Behavior Criminal Justice and

0 downloads 0 Views 162KB Size Report
paying for end-of-life care for inmates and ethical questions regarding the quality of medical ..... Bible studies during the week; in addition, religious services were held every weekend. .... emphysema, coma, and cancer (with and without pain).
Criminal Justice and Behavior http://cjb.sagepub.com/

Care Alternatives in Prison Systems: Factors Influencing End-of-Life Treatment Selection Laura L. Phillips, Rebecca S. Allen, Karen L. Salekin and Ronald K. Cavanaugh Criminal Justice and Behavior 2009 36: 620 DOI: 10.1177/0093854809334442 The online version of this article can be found at: http://cjb.sagepub.com/content/36/6/620

Published by: http://www.sagepublications.com

On behalf of:

International Association for Correctional and Forensic Psychology

Additional services and information for Criminal Justice and Behavior can be found at: Email Alerts: http://cjb.sagepub.com/cgi/alerts Subscriptions: http://cjb.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://cjb.sagepub.com/content/36/6/620.refs.html

>> Version of Record - May 7, 2009 What is This?

Downloaded from cjb.sagepub.com at Airlangga University on October 11, 2013

CARE ALTERNATIVES IN PRISON SYSTEMS Factors Influencing End-of-Life Treatment Selection LAURA L. PHILLIPS Michael E. DeBakey VA Medical Center, Houston, Texas

REBECCA S. ALLEN KAREN L. SALEKIN University of Alabama, Tuscaloosa

RONALD K. CAVANAUGH Alabama Department of Corrections, Montgomery

The authors examined age at the end of prison sentence, race, and psychosocial factors on end-of-life treatment preferences among 73 male inmates (28 nonlifers, 45 lifers) from the Alabama Aged and Infirmed Correctional Facility. All measures (e.g., Brief Symptom Inventory, Death Anxiety Scale) were administered in an interview format. A significant amount of variance in treatment preferences for cardiopulmonary resuscitation, feeding tube, and palliative care was predicted by race, lifer status, and death anxiety. Inmates who were members of minority groups, nonlifers, and those with high death anxiety expressed greater desire for a feeding tube, whereas inmates who were Caucasian or lifers expressed a greater desire for palliative care. Given the aging of the inmate population and increasing health care costs, further exploration of end-of-life treatment preferences among older inmates is warranted. Keywords:  end of life; treatment preferences; prison

G

iven the aging of the baby boomer generation, a greater number of individuals are facing decisions about their final years of life, prompting increased research focus on health issues and end-of-life care. This trend is seen within the prison system as a result of increased numbers of older inmates and the increasing length of sentences. The aging prison population has more than tripled since the early 1990s, representing one of the most dramatic changes in the American correctional system. One of every 23 inmates in prison today is age 55 or older, an 85% increase in the past 10 years. A 2004 report by the Sentencing Project found an 83% increase in the number of inmates serving a life sentence since 1992, yielding a total of 127,000 lifers (Mauer, King, & Young, 2004). AUTHORS’ NOTE: Portions of this article were presented at the American Psychological Association Board for the Advancement of Psychology in the Public Interest Symposium at the American Psychological Association Annual Meeting in Washington, D.C., in August 2005 and at the 58th annual scientific meeting of the Gerontological Society of America in Orlando, Florida, in November 2005. This research was supported by funding from the Center for Mental Health and Aging, University of Alabama, to R. S. Allen. Special thanks are extended to Michelle M. Hilgeman for assistance in data collection and to all of the inmates, correctional officers, and staff at the Alabama Aged and Infirmed Correctional Facility who gave their time and energy to this project. Address correspondence to Rebecca S. Allen, Ph.D., Department of Psychology/Center for Mental Health and Aging, University of Alabama, Box 870315, Tuscaloosa, AL 35487-0315. CRIMINAL JUSTICE AND BEHAVIOR, Vol. 36 No. 6, June 2009 620-634 DOI: 10.1177/0093854809334442 © 2009 International Association for Correctional and Forensic Psychology

620

Phillips et al. / CARE ALTERNATIVES IN PRISON SYSTEMS   621

The correctional system confronts the same issues the larger population encounters. However, this system—with few financial resources—is faced with the added dilemma of paying for end-of-life care for inmates and ethical questions regarding the quality of medical treatment offenders deserve (Aday, 2003). Notably, very little research has been conducted among older adults in the correctional system, partially due to inmates’ limited treatment options (Enders, Paterniti, & Meyers, 2005). This study seeks to provide information on the impact that age at the end of prison sentence, racial status, and psychosocial factors (i.e., emotional health, death anxiety, religiousness/spirituality) may have on end-of-life treatment preferences within the correctional system. PHYSICAL HEALTH AND AGE IN PRISONS

A 10-year difference between the health of inmates and that of the general population in Australia, making inmates’ physical health resemble that of someone approximately 10 years older than their chronological age, has been identified by Dawes (2002). This difference is believed to have developed due to inmates’ excessive drug and alcohol use, poor nutrition and eating habits, stressful life experiences, and lower socioeconomic status in comparison with nonoffenders. Unfortunately, there has not been any empirical evidence in the United States to support these hypotheses. However, the literature tends to use this 10-year difference to identify “older” (50 to 64 years) versus “elder” (65 years or older) offenders (Dawes, 2002; Florida Corrections Commission Annual Report, 1999; Gallagher, 1990, 2001; Steffensmeier & Motivans, 2000; Yates & Gillespie, 2000). Research as to the impact age has on inmates’ health varies based on the construct measured (i.e., subjective or objective health) and the comparison group (i.e., comparing older inmates to older community-residing adults or younger inmates). Initially, Gallagher (1990) found that 40% of Canadian older inmates saw the doctor three or more times in a 6-month period compared to 28% of inmates younger than 50 years of age. However, the objective health status of older inmates was similar to younger inmates when vision and hearing difficulties were discounted (Gallagher, 1990). A later study reported a difference in health service utilization based on offense history, with older inmates using more health services than older nonoffenders (Gallagher, 2001). The majority of inmates also have indicated that they believe themselves to be in “poor” health in comparison to same-aged nonoffenders (Aday, 1994). EMOTIONAL HEALTH IN PRISONS

Nearly 40% of state inmates 55 or older have a recent history or symptoms of mental health disorders (James & Glaze, 2006). Researchers have documented significant mental health problems among older inmates in U.S. prisons (Colsher, Wallace, Loeffelholz, & Sales, 1992; Loeb & AbuDagga, 2006; Regan, Alderson, & Regan, 2002) and among those of other nations (Fazel & Grann, 2002, 2004; Fazel & Lubbe, 2005). Older inmates may express increased levels of psychological distress when compared to younger inmates or community-residing older adults (D. Jones, 1976). Older inmates report feelings of depression, guilt, and psychological stress (Aday, 1994, 2003). Inmates report being away from their family and the stigma associated with their crime as two of the largest factors in their problems with emotional health (Aday, 1994).

622   CRIMINAL JUSTICE AND BEHAVIOR

Cultural differences and advanced age increase the heterogeneity of health care needs of older offenders, with differences emerging between those inmates who have been recently incarcerated and those who have spent extended amounts of time in prison (Yates & Gillespie, 2000). Although emotional health has not been examined as a factor influencing end-of-life treatment preferences within the prison setting, community-dwelling older adults with higher depression scores prefer more life-sustaining treatments than those with lower scores (Garrett, Harris, Norburn, Patrick, & Danis, 1993). Thus, it is possible that depression would affect the treatment preferences of older inmates, particularly given that these older adults are faced with the strain of being imprisoned and, potentially, dying within the correctional system (Aday, 2003). DEATH ANXIETY AMONG PRISON INMATES

We found only two studies that examined the death anxiety of inmates, neither of which were conducted with older adults. Stacey and Markin (1956) compared the death anxiety of prisoners, college students, foresters, and lawyers. This study found prisoners thought about their own deaths more frequently and vividly and that they were more concerned with their own deaths in comparison with the other groups. In addition, inmates score higher on the Death Anxiety Scale than most community populations (Templer, Barthlow, Halcomb, Ruff, & Ayers, 1979). Unfortunately, these two studies are dated, do not deal specifically with older inmates, and do not address the influence of death anxiety on end-of-life treatment preferences. Thus, more research is needed to evaluate the prevalence and potential influence of death anxiety among older inmates with regard to end-of-life medical care. The literature examining death anxiety has shown that older community-dwelling adults do not differ from other age groups on this issue (Fortner, Neimeyer, & Rybarczyk, 2000). There are differences, however, among subgroups of older adults. Specifically, those who were institutionalized and those with more physical problems expressed greater levels of death anxiety, whereas older adults who have stronger religious beliefs but not religious behavior had lower levels of death anxiety. RELIGIOUSNESS AND SPIRITUALITY IN PRISONS

Allen and colleagues (Allen, Phillips, Roff, Cavanaugh, & Day, 2008) examined religiousness/spirituality and mental health within the current sample (N = 73) of older male inmates. Their results expanded previous findings regarding this topic in three ways: (a) Participants were incarcerated primarily for murder and sexual crimes; (b) participants were 6 years older, on average, than those in prior research (Koenig, 1995); and (c) anxiety and desire for hastened death were examined in addition to depression. Allen and colleagues (2008) found that better self-reported health was associated with less anxiety and depression among older male inmates. More self-reported years of incarceration were associated with lower levels of experienced forgiveness (e.g., self-forgiveness, forgiveness of others, forgiveness by God). Interestingly, forgiveness, private religious practices, and religious meaning did not explain unique variance in anxiety, depression, or desire for hastened death. However, having more daily spiritual experiences was associated with less depression and less desire for hastened death, and feeling abandoned by God was associated with more symptoms of depression and a greater desire for hastened death. It appears

Phillips et al. / CARE ALTERNATIVES IN PRISON SYSTEMS   623

that the relation between religiousness/spirituality and mental health among older male inmates depends largely on whether the individual feels connected with or abandoned by a higher power. Koenig (1995) conducted interviews that combined several measures of religiosity to determine if older inmates differed from younger inmates with respect to religion. There were no age differences with respect to orthodoxy of belief, organized religious activity, religious experience, or religious coping. Although older inmates were not found to be substantially different from their younger peers, religious services are one of the few activities regularly available to inmates. Some inmates credit religious/spiritual conversion as the reason they have made changes in their drug-use habits or have decided to abide by society’s laws. Thus, it is conceivable that religious and/or spiritual beliefs and behavior may affect older inmates’ stated preferences regarding end-of-life medical treatments. Among community-dwelling older adults, DeLuca Havens (2000) found that religiosity influenced the execution of advance directives. Terminally ill participants who rated religion as very important in their lives expressed a greater desire for treatment (Garrett et al., 1993). In contrast, religious beliefs and practices among Jewish nursing home residents were not significant predictors of life-sustaining treatment preferences (Cohen-Mansfield et al., 1991). Moreover, church attendance was not found to predict preferences for aggressive medical interventions among ethnically diverse community populations (Sulmasy, Terry, & Miller, 1998). END-OF-LIFE ISSUES AND OLDER PRISON INMATES

Prior to implementing end-of-life services such as hospice or palliative care, it is important to understand what treatments are desired. Hospice care includes palliative and supportive services that provide physical, psychological, social, and spiritual care for dying persons and their families (A. Jones & Strahan, 1997). In contrast, palliative care extends the principles of hospice care to a broader population that could benefit from receiving aggressive comfort care earlier in the disease process (National Consensus Project for Quality Palliative Care, 2004). Palliative care is not restricted to individuals with a life expectancy of 6 months or less, and provision of palliative care is not exclusive of receipt of life-prolonging medical treatments. Inmates have indicated a desire for hospice within the prison system, including individualized treatment focused on pain management and providing ongoing emotional support until the time of death (Taylor, 2002). If educated regarding hospice, inmates express interest in receiving this care and in providing volunteer hospice services to other inmates (Dawes, 2002). However, although Taylor’s (2002) study focused on inmates’ desire to have hospice available to them, it did not examine other possible end-of-life treatment preferences (e.g., desire for cardiopulmonary resuscitation [CPR] or tube feeding). Presuming hospice is the preferred method of end-of-life care within the correctional system neglects substantial racial and ethnic differences (Kwak & Haley, 2005). CARE ALTERNATIVES IN PRISON SYSTEMS

Given the variability of findings regarding health and psychological factors in the community and the absence of prior research regarding end-of-life treatment preferences in the prison

624   CRIMINAL JUSTICE AND BEHAVIOR

setting, we conducted a study called Care Alternatives in Prison Systems to explore these variables. Life-sustaining medical care is an expensive enterprise, and prisons are cash-strapped systems with limited health care resources. Greater knowledge of end-of-life treatment preferences among older inmates may inform new policies and procedures for the disbursement of medical care resources within this complex setting burgeoning with older adults. The hypotheses for Care Alternatives in Prison Systems revolved around previously identified factors influencing the end-of-life treatment preferences of older adults residing within the community. This study focused on assessing similarities and differences between inmates whose sentence would end after the age of 75 (lifers) and those who would be released prior to turning 75 years old (nonlifers). The impact of age, physical health, emotional health, death attitudes, and religiousness/spirituality was examined for treatment preferences at the end of life. Regression models examining desire for CPR, desire for feeding tube, and desire for palliative care were utilized to determine the impact of demographic and psychosocial variables. Specifically, we hypothesized that inmates whose sentence would end after they turned 75 years of age would express less desire for life-prolonging treatments than those who would be released prior to becoming 75 years of age. We expected that older inmates who might live to be released from prison would prefer life-sustaining measures, allowing them to die on the outside. Second, we expected that members of minority racial groups would have a greater desire for life-sustaining treatments and a lower desire for palliative care than Caucasians (Allen, Haley, Roff, Schmid, & Bergman, 2006; Allen-Burge & Haley, 1997; Kwak & Haley, 2005). Third, in light of research among community-dwelling older adults, we predicted that inmates with more physical problems would express greater desire for palliative care measures and less desire for life-prolonging treatments than those in good physical health. Finally, we predicted that inmates with depression or anxiety would express a greater desire for life-sustaining treatments as would inmates with higher levels of death anxiety. We offered no specific predictions regarding the influence of religiousness and spirituality due to the conflicting results of prior studies but included these variables in our analyses. METHOD SETTING AND PARTICIPANTS

This study was approved by the University of Alabama institutional review board and officials with the Alabama Department of Corrections. Care Alternatives in Prison Systems was conducted in the Alabama Aged and Infirmed Correctional Facility in Hamilton, Alabama (Hamilton A & I), which is the facility dedicated to the incarceration of older and frail inmates. Approximately 300 adult males (81 were aged 50 or older) were housed at Hamilton A & I at the time of data collection (May 2004 to August 2004). Notably, Hamilton A & I did not have a special long-term care wing nor any special services for frail older inmates. However, in an effort to provide quality care for ill inmates, many life-prolonging treatments were provided, and there was also a hospice program. Inmates were placed into the hospice program if they had a terminal diagnosis and had decided to stop lifeprolonging treatments. With respect to religious or spiritual opportunities, all inmates who

Phillips et al. / CARE ALTERNATIVES IN PRISON SYSTEMS   625 TABLE 1:   Type of Crime Committed for Current Incarceration (N = 73) Type of crime Murder Sexual crime Financial crime Drug offense Assault Arson Othera

Frequency

Percentage

28 23 7 8 4 1 2

38.4 31.5 9.6 11.0 5.5 1.4 2.7

a. One prisoner refused to state why he was sentenced to life, and the other solicited someone to commit murder.

wished so at Hamilton A & I had a religious text at all times, even in administrative segregation. Two volunteer chaplains and members of several area churches visited the prison to hold Bible studies during the week; in addition, religious services were held every weekend. Inclusion criteria were that inmates must be aged 50 or older and have a Mini-Mental State Examination (MMSE) score of 15 or greater (Newman, 2003). Prior to beginning the interview, the reading subtest of the Wide Range Achievement Test, Third Edition (WRAT), was administered to obtain an estimation of literacy level (Manly, Byrd, Touradji, & Stern, 2004). Of the 81 inmates aged 50 or older at the time of data collection, 76 participants were recruited. Three inmates were discontinued because they scored less than 15 on the MMSE (Newman, 2003). Of the remaining 73 participants, 28 inmates’ release dates were prior to their 75th birthdays (nonlifers), and 45 inmates’ sentences either ended after their 75th birthdays or were life sentences (lifers). Out of the 73 participants, 49 were Caucasian (67%) and 24 (33%) identified with a minority racial group (primarily African American). Lifers versus nonlifers did not differ in the proportion of Caucasian inmates and inmates who identified themselves with a minority racial group, χ2(1, 72) = .166, p = .684, φ = .048. The majority of participants committed either murder (n = 28, 38%) or a sexual offense (n = 23, 32%). See Table 1 for additional information on the type of crime committed. Potential differences in the amount of experience with end-of-life issues, planning for end of life, and differences in sociodemographic variables were evaluated (see Table 2), with no significant differences found between lifers and nonlifers with the exception of age at the time of the interview, F(1, 72) = 12.74, p < .001, η2 = .152. Lifers were significantly older than nonlifers at the time of the interview. There were no significant differences between lifers and nonlifers (based on age at end of sentence) on any of the psychosocial measures. Means and standard deviations of each psychosocial measure by lifer and nonlifer status are shown in Table 3. MEASURES AND MATERIALS

MMSE (Folstein, Folstein, & McHugh, 1975). The MMSE is a brief cognitive assessment tool that reliably separates individuals with and without cognitive deficits. A cutoff of 15 out of 30 points was used due to the potential influence of educational level and lower socioeconomic status (Newman, 2003). The test–retest and interevaluator reliabilities are .89 and .83, respectively. The MMSE was included to exclude participants whose selfreport may be of questionable validity due to cognitive impairment.

626   CRIMINAL JUSTICE AND BEHAVIOR TABLE 2:   Sociodemographic Variables, by Nonlifer and Lifer Status Demographic Variable

Nonlifer (EOS Before Age 75)

Lifer (EOS After Age 75)

Categorical Variable

n

%

n

%

Experience with end of life Informal advance care planning    discussions with family Possesses a living will Has appointed a health    care proxy

20 8

27.4 11.0

27 16

37.0 21.9

8 8

11.0 11.1

9 17

12.3 23.6

Continuous Variable

M

SD

M

SD

11.07 59.29**

4.19 6.89

11.13 66.42**

3.76 9.07

Years of education Age at interview Note. EOS = end of sentence. **p < .01.

TABLE 3:   Means and Standard Deviations of Psychosocial Variables, by Nonlifer and Lifer Status Psychosocial Measure MMSE score WRAT absolute score Physical Health subscale Subjective Health Comorbidity Symptoms Stress Symptoms Lifestyle Health Care Utilization Emotional Health subscales BSI anxiety BSI depression Death anxiety Religiousness/spirituality Religious behaviors Religious/spiritual beliefs

Nonlifer (EOS Before Age 75)

Lifer (EOS After Age 75)

M

SD

M

SD

25.71 501.25

4.06 31.54

26.29 510.71

3.76 22.05

11.39 2.55 1.79 1.73 6.97

4.00 1.56 1.71 0.95 4.37

10.31 2.74 1.97 1.52 6.51

3.73 1.63 1.71 1.14 3.36

3.18 3.39 4.89

3.97 4.12 2.44

3.96 4.40 4.89

4.32 4.18 3.29

35.50 61.24

9.32 8.08

36.44 59.23

10.76 14.10

Note. EOS = end of sentence; MMSE = Mini-Mental State Examination; WRAT = Wide Range Achievement Test, Third Edition; BSI = Brief Symptom Inventory. There are no significant differences between lifers and nonlifers on these scales (p < .05).

WRAT (Wilkinson, 1993). The WRAT reading subtest provides an estimated grade-level equivalence of reading ability. The median alpha coefficients were .94 (range = .91 to .95) for the Blue Reading subtest and .93 (range = .89 to .94) for the Tan Reading subtest for ages 45 to 74. The median correlation between these alternate forms for ages 45 to 74 was .94 (range = .91 to .96). The WRAT reading subtest was included as a better measure of minority inmates’ ability to provide valid responses to self-report questions administered in interview format because the MMSE is highly related to educational attainment (Manly et al., 2004).

Phillips et al. / CARE ALTERNATIVES IN PRISON SYSTEMS   627

Demographics. Participants meeting inclusion criteria were asked to provide their age, race, marital status, years of education, prison offense, prison sentence, and experience with end-of-life issues including previous caregiving or decision-making experiences for individuals (including the self) near the end of life and informal advance care planning discussions with family, possession of a living will, or appointment of a health care proxy. Physical Health (Stewart, Hayes, & Ware, 1994). This measure is an adapted version of the 12-item Medical Outcomes Study Short-Form General Health Survey (Stewart et al., 1994) used in the REACH I study (Schulz et al., 2003). The Medical Outcomes Study has good internal consistency (greater than .85) and good evidence for validity. Questions regarding the frequency of doctor visits, injuries, and number of medications taken were added to create the measure titled Physical Health. This measure consists of five subscales: Subjective Health, Comorbidity Symptoms, Stress Symptoms, Lifestyle, and Health Care Utilization. Emotional health: Brief Symptom Inventory, Third Edition (Derogatis, 1993). Emotional health was measured by the Depression (6 items) and Anxiety (6 items) subscales of the Brief Symptom Inventory to reduce the number of items administered (12 items). Reliability estimates for these subscales are good, with an alpha ranging from .80 to .85. Test–retest reliability for these scales ranges from .68 to .84. Construct validity estimates are also uniformly high for these subscales, with factor loadings ranging from .42 to .65 for Depression and .49 to .57 for Anxiety (Derogatis, 1993). Death Anxiety Scale (Templer, 1970, 1995). The Death Anxiety Scale consists of 15 true/ false items with statements varying among the fear evoked when discussing death, thinking about how one wishes to die, and the emotions generated when thinking about illness. The Death Anxiety Scale has strong internal consistency (Cronbach’s alpha = .76) and test– retest reliability following a 3-week period (.83 among college students; Templer, 1970). In Templer et al.’s (1979) examination of inmates (age range = 17 to 53), the average death anxiety score was 6.07 (SD = 3.63), whereas the average death anxiety score of older (age range 60 to 83) community-dwelling males (Stevens, Cooper, & Thomas, 1980) was 4.91 (SD = 2.79). Religiousness and spirituality: Brief Multidimensional Measure of Religion and Spirituality (Fetzer Institute, 1999). This instrument was designed for use in health research and measures the impact of spirituality and religion on daily life. The original Multidimensional Measure of Religion and Spirituality consists of 38 items in 11 subscales. We used 7 subscales of the Brief Multidimensional Measure of Religion/Spirituality consisting of 26 Likert-type items: Daily Spiritual Experiences (α = .91), Values/Beliefs (α = .64), Forgiveness (α = .66; self-forgiveness, forgiveness by others, forgiveness by God), Private Religious Practices (α = .72), Religious and Spiritual Coping (positive α = .81, negative α = .54), Religious Intensity (α = .77), and Meaning (no α provided; Fetzer Institute, 1999). Due to substantial multicollinearity among these 6 subscales, an exploratory factor analysis with varimax rotation was performed, resulting in two factors with little shared variance. Based on examination of the items loading on each of these factors, we labeled them religious behaviors and religious/spiritual beliefs. These two factors are used throughout our analyses.

628   CRIMINAL JUSTICE AND BEHAVIOR

Life-Support Preferences/Predictions Questionnaire (Ditto et al., 2001). The LifeSupport Preferences/Predictions Questionnaire describes nine illness situations covering conditions varying in severity of illness, prognosis, and level of pain. This measure has an eighth-grade reading level and has been used in several community studies (Bookwala et al., 2001; Coppola et al., 1999; Ditto et al., 2001; Fagerlin, Ditto, Danks, & Houts, 2001). This is the first use of the Life-Support Preferences/Predictions Questionnaire among prison inmates. Therefore, we attempted to decrease the complexity of this measure by administering only four of the nine original scenarios: Alzheimer’s disease, emphysema, and cancer with and without pain. Treatment options for each of the four scenarios consisted of CPR, feeding tube, and palliative care. Interviewers paid particular attention to assisting inmates in understanding the term “palliative care,” but this attempt may not have been successful. Inmates selected a number between 1, indicating low preference for the treatment, and 5, indicating high preference for the treatment, to represent their desire for that specific treatment for each of the four illness scenarios. The treatment options were selected based on prior research (Allen-Burge & Haley, 1997) and because they offered varying levels of intervention for each illness. Coppola and colleagues (1999) found high internal consistency (Cronbach’s α = .86 to .96) among preference for life-sustaining treatments, by scenario, for Alzheimer’s disease, emphysema, coma, and cancer (with and without pain). Therefore, we averaged inmates’ treatment preferences across these four illness scenarios for each end-of-life treatment in all of our subsequent analyses, resulting in a preference range of 1 to 5 for each treatment. PROCEDURE

All inmates aged 50 years or older were invited to participate in the screening process for the study. Each potential participant was given an informed consent explaining the nature of the study and requesting his participation. On signing the consent form, the inmate was screened using the MMSE. If an inmate scored less than 15 on the MMSE, the interview ended and the inmate was thanked for his time. If the inmate scored 15 or greater, the interview continued (Newman, 2003). The WRAT was given immediately following the MMSE to obtain an estimate of achieved reading ability based on the recommendations of Manly and colleagues (2004) as a potential estimate of an inmate’s understanding of selfreport items. All measures were administered in interview format and were read aloud by the interviewer. Response cards for Likert-type scales were given to inmates to help facilitate accurate responding. Interviews lasted between 60 and 90 minutes. All information gathered during the interview was strictly confidential. Neither inmates nor the prison system had access to specific responses. DATA ANALYSIS

Prior to descriptive and inferential analyses, each variable was evaluated for internal consistency (see Table 4), normality, missing data, and outliers to ensure all assumptions were met for each statistical test utilized. None of the participants had more than 1 missing item on any scale. Missing data were imputed using the average of the scale for the missing value. To obtain a single indicator of treatment preference, preferences for CPR, feeding tube, and palliative care were averaged across illness scenarios, as in prior research (Bookwala et al., 2001; Ditto et al., 2001).

Phillips et al. / CARE ALTERNATIVES IN PRISON SYSTEMS   629 TABLE 4:   Internal Consistency of Scales Scale Physical Health subscale Subjective Health Comorbidity Symptoms Stress Symptoms Lifestyle Health Care Utilization Emotional Health subscale BSI anxiety BSI depression Death anxiety Religiousness/spirituality Religious behaviors Religious/spiritual beliefs

Number of Items

Cronbach’s Alpha

Range of Interitem Correlations

4 7 3 4 5

.65 .48 .57 .26 .62

.277 .129 .324 .031 .025

6 5 15

.77 .80 .71

.441 to .621 .363 to .745 .129 to .567

7 16

.83 .93

.425 to .714 .499 to .802

to to to to to

.579 .354 .417 .290 .392

Note. BSI = Brief Symptom Inventory.

TABLE 5:   Correlations Between Desired Treatments and Psychosocial Variables Psychosocial Variable

Cardiopulmonary Resuscitation

Age at interview Physical Health subscales Subjective Health Comorbidity Symptoms Stress Symptoms Lifestyle Health Care Utilization Emotional Health subscales BSI anxiety BSI depression Death anxiety Religiousness/spirituality Religious behaviors Religious/spiritual beliefs

Feeding Tube

Palliative Care

–.156

–.113

.131

.050 –.138 –.085 .042 –.171

.225 –.135 –.163 .002 –.192

–.058 .087 .060 –.032 .221

–.012 –.204 .231*

–.050 –.198 .325**

.140 .178 –.096

.146 .263*

.283* .386**

–.255* –.247*

Note. BSI = Brief Symptom Inventory. *p < .05. **p < .01.

Based on findings among community-dwelling older adults, we predicted that age at the end of an inmate’s sentence, racial group, physical health, emotional health, death anxiety, and religiousness and spirituality would influence end-of-life treatment selection. To explore our specific hypothesis regarding each of these variables, bivariate correlations were examined to verify that these variables were significantly associated with treatment preferences for CPR, feeding tube, and palliative care. Age at the time of interview was also correlated with end-of-life treatment preferences due to the significant differences between the lifer and nonlifer groups. The correlation matrix is shown in Table 5. Notably, age at the time of the interview was not significantly associated with CPR, feeding tube, or palliative care preferences. Thus, this variable was dropped from subsequent analyses. Only variables that correlated at p < .05 were included in each of the three regression models predicting preferences for CPR, feeding tube, and palliative care. The final regression

630   CRIMINAL JUSTICE AND BEHAVIOR TABLE 6:   Summary of Regression Analysis for Variables Predicting Treatment Preference (N = 73) Treatment Cardiopulmonary resuscitation Feeding tube Palliative care

Variable

Standard β

t

Age at end of sentence Racial group Death anxiety Religious/spiritual beliefs

–.104 .178 .210 .196

–0.924 1.559 1.872 1.708

.359 .124 .066 .092

Age at end of sentence Racial group Death anxiety Religious/spiritual beliefs Religious behaviors

–.204 .236 .302 .215 .123

–2.025 2.309 3.020 1.694 0.976

.047* .024* .004** .095 .332

Age at end of sentence Racial group Religious/spiritual beliefs Religious behaviors

.238 –.258 –.067 –.178

2.174 –2.322 –0.487 –1.305

.033* .023* .628 .196

p

Note. Cardiopulmonary resuscitation F(4, 68) = 3.127, p = .020, R 2 = .155. Feeding tube F(5, 67) = 6.948, p < .001, R 2 = .341. Palliative care F(4, 68) = 4.470, p = .003, R 2 = .208. *p < .05. **p < .01.

model for each of our dependent variables (e.g., CPR, feeding tube, palliative care) included age at the end of an inmate’s sentence (e.g., lifer or nonlifer), racial status, and any psychosocial variable displaying a significant bivariate association with that dependent variable. RESULTS

The results of each of the three regression analyses are shown in Table 6. The full model predicting preferences for CPR included age at the end of an inmate’s sentence (e.g., lifer or nonlifer), racial group, death anxiety, and religious/spiritual beliefs. This model was significant, F(4, 68) = 3.127, p = .020, R2 = .155. None of the individual predictors, however, had a uniquely significant association with preference for CPR. The full model predicting preference for a feeding tube included age at the end of an inmate’s sentence (e.g., lifer or nonlifer), racial group, death anxiety, religious behaviors, and religious/spiritual beliefs. This model was significant, F(5, 67) = 6.948, p < .001, R2 = .341. Significant predictors included age at end of sentence (β = –.204, t = –2.025, p = .047), racial group (β = .236, t = 2.309, p = .024), and death anxiety (β = .302, t = 3.020, p = .004). As predicted, inmates who were nonlifers or members of minority racial groups expressed a greater desire for a feeding tube. Furthermore, inmates reporting high death anxiety also expressed a greater desire for a feeding tube. The full model predicting preference for palliative care included age at the end of an inmate’s sentence (e.g., lifer or nonlifer), racial status, religious behaviors, and religious/ spiritual beliefs. This model explained a significant amount of variance in preference for palliative care, F(4, 68) = 4.470, p = .003, R2 = .208, and was influenced by age at the end of sentence (β = .238, t = 2.174, p = .033) and by racial group (β = –.258, t = –2.322, p = .023). As predicted, lifers and Caucasians expressed a greater desire for palliative care.

Phillips et al. / CARE ALTERNATIVES IN PRISON SYSTEMS   631

DISCUSSION

The unique setting of prison creates an important factor in end-of-life decisions— whether the inmate will die as a free individual. Our study was the first to examine end-oflife treatment preferences of older male prison inmates. We offered five specific predictions regarding the associations among age at the end of an inmate’s sentence, racial group, physical health, emotional health, and death anxiety with end-of-life treatment preferences for three specific medical treatments: CPR, feeding tubes, and palliative care. We found support for our first hypothesis, that inmates whose sentence would end after they turned 75 years of age would express less desire for life-prolonging treatments than those who would be released prior to becoming 75 years of age. As expected, those who might be released from prison desired life-prolonging measures whereas those who were likely to die within the prison setting expressed preferences for palliative care. However, inmates’ age at the time of the interview was unrelated to any of our end-of-life medical treatment preferences. This difference possibly represents the desire to die outside of prison. If dying outside of prison is unlikely, a preference shift occurs, with lifers electing to be kept out of pain and letting life, or death, take its course within the prison setting. Thus, future research and policy developments within the prison system should consider whether an inmate has the opportunity to be released from prison in designing end-of-life care alternatives and allocating resources to these treatment options. We also found support for our second hypothesis, that members of a minority racial group would have a greater desire for life-sustaining treatments and a lower desire for palliative care than Caucasians. This finding coincides with previous research conducted with community-dwelling older adults (Kwak & Haley, 2005). Specifically, members of minority racial/ethnic groups were more likely to prefer feeding tubes whereas Caucasians were more likely to express preferences for palliative care. Future research should attempt to identify and examine specific cultural factors that may explain these racial/ethnic differences. For example, it would be interesting to explore trust in the prison health care system and cultural justifications for end-of-life treatment selection in future studies. Our third hypothesis concerning physical health and our fourth hypothesis concerning emotional health were not supported. We expected inmates with more physical problems to express less desire for life-prolonging measures and greater desire for palliative care. We also expected that inmates with depression or anxiety would express greater desire for lifesustaining treatments and less desire for palliative care. However, we found no significant bivariate associations among physical or emotional health measures and desire for end-oflife medical treatments. Notably, the internal consistency of some of the physical health subscales in the measure used in this study was poor. Future studies should use standardized and well-validated measures such as the Short-Form 36 (SF-36) to explore the relations among physical and emotional health and end-of-life treatment selection. A unique contribution of this study was our inclusion of a death anxiety measure in predicting the end-of-life treatment preferences of older prison inmates. We expected and found that inmates with high levels of death anxiety would report greater preferences for life-sustaining measures. Death anxiety significantly predicted desire for a feeding tube (29% explained variance) and was included in the model predicting desire for CPR wherein no individual variable was found to offer unique variance in explaining desire for treatment. It could be that as we predicted and in accordance with prior research (Aday, 2003),

632   CRIMINAL JUSTICE AND BEHAVIOR

inmates fear dying within the prison. Notably, inmates at Hamilton A & I were familiar with feeding tubes due to the presence of an inmate with an enteral feeding tube within the population. Because death anxiety has not been widely explored as a factor influencing end-of-life treatment selection, future research should consider this important variable— particularly among correctional populations. Finally, we explored the relation between religious behaviors and religious/spiritual beliefs found to be important predictors of mental health among aging inmates (Allen et al., 2008; Koenig, 1995) with end-of-life treatment preferences. We found that although religious behaviors and religious/spiritual beliefs had significant bivariate associations with all of our end-of-life treatments (e.g., CPR, feeding tube, palliative care), these variables were not significant predictors in any of the regression models. Thus, our findings coincide with prior research among Jewish nursing home residents (Cohen-Mansfield et al., 1991) and ethnically diverse community populations (Sulmasy et al., 1998), offering limited support for the influence of religious or spiritual variables on desires for specific end-of-life medical treatments. Future research targeting explanations of racial/ethnic differences in end-of-life care preferences may need to examine variables other than religiousness or spirituality. Although this study makes a significant contribution to understanding end-of-life treatment preferences within the prison context, a few caveats should be noted. First, the prison where this study was conducted contains predominately Caucasian and African American inmates. Thus, our findings should not be extrapolated to other racial groups. Furthermore, this study was completed at one facility, which is a designated facility in Alabama for aged and infirm inmates. Thus, inmates retained at Hamilton A & I are exposed to terminally ill offenders and witness inmates dying while imprisoned on a regular basis, perhaps increasing their familiarity with end-of-life illnesses and treatments even if such circumstances represent, arguably, the worst way to die (Aday, 2003). Notably, however, many inmates had to ask for explanations regarding the differences between palliative care and life-sustaining treatment during the interviews. From further comments made by the inmates, it is questionable whether they fully understood the ramifications of the treatment preferences they expressed, even though we tried to facilitate understanding through screening with the MMSE and the WRAT reading subtest. There is a need to educate inmates as to the benefits and risks of end-of-life treatments to ensure informed decisions (Kwak & Haley, 2005). It seems unlikely that mere exposure to terminal illnesses or end-of-life treatments increased inmates’ knowledge, although it indeed may have sensitized inmates to what dying in a prison would mean. Future research and policy initiatives within the prison setting should target prisoner education regarding illness trajectories and benefits and drawbacks of specific end-of-life medical treatments. Finally, like all studies that seek to examine the impact of a variety of factors on end-of-life treatment preferences, the findings of this study are based on vignettes and not actual clinical illnesses experienced by these inmates. Thus, this study offers limited insight into the actual treatment preferences of older male inmates during the course of a life-threatening illness. Despite these limitations, the study provides information that has not heretofore been available. The need for chronic and end-of-life health care has increased as more offenders are aging and dying in prison. Health care resources, however, have not increased within the prison system during this period of time. Thus, understanding the end-of-life treatment preferences of older male inmates is a critical first step in developing future policies for implementing quality end-of-life care within this setting.

Phillips et al. / CARE ALTERNATIVES IN PRISON SYSTEMS   633

REFERENCES Aday, R. H. (1994). Aging in prison: A case-study of new elderly offenders. International Journal of Offender Therapy and Comparative Criminology, 38, 79-91. Aday, R. H. (2003). Aging prisoners: Crisis in American corrections. Westport, CT: Praeger. Allen, R. S., Haley, W. E., Roff, L. L., Schmid, B., & Bergman, E. J. (2006). Responding to the needs of caregivers near the end of life: Enhancing benefits and minimizing burdens. In J. L. Werth Jr. & D. Blevins (Eds.), Psychosocial issues near the end of life: A resource for professional care providers (pp. 183-201). Washington, DC: American Psychological Association. Allen, R. S., Phillips, L. L., Roff, L. L., Cavanaugh, R., & Day, L. (2008). Religiousness/spirituality and mental health among older male inmates. The Gerontologist, 48, 692-697. Allen-Burge, R., & Haley, W. E. (1997). Individual differences and surrogate medical decisions: Differing preferences for life-sustaining treatments. Aging & Mental Health, 1, 121-131. Bookwala, J., Coppola, K. M., Fagerlin, A., Ditto, P. H., Danks, J. H., & Smucker, W. D. (2001). Gender differences in older adults’ preferences for life-sustaining medical treatments and end-of-life values. Death Studies, 25, 127-149. Cohen-Mansfield, J., Rabinovich, B. A., Lipson, S., Fein, A., Gerber, B., Weisman, S., et al. (1991). The decision to execute a durable power of attorney for health-care and preferences regarding the utilization of life-sustaining treatments in nursing-home residents. Archives of Internal Medicine, 151, 289-294. Colsher, P. L., Wallace, R. B., Loeffelholz, P. L., & Sales, M. (1992). Health status of older male prisoners: A comprehensive survey. American Journal of Public Health, 82, 881-884. Coppola, K. M., Bookwala, J., Ditto, P. H., Lockhart, L. K., Danks, J. H., & Smucker, W. D. (1999). Elderly adults’ preferences for life-sustaining treatments: The role of impairment, prognosis, and pain. Death Studies, 23, 617-634. Dawes, J. (2002). Dying with dignity: Prisoners and terminal illness. Illness, Crisis & Loss, 10, 188-203. DeLuca Havens, G. A. (2000). Differences in the execution/nonexecution of advance directives by community dwelling adults. Research in Nursing & Health, 23, 319-333. Derogatis, L. R. (1993). Brief Symptom Inventory: Administration, scoring, and procedures manual (3rd ed.). Minneapolis, MN: National Computer Systems. Ditto, P. H., Danks, J. H., Smucker, W. D., Bookwala, J., Coppola, K. M., Dresser, R., et al. (2001). Advance directives as acts of communication: A randomized controlled trial. Archives of Internal Medicine, 161, 421-430. Enders, S. R., Paterniti, D. A., & Meyers, F. J. (2005). An approach to develop effective health care decision making for women in prison. Journal of Palliative Medicine, 8, 432-439. Fagerlin, A., Ditto, P. H., Danks, J. H., & Houts, R. M. (2001). Projection in surrogate decisions about life-sustaining medical treatments. Health Psychology, 20, 166-175. Fazel, S., & Grann, M. (2002). Older criminals: A descriptive study of psychiatrically examined offenders in Sweden. International Journal of Geriatric Psychiatry, 17, 907-913. Fazel, S., & Grann, M. (2004). Psychiatric morbidity among homicide offenders: A Swedish population study. American Journal of Psychiatry, 161, 2129-2131. Fazel, S., & Lubbe, S. (2005). Prevalence and characteristics of mental disorders in jails and prisons. Current Opinion in Psychiatry, 18, 550-554. Fetzer Institute/National Institute on Aging Work Group. (1999). Multidimensional Measurement of Religiousness/Spirituality for use in health research. Kalamazoo, MI: John E. Fetzer Institute. Florida Corrections Commission annual report. (1999). Retrieved August 8, 2005, from http://www.fcc.state.fl.us/fcc/ reports/final99/1eld.html Folstein, M. F., Folstein, S. E., & McHugh, R. R. (1975). Mini-Mental State: A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, 189-198. Fortner, B. V., Neimeyer, R. A., & Rybarczyk, B. (2000). Correlates of death anxiety in older adults: A comprehensive review. In A. Tomer (Ed.), Death attitudes and the older adult: Theories, concepts, and applications (pp. 95-108). New York: Brunner-Routledge. Gallagher, E. M. (1990). Emotional, social, and physical health characteristics of older men in prison. International Journal of Aging and Human Development, 31, 251-265. Gallagher, E. M. (2001). Elders in prison: Health and well-being of older inmates. International Journal of Law and Psychiatry, 24, 325-333. Garrett, J. M., Harris, R. P., Norburn, J. K., Patrick, D. L., & Danis, M. (1993). Life-sustaining treatments during terminal illness: Who wants what. Journal of General Internal Medicine, 8, 361-368. James, D. J., & Glaze, L. E. (2006). Mental health problems of prison and jail inmates. Washington, DC: Bureau of Justice Statistics. Jones, A., & Strahan, G. (1997). The national home and hospice care survey: 1994 summary. Vital Health Statistics 13, 126, 1-124. Jones, D. (1976). The health risks of imprisonment. Toronto, Canada: Lexington Books.

634   CRIMINAL JUSTICE AND BEHAVIOR Koenig, H. G. (1995). Religion and older men in prison. International Journal of Geriatric Psychiatry, 10, 219-230. Kwak, J., & Haley, W. E. (2005). Current research findings on end-of-life decision making among racially or ethnically diverse groups. The Gerontologist, 45, 634-641. Loeb, S. J., & AbuDagga, A. (2006). Health-related research on older inmates: An integrative review. Research in Nursing & Health, 29, 556-565. Manly, J. J., Byrd, D. A., Touradji, P., & Stern, Y. (2004). Acculturation, reading level, and neuropsychological test performance among African American elders. Applied Neuropsychology, 11, 37-46. Mauer, M., King, R., & Young, M. C. (2004). The meaning of “life”: Long prison sentences in context. Retrieved August 8, 2006, from http://www.sentencingproject.org/Publications.aspx National Consensus Project for Quality Palliative Care. (2004). Clinical practice guidelines for quality palliative care. Retrieved February 22, 2009, from www.nationalconsensusproject.org Newman, J. (2003, November 7). Information processing deficits in psychopathic offenders: Implications for self-regulation and violence. Paper presented at the University of Alabama Psychology Colloquium Series, Tuscaloosa. Regan, J. J., Alderson, A., & Regan, W. M. (2002). Psychiatric disorders in aging prisoners. Clinical Gerontologist, 26(1/2), 117-124. Schulz, R., Belle, S. H., Czaja, S. J., Gitlin, L. N., Wisniewski, S. R., Ory, M. G., & REACH Investigators. (2003). Introduction to the special section on Resources for Enhancing Alzheimer’s Caregiver Health (REACH). Psychology and Aging, 18, 357-360. Stacey, C. L., & Markin, K. (1956). The attitudes of college students and penitentiary inmates toward death and future life. Psychiatric Quarterly, 26(Suppl. 1), 27-32. Steffensmeier, D., & Motivans, M. (2000). Older men and older women in the arms of criminal law: Offending patterns and sentencing outcomes. Journal of Gerontology, 55B, S141-S151. Stevens, S. J., Cooper, P. E., & Thomas, L. E. (1980). Age norms for Templer’s Death Anxiety Scale. Psychological Reports, 46, 205-206. Stewart, A. L., Hayes, R. D., & Ware, J. E. (1994). The medical outcomes study short-form general health survey. In J. E. Mezzich, M. R. Jorge, & J. M. Sallom (Eds.), Psychiatric epidemiology: Assessment concepts and methods (pp. 368-382). Baltimore: Johns Hopkins University Press. Sulmasy, D. P., Terry, P. B., & Miller, D. J. (1998). Accuracy of substituted judgments in patients with terminal diagnoses: In response. Annals of Internal Medicine, 129, 1083-1083. Taylor, P. B. (2002). End-of-life care behind bars. Illness, Crisis & Loss, 10, 233-241. Templer, D. I. (1970). Construction & validation of Death Anxiety Scale. Journal of General Psychology, 82, 165-177. Templer, D. I. (1995). Anxiety disorders: Death Anxiety Scale. In N. S. Schutte & J. M. Malouff (Eds.), Sourcebook of adult assessment strategies (pp. 174-176). New York: Plenum. Templer, D. I., Barthlow, V. L., Halcomb, P. H., Ruff, C. F., & Ayers, J. L. (1979). The death anxiety of convicted felons. Corrective & Social Psychiatry & Journal of Behavior Technology, Methods, & Therapy, 25, 18-20. Wilkinson, G. S. (1993). The Wide Range Achievements Test—Third edition (WRAT-III). Wilmington, DE: Wide Range. Yates, J., & Gillespie, W. (2000). The elderly and prison policy. Journal of Aging and Social Policy, 11, 167-175.

Laura L. Phillips, PhD, received her PhD from The University of Alabama in 2007. Her primary research and clinical interests are program evaluation and interventions to improve the quality of life of veterans, prison inmates, and older adults living in rural areas. Rebecca S. Allen, PhD, received her PhD from Washington University in St. Louis in 1994. Her primary research and clinical interests are interventions to reduce the stress of family and professional caregivers for older adults with terminal and chronic illness. Karen L. Salekin, PhD, received her doctorate from the University of North Texas in 1997, and was a postdoctoral fellow at the University of Massachusetts Medical Center, Department of Psychiatry. Her major research interest focuses on forensic assessment and includes issues of mental retardation, capital mitigation, competence to stand trial, competence to waive Miranda rights, criminal responsibility, and risk and protective factors in juvenile delinquency. Ronald K. Cavanaugh, PsyD, is the chief psychologist for the Department of Corrections in the State of Alabama.