| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Address correspondence to Lawrence Schonfeld, PhD, Department of Aging & Mental Health, Louis de la Parte Florida Mental Health Institute, University of South Florida, 13301 Bruce B. Downs Blvd., Tampa, FL 33612. E-mail: schonfel{at}fmhi.usf.edu
 |
Abstract |
|---|
 TOP Abstract MethodDiscussionReferences |
|---|
Key Words: Elder abuse • Mental health • Substance abuse • Medicaid • Medicare
According to the Administration on Aging (2005), elder abuse is an umbrella term that describes physical abuse, sexual abuse, emotional or psychological abuse, financial or material exploitation, neglect by a caretaker, or self-neglect by the older adult. Much of the published research has focused on the frequency of these categories of abuse and the demographics of the victims, whether relying on national-level data (National Center on Elder Abuse [NCEA], 1998) or state-level data (Neale, Hwalek, Goodrich, & Quinn, 1996; Pavlik, Hyman, Festa, & Bitondo-Dyer, 2001; Reynolds & Schonfeld, 2004). In the National Elder Abuse Incidence Study (NEAIS), trained "sentinels," consisting of providers of more elder-specific services and contacts, more often identified verified cases of abuse than did the traditional adult protective services (APS) system (NCEA, 1998). Despite the value of including trained sentinels to assist in the identification process, most of the identification and verification falls on APS investigators who rarely assess for mental health or substance-use problems, nor do they provide direct mental health services (Teaster, Nerenberg, Stanis, & Stansbury, n.d.).
Mental health or substance-use disorders are often unrecognized and untreated in abused older adults (Kessler, Costello, Merikangas, & Ustun, 2001; Marshall, Benton, & Brazier, 2000; NCEA, 1998). Only a few studies have provided some evidence that mental disorders are more frequent among abused elders in comparison to individuals not abused. In a study conducted in the Netherlands, Comijs, Penninx, Knipscheer, and van Tilburg (1999) found significantly higher levels of distress among 77 cases of abuse in comparison to 147 nonabused elders. Dyer, Pavlik, Murphy, and Hyman (2000) compared 47 elders referred for treatment at a medical clinic due to abuse or neglect with a random sample of 97 elderly patients referred for other reasons. Prevalence of depression among victims of self-neglect was 62% compared to only 12% in a comparison group. Wolf (1999–2000) noted that while such studies demonstrate more depression among abused in comparison to nonabused older adults, other forms of emotional distress such as fear, shame, guilt, alienation, and posttraumatic stress have not been investigated.
It is not surprising, therefore, that the 2001 National Policy Summit on Elder Abuse identified mental health issues as the major service gap in assisting victims. The panel recommended that "age-appropriate specialized mental health services need to be available and accessible to include aggressive outreach, intensive case management and specialized clinicians to provide acute and ongoing services for victims" (NCEA, 2001).
There are a number of barriers to implementing such recommendations. As noted in the U.S. Surgeon General's report on mental health, older adults in general do not utilize behavioral health services for reasons such as stigma toward mental illness, ageism, and a lack of awareness of the problem by the elders themselves or their family members (U.S. Department of Health and Human Services, 1999). A lack of mental health services to victims of abuse may be further compounded because few cases of suspected mistreatment are actually confirmed following the APS investigation process, resulting in no further screening or follow-up. Furthermore, self-neglect, rather than abuse by a perpetrator, is often among the most frequent reasons for referrals made to the APS system (NCEA, 1998; Reynolds & Schonfeld, 2004), suggesting that this population's signs of deterioration are rarely observed by others and that they rarely seek social services, including behavioral health. Finally, as noted in a survey of APS providers, assessment protocols for APS rarely include formal screening of mental health or substance abuse problems (Teaster et al., n.d.).
The impetus for the present study began with a request by the then Secretary of Florida's Department of Elder Affairs (DOEA) to the first author to explore the relationship between elder abuse and behavioral issues of depression and substance abuse, as well as the extent to which victims of abuse were screened and treated for behavioral health problems. Given the lack of mental health assessment data within the protective-services system, we used alternative sources of data to investigate these issues. The purpose of the present study was to describe the frequency and costs of behavioral health services when comparing suspected cases of abuse and/or self-neglect with other older service users. We accomplished this through examination of Florida Medicaid and Medicare service claims data.
|
Method |
|---|
|
TOPAbstractMethodDiscussionReferences |
|---|
The study was approved by the University's Institutional Review Board. The Center for Medicaid and Medicare Services (CMS), the Agency for Health Care Administration (AHCA) in Florida, and the Florida Department of Children and Families (DCF) each approved access to, and analyses of, their respective agency's data through formal data-use agreements with the investigators. Data were stored and analyzed on a secure server at the Florida Mental Health Institute at the University of South Florida in Tampa.
Analyses involved comparison of older adults identified through a statewide abuse hotline database who also used health services with older adult service users not identified in the hotline database. Outcomes included proportions of behavioral health service users by measuring relative risk (risk ratios), comparison of actual costs between the two groups (independent mean comparisons), and comparison of penetration rates, that is, proportion of eligible cases demonstrating behavioral health service use. The datasets are described as follows.
Florida Abuse Hotline Information System (FAHIS)
The Florida DCF Adult Service Unit provided the FAHIS data. That unit originally provided data for the fiscal year July 1998 through June 1999. However, for the present study, we included only the first 6 months of FAHIS data for 1999 because only Medicare and Medicaid service claims data for that calendar year were available. The data variables provided by DCF were limited to clients' Social Security numbers, category of suspected abuse, and classification of the outcome of the investigation. DCF did not provide categories such as victims' place of residence, gender, race, location of the abuse, and any perpetrator data. The primary purpose for using the FAHIS data was to gain access to the Social Security numbers in order to crosslink the data to health services claims.
Reports to the FAHIS are derived from either written or toll-free telephone complaints from any source and for any suspected case of four classifications: abuse, neglect, exploitation, or self-neglect. Designated staff members or counselors screen allegations to determine whether the cases meet the criteria for an abuse report and provide technical assistance to local or district staff during the investigation process. Investigations are carried out at the local level by APS and reported to DOEA (Florida Legislature, Office of Program Policy Analysis and Government Accountability, 2001).
Medicare Data
Medicare data claims included 1999 beneficiary enrollment and demographic ("denominator") data and claims data obtained under data-use agreements with the Centers for Medicare and Medicaid Services (CMS). These claims data files or CMS's Standard Analytic Files (SAF) included final action, fee-for-service claims submitted to CMS for services provided to beneficiaries in the 1999 calendar year, and contained all in-state and out-of-state health care claims for all beneficiaries who were residents of Florida. Using a "finder file" provided by CMS, we obtained all persons' Social Security numbers and used them to match with the FAHIS data.
Medicaid Data
This dataset consisted of Medicaid enrollment and claims data also for calendar year 1999 as provided by the Agency for Health Care Administration, the Florida agency overseeing Medicaid and health quality assurance.
Analyses and Results
We identified a total of 9,383 nonduplicated cases, aged 65 and older, by Social Security number within the FAHIS data for the first 6-month period in 1999. Of these reports within the FAHIS database, a large proportion was not confirmed as coded in the FAHIS system. Outcomes of investigations showed that 38% of cases were "unfounded," 32% were judged as self-neglect, 24% were closed without classification, 3% were "proposed confirmed," 2% of cases were "confirmed," and 1% were coded as "no jurisdiction."
We used Social Security numbers to link FAHIS data with Medicaid data and separately with Medicare data for comparison with health service users not in the hotline database. Comparisons addressed total cost of services, costs of services per user, and services used as recorded within the Medicaid and Medicare datasets. We classified services as behavioral health or physical health services based on the primary diagnoses included on every claim for each person and then grouped them using the International Classification of Diseases 9th Revision Clinical Modification (ICD-9-CM) diagnosis code definitions. We classified behavioral health services as either mental health using ICD-9-CM codes 290, 293–302, and 306–315 or substance abuse using codes 291, 292 and 303–305. Physical health services comprised all other diagnostic codes. All participants in this study were service users with virtually all (more than 99%) using some form of physical health service.
We analyzed Medicaid and Medicare data separately. In the first set of analyses, we merged data from individuals aged 65 and older within the FAHIS data with Medicaid data based on Social Security numbers to identify nonduplicated cases of suspected elder mistreatment or self-neglect who were eligible for Medicaid services. We derived a comparison group by subtracting the first group from the pool of all 1999 Medicaid beneficiaries aged 65 and older. Similarly, we also created two such groups for Medicare data comparisons by merging the FAHIS and Medicare datasets.
Medicaid Data
There were 3,763 Medicaid enrollees (unduplicated cases) also identified within the FAHIS database, representing 40.2% of abuse hotline cases. Of these "abused" Medicaid enrollees, 3,198 (85%) used at least one service paid by Medicaid. The median age of this group was 81.2 (SD = 8.3), with 65% of cases being female. Primary categories of abuse reported were 38.8% for neglect, 21.4% for self-neglect, 29.8% for abuse, and 10.0% for exploitation.
Within the Medicaid files, there were 329,179 older service users in the second group, that is, those not identified in the FAHIS database. Of these, 205,031 (62%) used at least one Medicaid service in 1999. Their median age was 76.8 (SD = 8.6), with 70.2% being female.
Table 1 compares the categories of behavioral health services for both groups of Medicaid recipients. Within the FAHIS database, the most frequently used services were, in descending order: community mental health day treatment services; targeted case management, community mental health physician services, counseling, therapy, and treatment services; and rehabilitative services. Among service users not found in the FAHIS database, day treatment services were more frequently used by Medicaid claimants (37.2% vs 28.6%). However, all other percentages of the most frequently reported categories are remarkably similar.
|
|
|
There were 2,913,217 Medicare enrollees not identified in the FAHIS database. Their median age was 74.0 (SD = 7.4), with 57.0% being female. Of these enrollees, 58.5% used at least one Medicare service in 1999. Table 4 reveals a high penetration rate of 55% for behavioral health services among these service users. Although the penetration rate for individuals using only physical health services was higher among those not identified in the FAHIS database, the risk ratio again provided evidence that older adults were twice as likely to use behavioral health services if identified within the FAHIS database (RR = 4.69, p <.0001; 95% CI = 4.56–4.83).
|
|
|
|
Discussion |
|---|
|
TOPAbstractMethodDiscussionReferences |
|---|
The results lend support to the previous assumptions about the association between behavioral health and abuse. Individuals identified within the hotline data system more often utilized behavioral health services and appeared to incur greater costs compared to elders not reported to the hotline. Within the Medicaid files, the hotline database group had more than twice the penetration rate of service use, especially for those also using behavioral health services and at greater cost. Similarly, within the Medicare data, the rate of behavioral health services use among those in the FAHIS database was nearly 6 times greater than the other group's, with costs up to 3 times greater. Analysis of the Medicare data also suggests that those cases suspected of mistreatment or self-neglect may be using more expensive inpatient services as compared to outpatient or counseling and therapy services.
The conclusions of the study have a number of limitations that suggest the need for further investigation. First and foremost is the fact that most reports to abuse hotlines represent suspected, but not necessarily substantiated, cases by the APS system. As observed in the NEAIS, only 21% of cases were reported to and substantiated by APS agencies while 79% were reported by trained sentinels and not APS (NCEA, 1998). Such findings were similar to the rate of verification of the current study and indicate that identification and verification by APS is not an easy process. Recent research suggests that two thirds of claims of abuse being investigated by APS cannot be verified (Reynolds & Schonfeld, 2004). Thus, the hotline data are limited if used as the sole criteria for identifying abuse. Furthermore, the data from other service users not in the FAHIS database is likely to include "hidden" cases of mistreatment or self-neglect, that is, those who were never identified during the time period being investigated. This underidentification might reduce the differences observed in this study. Nevertheless, the penetration rates, odds ratios, and costs provide preliminary data supporting the human and economic cost of elder abuse, neglect, exploitation, and self-neglect.
A second limitation involves the quality of the data received. When matching FAHIS data with either Medicare or Medicaid data using Social Security numbers, a large portion of those numbers from the FAHIS system did not match up with Medicare data. One possibility is that the FAHIS data system may not verify the accuracy of Social Security numbers in many cases. A lesser possibility is the use of pseudo Social Security numbers for individuals reluctant to provide this information. A review of the Social Security numbers in the FAHIS data indicated that all the numbers appeared to have the appropriate number of digits (nine) and state prefixes (first three numbers). With a larger percentage of cases of suspected mistreatment or self-neglect, it is possible that the differences in service use, cost, and penetration rates would be even greater than that observed in this study. Another aspect of the data involves the skewness of the subgroup distributions. In most cases, the cost data were positively skewed, that is, the mean cost was higher than the median. The cost data is presented as potential evidence that individuals who are reported to the abuse hotline may require more care. Further investigation of cost data is certainly needed.
Finally, the limited number of variables within the FAHIS data provided to us did not allow us to pursue some interesting and relevant comparisons. We were not provided many of the basic descriptive data such as client demographic characteristics, the type of setting in which abuse occurred (e.g., institutional or community-based), and perpetrator characteristics. This limited our ability to identify which variables might best predict behavioral health service use.
Despite these limitations, an important implication for service providers and policy makers involves the need to enhance screening protocols to include behavioral health problems. Behavioral health screening by APS is rare (Teaster et al., n.d.). We believe that both behavioral health and elder abuse screening need to be implemented among a wider array of service providers, not only including APS investigators, but also aging services case managers, physicians, nurses, and other service providers. This combines the "trained sentinel" model recommended in the NEAIS (NCEA, 1998) with the Gatekeeper Model for behavioral health (Raschko, 1985). However, the incentive for providers to enhance their screening and referral protocols may have to be derived through state-level policy changes or through legislative mandates. Such changes have occurred for behavioral health in Florida. In 2000, Florida's Community Substance Abuse and Mental Health Services Act was amended by the Florida legislature (Florida Laws, Chapter 2000-349). The amendment required DCF to identify older persons with mental illness and/or substance abuse disorders as specific target populations for state funding. In 2003, in a special session, the legislature modified the state appropriations act by adding a requirement that the Florida DOEA and DCF implement a coordinated policy to identify and refer elderly persons with mental health or substance-abuse problems for treatment (Florida Laws, Chapter 2003-397). Unfortunately, implementing these mandates has been a difficult process.
As noted before, there is a need for further research given the limitations. Self-neglect should be more frequent among older adults living alone and may either result from, or contribute to, depression, alcohol problems, and other behavioral health problems. Mistreatment committed by others may lead to the decline in physical and mental health and result in increased need for services. Collection of more detailed victim information as well as perpetrator information (not available in the present study) might assist in determining this answer. Also, it would be helpful to investigate the extent to which acts of abuse by caregivers are antecedents to behavioral and physical health problems among victims or are reactions to the older adult's problems.
|
Footnotes |
|---|
1 Department of Aging & Mental Health, Louis de la Parte Florida Mental Health Institute, University of South Florida, Tampa.
2 Department Mental Health Law and Policy, Louis de la Parte Florida Mental Health Institute, University of South Florida, Tampa.
Decision Editor: Linda S. Noelker, PhD
Received for publication January 3, 2005. Accepted for publication July 11, 2005.
|
References |
|---|
|
TOPAbstractMethodDiscussionReferences |
|---|
20, s. 216.0166 Amendment of Community Substance Abuse and Mental Health Act. Retrieved November 30, 2005, from http://election.dos.state.fl.us/laws/00laws/convtbl.shtlm.This article has been cited by other articles:
 |
|
 |
|
  S. M. Strasser and T. Fulmer The Clinical Presentation of Elder Neglect: What We Know and What We Can Do Journal of the American Psychiatric Nurses Association, January 1, 2007; 12(6): 340 - 349. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
