of risk of abuse for all grantees offers the opportunity to uniformly assess change in risk of abuse across interventions.

Cost-effectiveness studies would also be helpful to establish how addressing the risk of abuse, neglect, and exploitation for older adults proactively through prevention interventions may result in savings for social service and health care spending. A cost study was not conducted for this evaluation, although an original intent of the project was to set the stage for a future analysis of post-intervention health and well-being as well as health care utilization costs using administrative claims data from Medicaid and Medicare. While the FOA for the grantees' prevention intervention called for each project to be cost-effective and programmatically efficient, cost-effectiveness analyses were not required.19 According to Neta et al. (2015), "Information on the costs and resources required to deliver an intervention are essential" in implementation research, both to determine cost-effectiveness and return on investment and to inform decisions by policy-makers and program adopters about promising strategies to adopt in the field.