A Population-Based Study of Physical Function and Risk for Elder Abuse Reported to Social Service Agency

Abstract

We examined the association between physical function and the risk for reported elder abuse. In the Chicago Health and Aging Project (N = 8,932), 238 participants had reported elder abuse. The independent variable was objectively assessed physical function using both directly observed physical performance testing and self-reported physical function (Katz activity of daily living scale, Nagi physical activity scale, and Rosow Breslau mobility scales). Outcomes were elder abuse and specific subtypes of elder abuse. After adjusting for confounders, lower levels of physical performance testing (OR, 2.71[1.58-4.64]), Katz impairment (OR, 1.84[1.29-2.59]), Nagi impairment (OR, 1.65[1.15-2.37]) and Rosow Breslau (OR, 1.76[1.26-2.47]) were associated with increased risk for elder abuse. Lowest levels of physical performance testing were associated with increased risk for psychological abuse (OR, 2.69[1.27-5.71]), caregiver neglect (OR, 2.66[1.22-5.79]), and financial exploitation (OR, 2.35 [1.21-4.55]). Our results may have important implications to healthcare professional, social services and other disciplines to prevent and treat elder abuse.

Keywords

elder abuse physical function epidemiological study

Introduction

The World Health Organization (WHO) has declared that elder abuse is a violation of one of a human being’s most basic fundamental rights, to be safe and free of violence (World Health Organization, 2002). Elder abuse includes physical abuse, sexual abuse, psychological abuse, caregiver neglect, and financial abuse. Available data suggest that one out of 10 U.S. elderly persons experience abuse each year, and many of them experience it in multiple forms (Acierno et al., 2010; Beach, Schulz, Castle, & Rosen, 2010). In addition, recent data from U.S. Adult Protective Services Agencies depict an increasing trend in the reporting of elder abuse (Teaster et al., 2007). This trend is particularly alarming as the literature suggests that elder abuse is associated with increased risk of morbidity and mortality (Dong, 2005; Dong et al., 2009; Dong et al., 2011). U.S. National Research Council has urgently called for rigorous research on all aspects of elder abuse, especially through population-based epidemiological studies (National Research Council, 2003), as our current understanding of the risk factors for elder abuse in the general population remains limited.

This study followed the conceptual framework of sociocultural context suggested by the National Research Council (National Research Council, 2003). Elder abuse is defined as “intentional actions that cause harm or create a serious risk of harm, whether or not intended, to a vulnerable elder by a caregiver or other person who stands in a trust relationship to the elder; or failure by a caregiver to satisfy the elder’s basic needs or to protect the elder from harm.” Three major components need to be present to constitute elder abuse: an older person, a trust relationship, and vulnerability. Older person includes all persons over the age of 65 years. Trust relationship denotes a caregiving relationship or other familial relationship where a person has the responsibility of caring for or protecting the interests of older persons. Vulnerability, another core concept in elder abuse, refers to dependence on others or impaired ability for self-care or self-protection. This vulnerability refers to the cluster of risk factors (i.e., physical function impairment) associated with increased likelihood of elder abuse.

The sociocultural context model focuses on the comprehensive assessment of vulnerability factors, medical comorbidities, relationships, and socioeconomic status inequity while considering the sociocultural context and social embeddedness in which specific forms of elder abuse takes place. This overarching model highlights the importance of these interactions created by vulnerability, especially those with physical impairment. Social embeddedness refers to the set of people in the social well-being (social network and social participation) of the older adult and trusted others that may overlap. Vulnerability factors refer to physical health status (e.g., medical conditions, physical impairment, cognitive impairment), and psychological status (e.g., depression). Power and exchange dynamics refers to the negotiation of social needs for the older adults. The interactions among these components may vary for specific types of elder abuse and may potentially lead to the risk for specific subtypes of elder abuse, while at the same time guiding the list of covariates for the proposed analyses.

Physical function impairment has been associated with substantial morbidity and mortality (Johnson, 2000; Olsen & Jeune, 1980). Prior studies have found conflicting associations between levels of self-reported physical function and the risk for elder abuse (Comijs, Smit, Pot, Bouter, & Jonker, 1998; Compton, Flanagan, & Gregg, 1997; Fulmer et al., 2005; Lachs, Berkman, Fulmer, & Horwitz, 1994; Lachs, Williams, O’Brien, Hurst, & Horwitz, 1997; Paveza et al., 1992; Pillemer & Finkelhor, 1989). We are not aware of any epidemiological study that has used directly observed physical performance testing to examine its association with elder abuse. Moreover, we are not aware of any epidemiolgoical study that has examained the association between physical function and specific subtype of elder abuse and multiple forms of elder abuse.

In this manuscript, we build on the existing literature and the guiding conceptual framework to examine the association between the directly observed physical performance testing and self-reported physical function and the risk for elder abuse and its subtypes. Our underlying hypothesis is that a key vulnerability factor: physical function impairment as measured by the lower levels of physical performance testing, greater impairments in Katz, Nagi, and Rosow Breslau abilities are associated with increased risk for elder abuse reported to social services agency. In addition, given the complex interactions among different vulnerability factors as guided by the conceptual framework, we hypothesize that the degrees of association between different physical function measures vary across different subtypes of elder abuse in this community-dwelling population.

Method

Design and Participants

The study population consists of participants in the Chicago Health and Aging Project (CHAP), a population-based study of a geographically defined community. Details of the CHAP study design have been described previously (Bienias, Beckett, Bennett, Wilson, & Evans, 2003; Evans et al., 2003). Briefly, the study population enrolled residents aged 65 years and older in four adjacent neighborhoods on the south side of Chicago, after a complete census of the community area. Data collection includes an in-person interview conducted in participant’s homes. The interviews include standardized questionnaires and tests for the assessment of health history, physical function, health behaviors, and social factors. As of 2010, a total of 8,932 older adults had participated in the CHAP study. Written informed consent was obtained and the study was approved by the Institutional Review Board at Rush University Medical Center.

Elder Abuse Record Linkage

Reports of elder abuse to social services agencies can come from a variety of sources, including healthcare and legal professionals, community organizations, city workers (e.g., postal worker, utility worker, etc.), family members, or concerned neighbors or friends who have contact with seniors. In Illinois Adult Protective Services (APS), the definition used for abuse includes physical abuse, sexual abuse, psychological abuse, neglect, and financial exploitation. Definitions of elder abuse subtypes are included in the Table 4.

We matched data from CHAP participants to elder abuse cases reported to social services agencies from 1993 through 2010. Matching was based on an algorithm that compared the following information: date of birth, sex, race, exact home address, zip codes, and the home phone number. Matching was performed twice to increase accuracy. This resulted in 238 older CHAP participants who matched a social service agency record (126 were confirmed cases of elder abuse). If a CHAP participant was found to be reported more than once, we selected the first report.

Assessment of Physical Function

Physical function implies the assessment of specific activities and tasks, and impairment threatens one’s ability to live independently. Physical function was assessed by direct performance testing, which provided a comprehensive assessment of lower-extremity function, which consisted of measures of tandem stand, timed walk, tandem walk, and ability to rise to a standing position from a chair (Guralnik et al., 1994). Summary measures of these above tests were created as physical performance test scores (range 0-15). Lower score indicate impairment in these above activities and tasks which are often needed for independent living and may contribute toward physical disability. Because of their reliance on direct observation of actual performance, the tests possess high face validity, and have also been shown to have good internal consistency and reliability (Cronbach’s alpha = 0.76), and excellent test-retest reliability (0.88-0.92; Ostir, Volpato, Fried, Chaves, & Guralnik, 2002).

In addition, self-reported physical functions were assessed using three well-established measures using categorical answers. The first measure used was the Katz Index of Activities of Daily Living (Katz ADL) (range 0-6), which measures limitations in an individual’s ability to perform six basic self-care tasks (Katz & Akpom, 1976). Impairment in any of the ADL measures has great clinical relevance and could indicate the older person’s difficulties with basic activities of daily living and reliance on others to provide these assistances. Impairment in all ADL measure would suggest that the older person will require total care by another person unable to engage in any meaning physical activity.

The second measure was an index of mobility (range 0-3), based on work by Rosow and Breslau (1966). It is composed of three items measuring the ability to do heavy work around the house, walk up and down stairs, and walk half a mile. The third measure used in this study was an index of basic physical activities, and is based on work by Nagi (1976). It measures five activities of upper or lower-extremity function (range 0-5). Impairment in any of these mobility and physical activities measures could indicate the older adult’s difficulties with daily housework or out of the house activities. The threshold for physical disability from these functional impairments may vary; depend on the availability of external help and/or assistive devices to the older adults to perform these activities.

Study Variables

Elder abuse status was separated into three groups for the cohort (reported, confirmed, and no elder abuse). Reported elder abuse consists of all reported cases to social services agencies whether confirmed or unconfirmed. Confirmed elder abuse consists of subset of reported elder abuse cases based on the above indicators determined by social services agencies. Elder abuse subtypes include physical abuse, sexual abuse, psychological abuse, financial exploitation, and caregiver neglect (summary of neglect, willful deprivation, and confinement).

Demographic variables used in these analyses were age, sex, and race. In addition, we included socioeconomic status measures of education and income. Self-reported medical conditions of hypertension, diabetes mellitus, stroke, heart disease, hip fracture, and cancer were collected. Psychosocial factors included assessment of depressive symptoms, social network, and social engagement. Symptoms of depression were measured using a modified version (Kohout, Berkman, Evans, & Cornoni-Huntley, 1993) of the Center for Epidemiologic Studies of Depression Scale (CES-D; Radloff, 1977). Social network was summarized as the total number of children, relatives, and friends seen at least monthly (Cornoni-Huntley, Brock, Ostfeld, Taylor, & Wallace, 1986). Social engagement was assessed by asking how often older adults participate in social activities outside of house; such as religious activities, museums, library, and senior centers.

Analytic Approach

Descriptive univariate analyses were conducted by elder abuse status across age, sex, race, education, income, medical conditions, depressive symptoms, social network, and social participation. Corresponding Chi-squared, t-test and z-test were used to compare the bivariate differences between the elder abuse and no elder abuse groups across these above characteristics. Similar procedures were used to compare the differences in the physical performance test, Katz ADL, Nagi, and Rosow Breslau tests.

To examine the association between physical function and elder abuse, we used a series of logistic regression models to systematically consider potential confounding factors as guided by the conceptual framework. The core model (Model A) was adjusted for age, sex, and race. The next model (Model B) included additional socioeconomic variables including education and income. In Model C, we added to the previous model, the medical conditions of hypertension, heart disease, diabetes, stroke, cancer, and hip fracture. In Model D, we included depressive symptoms as a potential confounder in the relationship between physical function and elder abuse. Finally, Model E included additional measures of social well-being, namely social network and social participation. All of the above models were conducted for physical performance testing, Katz, Nagi, and Rosow Breslau abilities. As a sensitivity analysis, we repeated all above models for confirmed cases of elder abuse.

Moreover, all of the above models (Models A-E) were repeated using tertiles of physical performance testing (tandem, walk, and chair stands abilities) and self-reported physical function measures with respect to elder abuse outcomes. All above models were repeated for specific subtypes of elder abuse, namely physical abuse, psychological abuse, caregiver neglect, and financial exploitation as well as for participants who suffered multiple types of elder abuse. Due to the low number of sexual abuse cases, we did not have adequate power to examine these associations.

All models were validated graphically and analytically and model assumptions were adequately met. Odds ratio (OR), 95% confidence interval (CI), parameter estimate (PE), standard error (SE) and p values were reported for all analyses. Analyses were carried out using SAS^®, Version 9 (SAS Institute Inc., Cary, NC). All analyses used a 2-sided alternative with acceptable significance of P value less than 0.05. All analyses were performed using the PROC PHREG procedure in SAS® (SAS Institute Inc, 2004)].

Results

Baseline Characteristics

Of the 8,932 participants, 238 participants were reported for elder abuse. Overall, those with elder abuse were older and more women, and had lower levels of education, income, higher number of medical conditions, greater depressive symptoms, and lower levels of social network and social participation. The mean score for physical performance testing was 10.4 (3.8) for those without elder abuse and 6.5 (4.5) for those with elder abuse. For self-reported physical function, the differences between no-abuse versus abuse group were (Katz: 0.3 (1.1) versus 1.3 (2.0); Nagi (0.9(1.4) versus 2.1 (1.8); and Rosow (0.6 (0.9) versus 1.6 (1.3), respectively). The differences between specific physical function levels by elder abuse status are detailed in Table 1 .

Table 1.

Physical Function and Elder Abuse in the Study Population.

	EA N = 238	No EA N = 8,694	t-value	z-value	Chi-Squared	DF	P value
Physical performance test, mean, SD	6.5 (4.5)	10.4 (3.8)	14.87			8,456	<0.001
Tandem stands			12.4			8,882	<0.001
0-1	66 (28.3)	833 (9.6)
2-3	69 (29.6)	1,187 (13.7)
4-5	98 (42.1)	7631 (76.7)
Measured walk			11.3			8,911	<0.001
0-1	105 (44.1)	1,618 (18.7)
2-3	91 (38.2)	3,329 (38.4)
4-5	42 (17.6)	3,728 (42.9)
Chair stands			11.9			8,458	<0.001
0-1	111 (50.9)	1,502 (18.2)
2-3	70 (32.1)	3,080 (37.4)
4-5	37 (16.9)	3,660 (44.4)
Katz ADL impairment, mean, SD	1.3 (2.0)	0.3 (1.1)		13.29			<0.001
Walking	13 (5.2)	142 (1.6)			19.5	1	<0.001
Bathing	12 (5.0)	83 (0.9)			36.2	1	<0.001
Dressing	8 (3.3)	63 (0.7)			20.1	1	<0.001
Eating	6 (2.5)	48 (0.6)			14.7	1	<0.001
Transferring	8 (3.3)	70 (0.8)			17.2	1	<0.001
Toileting	11 (4.6)	77 (0.9)			32.6	1	<0.001
Nagi impairment, mean, SD	2.1 (1.8)	0.9 (1.4)		10.36			<0.001
Pulling or pushing large objects	135 (56.5)	2,136 (24.7)			123.9	1	<0.001
Stooping, crouching or kneeling	136 (56.9)	2,941 (33.9)			54.4	1	<0.001
Lifting or carrying 10 lbs weight	116 (48.3)	1,869 (21.6)			96.3	1	<0.001
Reaching or extending arms above shoulder	50 (20.8)	779 (8.9)			38.8	1	<0.001
Writing or handling or fingering small object	65 (27.1)	757 (8.7)			93.9	1	<0.001
Rosow Breslau impairment, mean, SD	1.6 (1.3)	0.6 (0.9)		12.96			<0.001
Heavy house work	152 (63.3)	2,612 (30.1)			120.2	1	<0.001
Walk up and down stairs	89 (37.1)	906 (10.4)			167.0	1	<0.001
Walk half a mile	133 (55.7)	1,762 (20.4)			171.6	1	<0.001

Note: ADL = activities of daily living. Tandem coding: 0 = unable to perform. 1 = ability to hold side-by-side stand < 10 s. 2 = ability to hold side-by side for 10 s, but unable to do semi-tandem. 3 = ability to hold side-by-side for 10 s and able to do semi-tandem for < 10 s. 4 = ability to hold semi-tandem for full 10 s. 5 = ability to hold full tandem for 10 s.

Measured walk 8 feet: 0 = Unable to perform. 1 = 6.1 s or more. 2 = 4.5 to 6.0 s; 3 = 3.7 to 4.4 s. 4 = 3.1 to 3.6 s. 5 = 1 to 3 s.

Chairs Stands: 0 = Unable to perform. 1 = Able to perform 1-4 times. 2 = Able to perform 5 times and took 18.6 s or more. 3 = Able to perform 5 times and took 15.3 s to 18.5 s. 4 = Able to perform 5 times and took 12.8 s to 15.3 s. 5 = Able to perform 5 times and took 1 s to 12.8 s.

Physical Function and Elder Abuse

In the fully-adjusted model (Table 2, Model E), every 1 point of lower levels in physical performance testing score is associated with increased risk for elder abuse (OR, 1.11, CI, 1.06-1.15). For Katz ADL, every 1 point of higher Katz impairment score was significantly associated with increased risk for elder abuse (OR, 1.12, CI, 1.02-1.22). Every 1 point higher in either Nagi (OR, 1.21, CI 1.09-1.33) or Rosow Breslau (OR, 1.43, CI 1.24-1.65) scales were also associated with increased risk for elder abuse. We repeated all above analyses for the confirmed cases of elder abuse, the results were similar (Table 3).

Table 2.

Association Between Physical Function and Reported Elder Abuse.

	A	B	C	D	E
Models	Odds ratio (95% confidence interval)for elder abuse outcome
Physical performance	1.18 (1.14-1.22)	1.19 (1.15-1.23)	1.17 (1.12-1.21)	1.13 (1.09-1.17)	1.11 (1.06-1.15)
Age	1.02 (1.00-1.04)	1.03 (1.01-1.05)	1.03 (1.01-1.05)	1.04 (1.01-1.06)	1.03 (1.01-1.06)
Men	0.62 (0.45-0.85)	0.65 (0.47-0.90)	0.65 (0.47-0.91)	0.72 (0.51-0.99)	0.66 (0.47-0.93)
African American	4.87 (3.23-7.34)	4.52 (2.93-6.97)	4.51 (2.92-6.96)	4.54 (2.88-7.15)	4.55 (2.88-7.17)
Education		1.02 (0.98-1.07)	1.02 (0.98-1.07)	1.03 (0.99-1.08)	1.04 (0.99-1.09)
Income		0.93 (0.86-0.99)	0.92 (0.86-0.99)	0.92 (0.85-0.99)	0.92 (0.85-0.99)
Medical conditions			1.27 (1.11-1.45)	1.25 (1.09-1.44)	1.25 (1.09-1.44)
CES-D				1.17 (1.10-1.24)	1.15 (1.08-1.22)
Social network					0.99 (0.96-1.02)
Social support					0.86 (0.78-0.96)
Katz ADL impairment	1.28 (1.19-1.37)	1.29 (1.21-1.39)	1.24 (1.15-1.34)	1.17 (1.07-1.27)	1.12 (1.02-1.22)
Age	1.05 (1.03-1.07)	1.05 (1.03-1.07)	1.05 (1.03-1.07)	1.05 (1.03-1.07)	1.05 (1.03-1.07)
Men	0.52 (0.39-0.71)	0.56 (0.41-0.77)	0.57 (0.41-0.78)	0.65 (0.47-0.89)	0.59 (0.43-0.83)
African American	5.74 (3.87-8.49)	4.84 (3.20-7.33)	4.81 (3.17-7.29)	4.92 (3.16-7.66)	4.92 (3.16-7.67)
Education		0.99 (0.96-1.04)	0.99 (0.96-1.04)	1.02 (0.97-1.06)	1.03 (0.99-1.08)
Income		0.91 (0.85-0.98)	0.91 (0.85-0.98)	0.91 (0.85-0.98)	0.92 (0.85-0.98)
Medical conditions			1.39 (1.23-1.59)	1.35 (1.18-1.54)	1.34 (1.17-1.53)
CES-D				1.17 (1.11-1.24)	1.14 (1.08-1.21)
Social network					0.99 (0.96-1.01)
Social support					0.83 (0.74-1.22)
Nagi impairment	1.39 (1.29-1.51)	1.41 (1.29-1.53)	1.33 (1.22-1.45)	1.25 (1.14-1.38)	1.21 (1.09-1.33)
Age	1.04 (1.03-1.06)	1.04 (1.03-1.06)	1.05 (1.03-1.07)	1.05 (1.03-1.07)	1.05 (1.03-1.07)
Men	0.58 (0.42-0.79)	0.61 (0.44-0.85)	0.61 (0.44-0.84)	0.68 (0.49-0.95)	0.63 (0.45-0.87)
African American	6.05 (4.09-8.95)	5.35 (3.54-8.09)	5.25 (3.47-7.95)	5.28 (3.39-8.22)	5.21 (3.34-8.11)
Education		1.00 (0.96-1.04)	1.00 (0.96-1.04)	1.02 (0.97-1.06)	1.03 (0.98-1.08)
Income		0.93 (0.87-0.99)	0.93 (0.86-0.99)	0.92 (0.86-0.99)	0.92 (0.86-0.99)
Medical conditions			1.32 (1.16-1.51)	1.29 (1.13-1.48)	1.28 (1.12-1.47)
CES-D				1.14 (1.08-1.21)	1.12 (1.06-1.19)
Social network					0.99 (0.96-1.01)
Social support					0.84 (0.76-0.93)
Rosow Breslau impairment	1.76 (1.56-1.98)	1.79 (1.59-2.02)	1.67 (1.47-1.89)	1.52 (1.32-1.74)	1.43 (1.24-1.65)
Age	1.03 (1.01-1.05)	1.03 (1.01-1.05)	1.04 (1.02-1.06)	1.04 (1.02-1.06)	1.04 (1.02-1.06)
Men	0.58 (0.43-0.79)	0.62 (0.45-0.85)	0.62 (0.45-0.85)	0.69 (0.50-0.96)	0.65 (0.46-0.90)
African American	5.66 (3.79-8.43)	5.15 (3.38-7.84)	5.12 (3.36-7.79)	5.22 (3.33-8.18)	5.19 (3.31-8.14)
Education		1.01 (0.97-1.05)	1.01 (0.97-1.05)	1.02 (0.98-1.07)	1.03 (0.99-1.08)
Income		0.93 (0.87-0.99)	0.93 (0.87-0.99)	0.92 (0.86-0.99)	0.93 (0.86-0.99)
Medical conditions			1.26 (1.10-1.44)	1.24 (1.08-1.42)	1.24 (1.08-1.42)
CES-D				1.14 (1.08-1.21)	1.12 (1.06-1.19)
Social network					0.99 (0.96-1.02)
Social support					0.86 (0.78-0.96)

Note: ADL = activities of daily living. CES-D = Center for Epidemiologic Studies of Depression Scale. Physical function measures were modeled as continuous variables.

Table 3.

Association Between Physical Function and Confirmed Elder Abuse.

	Odds ratio (95% CI)	Parameter estimate, standard error	P value
Physical performance testing
A	1.17 (1.11-1.22)	0.15, 0.02	<0.001
B	1.18 (1.12-1.23)	0.16, 0.02	<0.001
C	1.15 (1.09-1.21)	0.14, 0.03	<0.001
D	1.09 (1.04-1.16)	0.09, 0.03	<0.001
E	1.08 (1.02-1.14)	0.08, 0.03	<0.001
Katz impairment
A	1.24 (1.12-1.37)	0.21, 0.06	<0.001
B	1.26 (1.14-1.39)	0.23, 0.05	<0.001
C	1.19 (1.08-1.32)	0.18, 0.05	<0.001
D	1.07 (0.95-1.21)	0.07, 0.06	0.276
E	1.04 (0.91-1.18)	0.04, 0.06	0.586
Nagi impairment
A	1.39 (1.25-1.56)	0.33, 0.06	<0.001
B	1.42 (1.27-1.58)	0.35, 0.06	<0.001
C	1.32 (1.17-1.48)	0.28, 0.06	<0.001
D	1.21 (1.07-1.38)	0.19, 0.07	<0.001
E	1.18 (1.03-1.35)	0.17, 0.07	0.015
Rosow Breslau impairment
A	1.72 (1.47-2.03)	0.54, 0.08	<0.001
B	1.78 (1.51-2.10)	0.58, 0.08	<0.001
C	1.62 (1.36-1.92)	0.48, 0.09	<0.001
D	1.41 (1.16-1.70)	0.34, 0.09	<0.001
E	1.34 (1.10-1.64)	0.29, 0.10	0.004

Note: Model A = age, sex, race. Model B = A + education, income. Model C = B + medical conditions. Model D = C + depressive symptoms. Model E = D + social network, social participation.

Physical Function With Specific Types of Elder Abuse

In the fully-adjusted model (Table 4 Model E), every 1 point lower in physical performance testing score was associated with increased risk for physical abuse (OR, 1.12, CI, 1.02-1.23), psychological abuse (OR, 1.12, CI, 1.06-1.19), caregiver neglect (OR, 1.17, CI, 1.11-1.24), and financial exploitation (OR, 1,08, CI, 1.03-1.14). For Katz, every 1 point higher in Katz impairments was associated with increased risk for psychological abuse and caregiver neglect, but not for physical abuse or financial exploitation. For Nagi, every 1 point higher in Nagi impairment was associated with increased risk for all subtypes of elder abuse. For Rosow Breslau, every 1 point higher in Rosow Breslau impairment was associated with increased risk for psychological abuse, caregiver neglect, and financial exploitation ( Table 4 ).

Table 4.

Association Between Physical Function and Specific Subtypes of Elder Abuse.

	Physical abuse		Psychological abuse		Caregiver neglect		Financial exploitation
	OR (CI)	PE, SE	OR (CI)	PE, SE	OR (CI)	PE, SE	OR (CI)	PE, SE
Physical performance
A	1.16 (1.07-1.26)	0.15, 0.04***	1.18 (1.12-1.24)	0.16, 0.03***	1.25 (1.19-1.31)	0.22, 0.02***	1.14 (1.09-1.20)	0.13, 0.02***
B	1.18 (1.08-1.28)	0.16, 0.04***	1.17 (1.11-1.23)	0.16, 0.03***	1.26 (1.20-1.32)	0.23, 0.02***	1.15 (1.10-1.20)	0.14, 0.02***
C	1.19 (1.09-1.29)	0.17, 0.04***	1.15 (1.09-1.22)	0.14, 0.03***	1.23 (1.17-1.30)	0.21, 0.03***	1.12 (1.07-1.18)	0.12, 0.02***
D	1.16 (1.06-1.27)	0.15, 0.05**	1.13 (1.07-1.20)	0.12, 0.03***	1.19 (1.13-1.26)	0.18, 0.03***	1.10 (1.15-1.20)	0.09, 0.03***
E	1.12 (1.02-1.23)	0.12, 0.05*	1.12 (1.06-1.19)	0.12, 0.03***	1.17 (1.11-1.24)	0.16, 0.03***	1.08 (1.03-1.14)	0.08, 0.03**
Katz
A	1.16 (0.96-1.41)	0.15, 0.09	1.28 (1.15-1.42)	0.25, 0.05***	1.41 (1.28-1.54)	0.34, 0.05***	1.19 (1.08-1.32)	0.18, 0.05***
B	1.19 (0.99-1.45)	0.18, 0.09	1.27 (1.13-1.41)	0.24, 0.06***	1.43 (1.30-1.57)	0.36, 0.05***	1.21 (1.09-1.34)	0.19, 0.05***
C	1.17 (0.96-1.43)	0.16, 0.10	1.22 (1.09-1.36)	0.19, 0.06***	1.36 (1.23-1.49)	0.31, 0.05***	1.16 (1.05-1.29)	0.15, 0.05**
D	1.14 (0.93-1.41)	0.13, 0.11	1.20 (1.06-1.36)	0.19, 0.06**	1.27 (1.14-1.42)	0.24, 0.06***	1.14 (1.02-1.28)	0.13, 0.06*
E	1.07 (0.87-1.32)	0.07, 0.11	1.18 (1.04-1.33)	0.16, 0.06**	1.21 (1.08-1.36)	0.19, 0.06**	1.09 (0.95-1.23)	0.09, 0.06
Nagi
A	1.36 (1.12-1.65)	0.30, 0.09**	1.45 (1.29-1.64)	0.37, 0.07***	1.61 (1.44-1.81)	0.48, 0.06***	1.31 (1.18-1.46)	0.27, 0.05***
B	1.40 (1.15-1.71)	0.34, 0.10***	1.42 (1.26-1.61)	0.35, 0.06***	1.64 (1.46-1.84)	0.49, 0.06***	1.32 (1.18-1.47)	0.28, 0.06***
C	1.39 (1.13-1.72)	0.33, 0.11**	1.36 (1.19-1.54)	0.31, 0.07***	1.54 (1.36-1.74)	0.43, 0.06***	1.25 (1.12-1.41)	0.23, 0.06***
D	1.34 (1.07-1.68)	0.29, 0.11**	1.33 (1.16-1.53)	0.29, 0.07***	1.47 (1.29-1.68)	0.39, 0.07***	1.22 (1.08-1.38)	0.19, 0.06**
E	1.26 (1.00-1.58)	0.23, 0.12*	1.32 (1.14-1.52)	0.28, 0.07***	1.42 (1.24-1.62)	0.35, 0.07***	1.18 (1.04-1.34)	0.17, 0.06**
Rosow Breslau
A	1.58 (1.19-2.09)	0.46, 0.14**	1.77 (1.49-2.11)	0.57, 0.09***	2.10 (1.77-2.48)	0.74, 0.09***	1.61 (1.38-1.88)	0.47, 0.08***
B	1.66 (1.24-2.21)	0.51, 0.15***	1.74 (1.46, 2.08)	0.55, 0.09***	2.16 (1.82-2.56)	0.77, 0.09***	1.64 (1.39-1.92)	0.49, 0.08***
C	1.65 (1.22-2.23)	0.50, 0.15**	1.64 (1.36-1.98)	0.49, 0.09***	1.99 (1.66-2.39)	0.69, 0.09***	1.54 (1.30-1.82)	0.43, 0.09)***
D	1.53 (1.11-2.11)	0.43, 0.16**	1.57 (1.29-1.92)	0.45, 0.10***	1.80 (1.48-2.19)	0.59, 0.10***	1.47 (1.23-1.75)	0.38, 0.09***
E	1.35 (0.97-1.88)	0.29, 0.17	1.54 (1.25-1.89)	0.43, 0.11***	1.69 (1.38-2.07)	0.52, 0.10***	1.39 (1.16-1.68)	0.33, 0.09***

Note: PE = parameter estimates. SE =: standard errors.

Model A = age, sex, race. Model B = A + education, income. Model C = B + medical conditions. Model D = C + depressive symptoms. Model E = D + social network, social participation.

Table 4 demonstrates the association between physical function and the risk for specific subtypes of elder abuse. Physical function measures are modeled as a continuous variable and elder abuse is modeled as a categorical variable. Models A-E demonstrates the specific association between physical function measures and elder abuse, taking into consideration of the wide range of potential confounding factors.

Definition of the Subtypes of Elder Abuse:

Physical abuse is defined as inflicting physical pain or injury upon an older adult.

Sexual abuse is touching, fondling, intercourse, or any other sexual activity with an older adult, when the older adult is unable to understand, unwilling to consent, threatened or physically forced.

Psychological abuse involves verbal assaults, threat of abuse, harassment or intimidation.

Neglect is a caregiver’s failure to provide an older adult with life’s necessities, including food, clothing, shelter or medical care.

Financial exploitation includes the misuse, or withholding of an older adult’s resources by another, to the disadvantage of the elderly person or the profit or advantage of someone else.

p < .05. **p < .01. ***p < .001

Lastly, we repeated similar logistic regression models to examine the association between levels of physical function and multiple types of elder abuse. In the fully-adjusted model ( Table 5 ), the lowest tertiles in the 8-feet walking speed were associated with increased risk for three or more forms of abuse (OR, 3.30, CI, 1.20-9.08). For Nagi, the highest levels of impairment are associated with increased risk for three or more forms of abuse (OR, 2.80, CI, 1.18-6.65). For Rosow Breslau, the highest levels of impairment are associated with increased risk for three or more forms of abuse (OR, 2.28, CI, 1.02-5.09).

Table 5.

Association Between Levels of Physical Function and Specific Subtypes of Elder Abuse.

		Specific subtypes of elder abuse				Number of multiple types of elder abuse
	All cases of elder abuse	Physical abuse	Psychological abuse	Caregiver neglect	Financial exploitation	1 type	2 types	3 or more
	Odds ratio (95% confidence interval)
Physical performance
High	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0
Middle	1.38 (0.79-2.38)	0.96 (0.31-2.99)	1.22 (0.56-2.65)	1.05 (0.47-2.38)	1.45 (0.75-2.79)	1.27 (0.55-2.96)	1.99 (0.79-4.99)	0.87 (0.27-2.82)
Low	2.71 (1.58-4.64)	2.32 (0.77-7.08)	2.69 (1.27-5.71)	2.66 (1.22-5.79)	2.35 (1.21-4.55)	2.67 (1.18-6.05)	3.16 (1.25-7.99)	2.32 (0.76-7.10)
Tandem
High	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0
Middle	2.24 (1.58-3.19)	3.17 (1.43-7.02)	1.91 (1.13-3.24)	2.55 (1.51-4.32)	2.19 (1.41-3.41)	2.54 (1.53-4.21)	2.31 (1.30-4.09)	1.63 (0.67-3.98)
Low	2.18 (1.44-3.29)	2.49 (0.93-6.72)	2.17 (1.19-3.97)	3.69 (2.11-6.47)	1.62 (0.93-2.83)	1.89 (1.02-3.49)	2.69 (1.41-5.18)	2.16 (0.82-5.65)
Walking
High	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0
Middle	1.38 (0.93-2.05)	1.68 (0.68-4.15)	1.17 (0.66-2.07)	1.74 (0.91-3.34)	1.22 (0.75-1.97)	1.34 (0.74-2.42)	1.49 (0.79-2.78)	1.19 (0.45-3.16)
Low	1.87 (1.21-2.91)	1.74 (0.60-5.00)	2.01 (1.08-3.74)	3.63 (1.83-7.19)	1.39 (0.80-2.41)	1.72 (0.89-3.31)	1.61 (0.79-3.27)	3.30 (1.20-9.08)
Chair stand
High	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0
Middle	1.29 (0.85-1.95)	1.88 (0.71-4.99)	1.01 (0.56-1.84)	1.56 (0.80-3.05)	1.25 (0.75-2.09)	1.11 (0.59-2.07)	1.69 (0.85-3.39)	1.02 (0.39-2.62)
Low	2.29 (1.47-3.59)	2.89 (0.97-8.55)	2.31 (1.24-4.31)	3.41 (1.70-6.83)	2.05 (1.17-3.58)	1.98 (1.03-3.81)	3.03 (1.43-6.38)	2.04 (0.74-5.64)
ADL
No	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0
Yes	1.84 (1.29-2.59)	1.63 (0.71-3.74)	2.53 (1.52-4.19)	2.62 (1.61-4.27)	1.60 (1.01-2.53)	1.30 (0.78-2.17)	2.54 (1.46-4.41)	2.28 (0.99-5.30)
Nagi
No	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0
Yes	1.65 (1.15-2.37)	2.75 (1.17-6.46)	2.44 (1.42-4.20)	2.47 (1.46-4.19)	1.45 (0.91-2.29)	1.02 (0.59-1.73)	2.45 (1.35-4.44)	2.80 (1.18-6.65)
Rosow Breslau
No	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0
Yes	1.76 (1.26-2.47)	1.55 (0.72-3.35)	2.24 (1.36-3.68)	2.01 (1.22-3.31)	1.93 (1.26-2.98)	1.61 (0.98-2.65)	1.88 (1.09-3.23)	2.28 (1.02-5.09)

Note: ADL = activities of daily living.

Models Adjusted for: age, sex, race, education, income, medical conditions, depressive symptoms, social network, and social participation

1.0 in each of the categories designate as the reference comparison group.

Table 5 demonstrates the association between levels of physical function and the risk for specific subtypes of elder abuse as well as multiple forms of elder abuse. Physical function measures are modeled as either tertiles or categorical variables and elder abuse is modeled as categorical variables. Models A-E demonstrates the specific association between levels of physical function measures and elder abuse, taking into consideration of the wide range of potential confounding factors.

Discussion

In this population-based study of 8,698 older people, we found that lower levels of physical performance testing score as well as greater impairments in Katz, Nagi, and Rosow Breslau scores were associated with increased risk for elder abuse. In addition, we found that strength of association between the physical function measures above and elder abuse varies by the specific subtypes of elder abuse. Moreover, lower ability in walking speed, greater impairment in Nagi and Rosow Breslau was associated with increased risk of suffering multiple forms of elder abuse.

In our guiding conceptual framework of sociocultural context, physical function impairment was a key vulnerability factors that may predispose an elderly to be abused. Prior research suggests that there is an incomplete and often controversial evidence linking impairment in physical function to elder abuse. Fulmer et al. (2005) found that lower levels of self-reported functional status were associated with increased risk of elder abuse. Prior systematic review (Cooper, Selwood, & Livingston, 2008) found that 25% of dependent older adults reported significant levels of psychological abuse, 1% reported physical abuse, and 20% of older adults present to the emergency department had experienced caregiver neglect. Wang (2006) used the Psychological Elder Abuse Scale and found dependent older adults experience 22.6% of the psychological abuse. Beach et al assessed disabled older adult and their care givers and found that greater care recipient ADL and IADL needs was associated with significant potential harmful caregiver behaviors (OR, 1.12 (1.03-1.22) (Beach et al., 2005). However, other studies have found that physical dependence was not associated with increased risk of elder abuse (Comijs et al., 1998; Dong & Simon, 2010; Pillemer & Finkelhor, 1989, 1989). We are aware of one population-based study (Lachs et al., 1994), that found impairment in eating ability (OR, 3.5, 95% CI, 1.2-11.7) and greater impairment in ADL (OR, 1.3, 95% CI 1.1-1.8) was associated with increased risk of elder abuse (Lachs et al., 1997). Drawn from the sociocultural conceptual framework, our study examined a wide range of potential factors to explore the relationship between physical function and elder abuse in this community-dwelling population.

Our findings build on the results of other studies of physical function and elder abuse in a number of different ways. First, our study is the largest population-based study to examine the association between physical function and elder abuse, in a well characterized community for more than 17 years, which contribute to the generalizability of our study findings. Second, the present study further expands our existing knowledge about the associations of both directly observed physical performance testing as well as self-reported physical function measure and elder abuse, particularly among those with lowest levels of physical function.

Third, our study is also the first population-based study to examine the associations between physical function with the specific subtypes of elder abuse. Evidence suggests that different types of elder abuse may have different predisposing and precipitating risk factors (Garre-Olmo et al., 2009). Guided by our conceptual framework, interactions among each of the vulnerability factors may vary at the differential level according to specific types of elder abuse. In term, these interactions may increase the risk for certain types of abuse, while not for other types of abuse. Our study findings demonstrate that degree of association varies between physical function and different subtypes of elder abuse. This finding may have direct implications for future prevention and intervention strategies targeted to specific subset of the targeted populations. Lastly, our study found that those with lowest ability to walk 8-foot distance, and greatest impairment in Nagi and Rosow Breslau scales had substantially higher risk for suffering three or more forms of elder abuse. Educational initiatives should be developed to inform these professionals that more comprehensive assessment of physical function measures might be needed to identify a particularly higher risk group suffering multiple forms of abuse.

The temporal relations between physical function and elder abuse need further investigation. We considered a series of sociodemographic, socioeconomic characteristic, medical comorbidities, psychological, and social well-being. However, adjustments for these factors did not significantly influence the relationship between physical function and elder abuse. Although it is conceivable that physical decline could increase dependency and increase the risk for elder abuse, at the same time, it is also possible that abusive acts could induce physical injuries that could worsen physical function impairment. In addition, metabolic abnormalities, nutritional deficiencies, infections, injuries, or trauma may be other factors that account for the association between physical function and elder abuse, but these factors were not considered in this analysis. Severity of medical comorbidities could be another important factor in determining the causal mechanisms between physical function and elder abuse. It is conceivable that elder abuse could exacerbate the existing medical conditions, which could predispose greater impairment in physical function. However, we do not have measures in our existing data to further elucidate these relations.

Our study also has a number of limitations. First, elder abuse was not ascertained uniformly for all members of the CHAP population, but only for participants reported or substantiated to the social services agency because someone suspected problems. Elder abuse is underreported and in our study, there is likely underestimation of the potential effect of missing elder abuse case with respect to the relationship between physical function and elder abuse. Future study is needed to uniformly collect elder abuse measures in representative populations to rigorously examine these associations.

Second, this cross-sectional study cannot assess the temporal relation of physical function impairment and elder abuse. It is possible that physical function impairment could lead to elder abuse. The reverse causation could also be true in that higher levels of impairment are associated with greater levels of reporting. In addition, elder abuse cases in our study are not incident cases, which could create significant biases with significant methodological challenges in potentially conducting prospective analyses with respect to these cases. Third, in addition to the potential changes in physical function over time, some older adults transition in and out of physical disability states, which CHAP study will likely miss over the 3-year cycle. However, we did not have these measures to determine the extent of influence of these factors. Further work is needed to clarify this issue in representative populations.

This study has important practical and policy implications. For health care professionals who care for older adults, screening for elder abuse should occur alongside the screening for physical function impairment. In addition, particular attention for possible elder abuse should be paid to those with low levels of physical function, which may have substantial increased risk to be abused. Moreover, health care professionals should consider routinely screening for elder abuse among those with high levels of physical disability. Attention to these relationships could serve as important markers to preventive strategies for elder abuse. For social services agencies, most programs require that physical function screening determine the degree of vulnerability and unmet needs. “Red-flags” for potential elder abuse should be raised for vulnerable, functional impaired individuals, especially those without the support needed to meet their most basic day-to-day needs. Furthermore, health care professionals should integrate physical performance testing in addition to the routine assessment of ADL, which will likely capture at-risk population, particularly those who may suffer multiple forms of elder abuse. Improved understanding of these relationships could be critical for future prevention and intervention strategies to combat the issues of elder abuse.

Conclusion

We conclude that lower levels of physical performance testing and greater impairment in Katz, Nagi, and Rosow Breslau are associated with increased risk for elder abuse. The increased risks for elder abuse are particularly high for those with lowest levels of physical function. Moreover, the degree of association between physical function and elder abuse varies by the subtypes of elder abuse. Furthermore, those with greater impairment in an 8-feet walk, Nagi, and Rosow Breslau ability are at substantial higher risk for multiple forms of elder abuse. However, the potential causal association between physical function and elder abuse requires further longitudinal investigation. Future studies will also be necessary to determine the temporal relations between comprehensive assessment of physical function and elder abuse, and to explore the racial and ethnic differences in the levels of physical function with elder abuse in representative populations.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

XinQi Dong is the Director of Chinese Health, Aging and Policy Program and Associate Director of Rush Institute for Healthy Aging at Rush University Medical Center. His work involves the population-based study of elder abuse in diverse communities.

Melissa Simon is the Vice Chairman of Clinical and Policy Research and Associate Professor OB/GYN at Northwestern University. She is a women’s health expert and has a long history of working with disenfranchised and vulnerable populations.

Denis Evans is the Director of Rush Institute on Aging and Professor of Medicine at Rush University Medical Center. His work involves the longitudinal epidemiological study of common chronic disease in older adults and principle investigator of the Chicago Health and Aging Project.

References

Acierno

Hernandez

M. A.

Amstadter

A. B.

Resnick

H. S.

Steve

Muzzy

Kilpatrick

D. G.

(2010). Prevalence and correlates of emotional, physical, sexual, and financial abuse and potential neglect in the United States: The National Elder Mistreatment Study. American Journal of Public Health, 100, 292-297.

Beach

S. R.

Schulz

Castle

N. G.

Rosen

(2010). Financial exploitation and psychological mistreatment among older adults: Differences between African Americans and Non-African Americans in a population-based survey. Gerontologist, 50, 744-757.

Beach

S. R.

Schulz

Williamson

G. M.

Miller

L. S.

Weiner

M. F.

Lance

C. E.

(2005). Risk factors for potentially harmful informal caregiver behavior. Journal of .American. Geriatrics Society, 53, 255-261.

Bienias

J. L.

Beckett

L. A.

Bennett

D. A.

Wilson

R. S.

Evans

D. A.

(2003). Design of the Chicago Health and Aging Project (CHAP). Journal of Alzheimer’s Disease, 5, 349-355.

Comijs

H. C.

Smit

J. H.

Pot

A. M.

Bouter

L. M.

Jonker

(1998). Risk indicator of elder mistreatment in the community. Journal of Elder Abuse and Neglect, 9, 67-76.

Compton

S. A.

Flanagan

Gregg

(1997). Elder abuse in people with dementia in Northern Ireland: Prevalence and predictors in cases referred to a psychiatry of old age service. International Journal of Geriatric Psychiatry, 12, 632-635.

Cooper

Selwood

Livingston

(2008). The prevalence of elder abuse and neglect: A systematic review. Age and Ageing, 37, 151-160.

Cornoni-Huntley

Brock

D. B.

Ostfeld

Taylor

J. O.

Wallace

R. B.

(1986). Established populations for epidemiological studies of the Elderly Resource Data Book (Rep. No. NIH Publication No. 86-2443). Washington, DC: US Department of Health and Human Services.

Dong

(2005). Medical implications of elder abuse and neglect. Clinics in Geriatric Medicine, 21, 293-313.

10.

Dong

Simon

M. A.

(2010). Is impairment in physical function associated with increased risk of elder mistreatment? Public Health Reports, 125, 743-753.

11.

Dong

Simon

M. A.

Beck

Farran

McCann

Mendes de Leon

C. F.

. . . Evans

D. A.

(2011). Elder abuse and mortality: The role of psychological and social wellbeing. Gerontology, 57, 549-558.

12.

Dong

Simon

M. A.

Mendes de Leon

Fulmer

Beck

Hebert

. . .Evans

(2009). Elder self-neglect and abuse and mortality risk in a community-dwelling population. Journal of the American Medical Association, 302, 517-526.

13.

Evans

D. A.

Bennett

D. A.

Wilson

R. S.

Bienias

J. L.

Morris

M. C.

Scherr

P. A.

. . . Schneider

(2003). Incidence of Alzheimer disease in a biracial urban community: Relation to apolipoprotein E allele status. Archives of Neurology, 60, 185-189.

14.

Fulmer

Paveza

VandeWeerd

Fairchild

Guadagno

Bolton-Blatt

Norman

(2005). Dyadic vulnerability and risk profiling for elder neglect. Gerontologist, 45, 525-534.

15.

Garre-Olmo

Planas-Pujol

Lopez-Pousa

Juvinya

Vila

Vilalta-Franch

(2009). Prevalence and risk factors of suspected elder abuse subtypes in people aged 75 and older. Journal of American Geriatrics Society, 57, 815-822.

16.

Guralnik

J. M.

Simonsick

E. M.

Ferrucci

Glynn

R. J.

Berkman

L. F.

Blazer

D. G.

. . . Wallace

R. B.

(1994). A short physical performance battery assessing lower extremity function: association with self-reported disability and prediction of mortality and nursing home admission. Journal of Gerontology, 49, M85-M94.

17.

Johnson

N. E.

(2000). The racial crossover in comorbidity, disability, and mortality. Demography, 37, 267-283.

18.

Katz

Akpom

C. A.

(1976). A measure of primary sociobiological functions. International Journal of Health Services, 6, 493-508.

19.

Kohout

F. J.

Berkman

L. F.

Evans

D. A.

Cornoni-Huntley

(1993). Two shorter forms of the CES-D (Center for Epidemiological Studies Depression) depression symptoms index. Journal of Aging Health, 5, 179-193.

20.

Lachs

M. S.

Berkman

Fulmer

Horwitz

R. I.

(1994). A prospective community-based pilot study of risk factors for the investigation of elder mistreatment.Journal of the American Geriatrics Society, 42, 169-173.

21.

Lachs

M. S.

Williams

O’Brien

Hurst

Horwitz

(1997). Risk factors for reported elder abuse and neglect: A nine-year observational cohort study.Gerontologist, 37, 469-474.

22.

Nagi

S. Z.

(1976). An epidemiology of disability among adults in the United States. Milbank Memorial Fund Quarterly.Health Society, 54, 439-467.

23.

National Research Council (2003). Elder Mistreatment: Abuse, neglect and exploitation in an Aging America. Washington, DC: The National Academies Press.

24.

Olsen

Jeune

(1980). The mortality experience of early old-age and disability pensioners from unskilled- and semiskilled labour groups in Fredericia. Scandinavian Journal of Social Medicine Supplementum, 16, 50-52.

25.

Ostir

G. V.

Volpato

Fried

L. P.

Chaves

Guralnik

J. M.

(2002). Reliability and sensitivity to change assessed for a summary measure of lower body function: Results from the Women’s Health and Aging Study. Journal of Clinical Epidemiology, 55, 916-921.

26.

Paveza

G. J.

Cohen

Eisdorfer

Freels

Semla

Ashford

J. W.

. . . Levy

(1992). Severe family violence and Alzheimer’s disease: Prevalence and risk factors. Gerontologist, 32, 493-497.

27.

Pillemer

Finkelhor

(1989). Causes of elder abuse: Caregiver stress versus problem relatives. American Journal of Orthopsychiatry, 59, 179-187.

28.

Radloff

(1977). The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measure, 1, 385-401.

29.

Rosow

Breslau

(1966). A Guttman health scale for the aged. Journals of Gerontology, 21, 556-559.

30.

SAS Institute Inc (2004). SAS OnlineDoc, Version 9.1.3. Cary, NC: Author.

31.

Teaster

P. B.

Dugar

Mendiondo

Abner

E. L.

Cecil

K. A.

Otto

J. M.

(2007). The 2004 Survey of Adult Protective Services: Abuse of adults 60 years of age and older. Retrieved from www.elderabusecenter.org/pdf/research/apsreport030703.pdf [On-line].

32.

Wang

J. J.

(2006). Psychological abuse and its characteristic correlates among elderly Taiwanese. Archives of Gerontology and Geriatrics, 42, 307-318.

33.

World Health Organization (2002). World report on violence and health. http://www.who.int/violence_injury_prevention/violence/world_report/en/summary_en.pdf [On-line].