Paranoid Ideation Among Elderly African American Persons

Abstract

Purpose: The prevalence and correlates of paranoid ideation were investigated. Design and Methods: On the basis of a sample of 998 independently living elderly African American persons in a cross-sectional study, the study used the Brief Symptom Inventory to measure paranoid ideation and 14 independent variables, including demographic characteristics, cognitive deficit, depression, self-reported memory functioning, emotional and instrumental support, stressful life events, limitation of daily activities, self-rated health status, and self-rated hearing and vision. Results and Implications: Paranoid ideation (symptoms of paranoia) was found in 10% of this sample. A multiple regression analysis of the data revealed that of the 14 independent variables used in this study, 6 (income, instrumental support, hearing, stressful life events, self-reported memory deficit, and depression) showed a significant relationship with paranoid ideation.

Vernon L. Greene, PhD

The limited available epidemiological studies of paranoid ideation among elderly people indicate a steady increase in symptoms among this segment of the population over the past three decades (Christenson and Blazer 1984; Blazer, Hays, and Salive 1996; Forsell and Henderson 1998; Lowenthal 1964). The Diagnostic and Statistical Manual of Mental Disorders (4th ed., or DSM–IV: American Psychiatric Association 1994) makes only a limited reference to paranoia in elderly people. The terms paranoid ideation or paranoid symptom, as used in this article, refer to a distinct clinical entity: a suspicion or delusional disorder that is nonschizophrenic (Lewis 1970; Rabins 1992) and that is characterized by the tendency to view individuals or agencies with suspicion or as having harmful intentions. Studies suggest that only a small percentage of elderly persons suffering from paranoia are adequately diagnosed and treated (Forsell and Henderson 1998; Wykle, Segall, and Nagley 1992). This documented increase in symptoms, together with reports of lack of proper diagnosis and treatment of paranoid ideation among the rapidly growing elderly population, suggests the need for more careful examination of the problem.

The literature that has been published on this topic has stressed the clinical importance of the following as high-risk or contributing factors for late-life paranoid symptoms: sensory deficits; social isolation; medical illness; cognitive impairment; confusion; the aging process; poor self-rated physical and mental health (Jette and Winnett 1987); poor education and economic status (Terrell and Barrett 1979); sense of deprivation and injustice (Fenigstein and Vanable 1992); sense of loss of control over the environment (Mirowsky and Ross 1983); and psychiatric symptoms, particularly depressive symptoms (Almeida, Howard, Levy, and David 1995; Blazer et al. 1996; Christenson and Blazer 1984; Eisdorfer 1980).

It is now well established that the study of psychiatric disorders should be conducted in the context of both social and cultural perspectives (Fabrega, Mezzich, and Ulrich 1988; Strickland, Longobardi, and Gray 1999). In the current literature, both Fabrega and associates 1988 and Newhill 1990 have raised these issues, which are of primary importance to our study, namely the role of social and cultural background in addition to the individual psychopathology. These authors indicate the need to attend to sociocultural differences in the expression of paranoid symptomatology. For example, self-protective defensive behavior among African Americans may be perceived as paranoid symptomatology if the oppressive socioeconomic context is ignored (Newhill 1990).

The purpose of this exploratory study was to expand understanding of paranoid symptomatology among urban, elderly African American persons using senior center facilities by examining the independent impact of demographics, social support, cognitive deficit, and health status self-ratings on health, vision, hearing, memory, and personal characteristics on paranoid symptoms. The study of paranoid ideation and its correlates may yield data that are useful in establishing treatment equity for this population (Bazargan and Hamm-Baugh 1995).

Methods

Sample Selection

The study sample was composed of 1,114 elderly African Americans, ages 62 years and older, living independently in New Orleans, Louisiana. They were randomly selected from the service lists for 23 senior citizen centers in the city. The participants were notified by a personal letter describing the nature of the study and selection procedures for inclusion of participants. To encourage voluntary participation, each potential participant was offered remuneration. A few days after personal letters were mailed, selected participants were contacted by telephone and solicited for participation. For those selected whose telephone numbers were not listed (n = 70; 6%), we used door-to-door solicitation. During these solicitations, the study was once again described, and all questions regarding the research were answered. Appointments for survey interviews were made with all those who agreed to participate in this study. Face-to-face interviews were subsequently conducted in the homes of all consenting participants by trained African American interviewers at designated appointed times. The data were collected between February 1992 and July 1993. Only 12% of those selected declined to participate; the refusal rates for telephone and door-to-door solicitation were similar. Out of 1,114 interviews, a total of 998 were successfully completed.

Measurement

Table 1 displays the means, standard deviations, and percentages of the variables used in this study to evaluate the factors associated with paranoid ideation among our sample of elderly African American persons. We used the Brief Symptom Inventory (BSI), a 53-item self-reported symptom inventory (essentially the brief form of the revised Symptom Check List 90; Derogatis 1993) in the present study. The BSI was designed to reflect the psychological symptom patterns of psychiatric and medical patients as well as nonpatients (Derogatis 1993). Paranoid ideation and depression are among nine primary symptom dimensions that are measured by the BSI. The cardinal characteristics of projective thought, hostility, suspiciousness, grandiosity, centrality, fear of loss of autonomy, and delusions are viewed as primary aspects of paranoid ideation. The symptoms of the paranoid ideation dimension consist of the following beliefs:

1. Others are to blame for most of your troubles.

2. Most people cannot be trusted.

3. You are watched or talked about by others.

4. Others are not giving you proper credit for your achievements.

5. People will take advantage of you if you let them.

The internal consistency and test–retest reliability coefficients for the Paranoid Ideation scale have been reported as .77 and .91, respectively (Derogatis 1993). Each item of the BSI is rated on a 5-point scale of distress (0–4), ranging from not at all (0) at one pole to extremely (4) at the other. The alpha reliability of this composite indicator for our sample was .71.

The symptoms of the depression dimension reflect a representative range of the indications of clinical depression, and this dimension is composed of a set of symptoms and signs that are associated clinically with high levels of manifest depression (Derogatis 1993). The internal consistency coefficient for the depression index has been reported as .85 (Derogatis 1993). The Cronbach's alpha for this index among our sample was .76.

Cognitive impairment was measured by means of the Mini-Mental State Examination (MMSE; Folstein, Folstein and McHugh 1975). The MMSE has been extensively tested for its ability to identify cognitive impairments accurately (Tombaugh, McIntyre, and McIntyre 1992). However, a few studies showed that participants' educational level influenced the results of the MMSE (Escobar and colleagues 1986; Murden, McRae, Kaner, and Bucknam 1991). The MMSE used in this study consisted of five cognitive domains or functions: orientation to time, orientation to place, registration of three words, attention (spelling world backward, and recall of three words. Following Morris, Mohs, and Rogers 1988, we did not include the calculation function (the serial 7s), but used world as an alternative. The word world was spelled forward (or corrected) before spelling it backward.

Self-reported memory deficits were evaluated by the 27-item Memory Functioning Questionnaire (Gilewski and Zelinski 1988; Zelinski and Gilewski 1988). The 27-item scale used in this study has four subscales: General Rating of Memory, Retrospective Functioning, Frequency of Forgetting, and Remembering Past Events. The internal consistency estimate (Cronbach's alpha) for the overall summated 27-item scale was very high (α = .93), and none of the correlation coefficients among the items was less than .40. The total score ranged from 28 to 100 (SD = 12.23), and higher scores on this index represent a more positive self-appraisal.

Emotional and tangible social support was measured with a modified version of the inventory of social supportive behaviors (Barrera, Sandler, and Ramsay 1981; Krause 1986). Emotional support consisted of 11 items and was based on the personal qualities and behaviors of a support person, including empathy, caring, love, and trust. Instrumental support consisted of 9 items and was based on the instrumental behaviors performed by a helping person. The helping person intervenes by providing transportation, making a financial contribution, helping with work obligations, or providing some other form of material aid or direct relief (Krause 1986). A higher score on these indicators reflects greater emotional and tangible support. The reliabilities of the 11 and 9 items for emotional and instrumental support were .87 and .75, respectively.

This study included a measure of negative stressful life events as one of the potential correlates of paranoid ideation. The measure of stressful life events used in this study is a revised version of the scale developed by Holmes and Rahe 1967 and consists of 16 items. The items used for this index are stressful life events that range from the death of a spouse or a child to being the victim of a crime. Each stressful life event was coded dichotomously, with yes coded as 1, and no coded as 0. Therefore, the total score can range from 0 to 16, and a higher score on the scale indicates more stressful life events.

An index measuring self-reported health status was constructed, using three items: self-reported health rating, self-reported health status compared with 2 years ago, and self-reported health status compared with peers of the same age. Principal-components analysis of these three variables formed only one factor, and this factor was highly correlated with each of the variables (% of variance = 60.4; communality, .58 to .64).

Self-reported vision was evaluated with the single question "How is your vision [if Respondent wears glasses] after you put your glasses on?" with response categories of 1 = excellent, 2 = good, 3 = fair, and 4 = poor. Similarly, self-reported hearing was evaluated with the single question "How is your hearing?" with response categories of 1 = excellent, 2 = good, 3 = fair, and 4 = poor.

Limitation of daily activities was derived from 11 questions asking respondents to report how much their daily activities were limited by any given chronic illnesses (i.e., heart disease; hypertension; diabetes; arthritis; respiratory diseases; kidney, teeth and gum, and blood circulation problems; stroke; and cancer). Scores ranged from 0, not at all, to 3, a great deal. It is important to note that the limitation of daily activity scores are derived from the chronic condition questions, not the Activities of Daily Living index questions. This would probably underreport limitations in activities where the individual has limitations but does not attribute them to any of the 11 listed conditions. Age, sex, education, and monthly income are among selected demographic and socioeconomic status variables.

Analysis Plan.

In performing the multivariate analysis, we used two statistical techniques: multiple logistic regression and ordinary least square (OLS) regression. To do so, both dichotomous and continuous scores were created with the index of paranoid ideation. As mentioned in the measurement section, the five symptoms composing the paranoid ideation dimension were measured on a 5-point scale of distress (0–4), ranging from not at all (0) at one pole to extremely (4) at the other. The total score ranged from 0 to 17. The average of the index (total score divided by the number of items) ranged from 0 to 3.4 (M = 0.379; SD = 0.542), with the higher scores representing more paranoid ideation symptoms. Multiple regression analyses were conducted with this summated continuous raw score. Following procedures outlined by Lewis-Beck 1980, we submitted all of the statistical assumptions of OLS regression analyses to testing. No harmful multicollinearity was detected. However, we acknowledge that the index of paranoid ideation (dependent variable) is skewed (skewness = 1.937; SE = .077). Therefore, one of the assumptions of OLS regression is violated. Because our sample size was large (N = 998), unbiased estimates of regression coefficients are expected. Nonetheless, as an added safeguard, we also estimated the model for the dichotomous measure, using logistic regression analyses. For the purpose of logistic regression analyses, a score of 1.2 for the index of paranoid ideation is considered a cut-off point. The cut-off point score is more than 1.5 units of standard deviation above the average. This cut-off point score is similar to the mean for paranoid ideation among a sample of 1,002 psychiatric outpatients (M = 1.12; SD = 0.95) from the BSI scales' validation study (Derogatis 1993).

Results

Characteristics of the Sample

The mean age of the sample was 72.4 years; the range was 62 to 99 years. The sample consisted of 24% men and 79% women. Only 21% of the sample were currently married, 55% were widowed, and 25% were divorced, separated, or never married. The mean education was 8.7 years, and 49% of the participants indicated that they had no formal education beyond eighth grade. Slightly more than 24% reported completing high school. Seventy percent of the sample reported a monthly income of $750 or less ($9,000 per year). The average reported income was $680 per month ($8,160 per year).

Similarly to our data, census information indicated that 61% of elderly African Americans living in New Orleans had a yearly income of $9,999 or less; 51% had less than a 9th-grade education; and 24% had completed high school or had post–high school education (U.S. Bureau of the Census 1990).

With regard to self-reported health status, 8% rated their present health as excellent; 32% described their health as good, 42% as fair, and 18% as poor. A 1992 National Health Interview Survey from the Centers for Disease Control and Prevention (U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics 1995) found, in findings remarkably similar to our own, that 9.3% of African Americans ages 65 years and older reported their health status as excellent, compared with 17.6% who rated their health status as poor (N = 2,622).

Of our study population, 7% indicated no chronic illness. The number of chronic illnesses reported by the remainder of the sample ranged from 1 to 11; the average number of conditions reported was just over 3 (M = 3.3). The most frequently cited chronic illnesses were arthritis (66%), hypertension (62%), eye problems (47%), heart trouble (33%), diabetes (24%), and circulation problems (24%). Almost 38% of participants reported that at least one condition greatly limited their daily activities. Of the respondents, 25% represented no limitation in their daily activities. The findings of the 1986 Functional Limitations Supplement to the National Health Interview Survey, which indicated that almost 29% of elderly African Americans reported no difficulty in their daily activities (Blesch and Furner 1993), are similar to our study. Table 1 displays the means, percentages, standard deviations, ranges, and coding algorithms of all independent variables used in this study.

Paranoid Ideation

Our data classified 10% of elderly African American participants in this study as being above the cut-off point for having symptoms of paranoia. Table 1 displays the means and standard deviations of the paranoid ideation index and the five symptoms comprising its dimension. The mean for the paranoid ideation index was .379 (SD = .542). This is identical to the nonpatient normative sample of 974 of the BSI scale validation study by Derogatis 1993. Within our sample, 63% of the individuals with scores of 6 or more answered "quite a bit" or "extremely" to at least two symptoms of paranoid ideation. In addition, 86% of participants with scores of 6 or more answered "moderately," "quite a bit," or "extremely" to at least two symptoms of paranoid ideation.

Bivariate Relationship

The zero-order correlations among the index of paranoid ideation and each of the selected independent variables are displayed in Table 2 . This table indicates that 7 of 14 independent variables were significantly related to paranoid ideation among our sample of elderly African American persons. They were depression (r = .47), self-reported memory function (r = −.23), number of stressful life events (r = .22), limitation of daily activities (r = .15), self-rated health status (r = .10), instrumental social support (r = −.08), and income (r = .08). To control for possible multicollinearity, we examined all correlations of the independent variables. A further inspection of Table 2 indicates that the interrelations among the independent variables were modest, and no sign of harmful multicollinearity was detected. Of 105 bivariate correlation coefficients, only 4 were greater than .4 (Table 2 ). These interrelationships were as follows: between education and cognitive deficit (r = −.40); between self-reported memory function and depression (r = −.40); between emotional support and instrumental support (r = .50); and the largest, between self-rated health index and Activity of Daily Living scale (r = .53).

Multivariate Analysis

Table 3 contains the standardized multiple regression coefficients and standard error for predicting paranoid ideation, illustrating that with all independent variables in the equation, the adjusted R² was .252, F(14, 978) = 24.75, p < .0001. Of 14 independent variables, 6 showed a significant relationship with paranoid ideation. Examination of the data revealed that of 7 independent variables that showed significant relationships in bivariate levels, all except 1 (limitation of daily activities) remained significant in multiple regression analysis. As was expected, depression (β = −.43) was among the predominant correlates of paranoid ideation. Stressful life events, self-rated memory index, instrumental support, income, and self-rated hearing were also significantly related to paranoid ideation. These results indicate that a higher level of paranoid ideation was found among those with (a) higher levels of depression, (b) higher numbers of negative stressful life events, (c) more self-reported memory problems, (d) lower levels of instrumental social support, (e) lower levels of income, and (f) more hearing problems. Limitation of daily activities, which initially showed a significant relationship with paranoid ideation in bivariate analysis, lost its significance in the multiple regression as a result of the variance shared with depression.

We used a new model to test for the presence of interaction effects between the social support indices, depression scale, and stressful life events. Because this model included new product terms and in order to avoid multicollinearity between the independent variables (first-order terms) and the product terms, the independent variables were centered (put in deviation score form so their means were zero) and the product terms were formed by multiplying the centered independent variables of interest (Aiken and West 1991). No interaction effects were detected.

The results of a multivariate logit analysis connecting paranoid ideation to the 14 independent variables reveals that, similar to the multiple regression analysis, the same 6 variables registered a significant independent impact on paranoid ideation (table not presented here). Table 4 shows the predicted probability fluctuation (Hamilton 1990) of paranoid ideation for those 6 variables that were significant on the logit model. Fluctuation was calculated for two hypothetical respondents who were "perfectly average" in all relevant predictors except that one respondent was placed at the 10th percentile rank and the other was placed at the 90th percentile rank of each variable whose independent impact on paranoid ideation was under investigation. For example, we found that once all other relevant variables were held constant, elderly African American participants whose measured level of self-reported memory function index placed them in the 90th percentile were 7% less likely to be identified with paranoid ideation symptoms than those placed in the 10th percentile. The nature of the linkages between paranoid ideation and hearing impairment, income, stressful life events, instrumental support, and depression are also provided in Table 4 .

Discussion

In this study, we provide the first published data assessing the prevalence and associations of paranoid symptoms in a selected community sample of elderly African Americans. The model measured paranoid ideation with the BSI and included 14 independent variables: demographic characteristics, cognitive impairment, depression, self-reported memory deficits, social support, stressful life events, limitation in daily activities, health status, and self-rated hearing and vision.

Symptoms of paranoia were found in 10% of this sample. How does this compare with the general aged population? We know, for example, that in 1964 Lowenthal documented 2% paranoid ideation; 20 years later Christenson and Blazer 1984 reported 4% paranoid ideation symptoms in a community of elders; and 12 years later Blazer and associates 1996 reported that symptoms of paranoia were found in 9.5% of a community sample of elders. We do not know the historical trends for paranoid ideation among elderly African Americans. All we now know is that the rate we found (10%) is higher than that found in previous studies but appears comparable to the general aged population. Nevertheless, these data are alarming for African American populations. The alarm is raised because our results are derived from a sample drawn from persons in the community who might have been receiving greater social support from senior service facilities, a feature that might distinguish them from the general community of African American elderly persons. This would imply that rates in the general population of African American elderly persons may in fact be considerably higher than those in our sample. Regardless of this hypothesis, the rate of 10% points to an urgent emerging problem requiring further research.

Examination of the data using both OLS and logistic regression revealed that 6 of the 14 independent variables used in this study showed a significant relationship with paranoid ideation. The results indicated that a higher level of paranoid ideation was found among individuals with higher levels of depression, greater numbers of negative stressful life events, more self-reported memory problems, lower levels of instrumental social support, lower levels of income, and more hearing problems. These overall results confirm the most recent Swedish study (Forsell and Henderson 1998), which reported that paranoia symptoms were found among individuals with higher levels of depression, lower levels of social support, and poorer physical health.

Depression was identified as the strongest correlate of paranoid ideation symptoms among our sample of elderly African Americans. This finding was similar to data reported in previous cross-sectional and longitudinal studies using multivariate analysis. At first glance, the relationship between paranoid ideation and depression seems contradictory, as a classical observation in depressed people is attribution of negative events to self and attribution of positive events to circumstances or actions of others (Robbins and Hays 1995). However, the older adult is losing friends, resources, networks, income, and mobility. The struggle to replace what is being lost could perhaps lead to abnormal attention to and preferential recall of threat-related information (Bentall, Kaney, and Bowen-Jones 1995), or the depression may simply be associated with chronic illness due to aging or may be an appropriate response to a hostile environment (Blazer et al. 1996). The findings concerning depression are complex and suggest the need for more in-depth exploration through a longitudinal study.

A second finding of significance was the correlation between lower levels of instrumental social support and paranoid ideation. Jette and Winnett 1987 argued that vulnerable individuals who are unable to respond effectively to the challenges of advancing age, owing to lack of personal and social resources, often feel increasingly isolated, helpless, and misunderstood and experience a lack of acknowledgment or confirmation. Over time, their sensitivity may increase to the point that they become suspicious of others' intentions toward them.

Greater number of stressful life events was also a correlate of significance. These findings differ considerably from those of Blazer and associates 1996, which documented no connection between negative life events and paranoid ideation. One explanation for this discrepancy might be the difference between Blazer and colleagues' predominantly rural setting (52%) as compared with our urban one. Our data suggest that elderly African Americans living in the inner city are influenced by excessive crime, frequent loss (death of spouse, friends, relatives), lack of access to proper medical care, and serious financial difficulties. Thus, they may engage in a defensive strategy against depression and the resulting self-negation. This struggle leads to what is so frequently referred to in the African American community as "healthy paranoia."

There was notable association between self-reported memory problems and symptoms of paranoid ideation, even after other related variables, including depression and cognitive and sensory functioning, were held constant. Over 14% of this sample of aged African Americans reported that they had major problems with their memory functions. Moreover, 26% of participants described their memory function as worse than 1 year ago, and 13% of this sample said that they were very bad at remembering things that occurred last month. Among participants with symptoms of paranoid ideation, 24% reported that they had major problems with their memory functions; 38% described their memory function as worse than 1 year ago; and 19% said that they were very bad at remembering things that occurred last month. This relationship between subjective memory assessment and symptoms of paranoid ideation has significance in terms of broadening the role of primary care physicians in promoting mental health and quality of life among elderly persons.

In the last finding of note, we detected an independent impact of hearing problems on symptoms of paranoid ideation, a finding consistent with the literature (Almeida et al. 1995), but no connection between self-reported vision impairment and symptoms of paranoid ideation. Several research findings have suggested that hearing impairments increase self-reported memory problems (Bazargan and Barbe 1994) and diminish cognitive capacity (Granik, Kleban, and Weiss 1976; Lindenberger and Baltes 1994; Marsiske, Delius, Lindenberger, Scherer, and Tesch-Romer 1996; Sands and Meredith 1989; Thomas et al. 1983). Together these may reinforce a preexisting tendency to social isolation, withdrawal, and suspiciousness (Almeida et al. 1995). Reports of a significant reduction of psychotic activity after the fitting of hearing aids indicates that hearing impairment is at least a predisposing factor to the development of paranoid symptoms (Almeida 1993; Eastwood, Corbin, and Reed 1981; Khan, Clark, and Oyebode 1988).

This study suffered from three main limitations. First, interpretation of the findings is limited by the cross-sectional nature of the study. For example, taking the paranoid symptoms as the independent variable, the impact of depression, self-reported memory problems, and other independent variables on symptoms of paranoid ideation were discussed. However, it is plausible that paranoid ideation induces depression. Similarly, the association of self-reported memory problems and paranoid ideation could reflect complaining behavior rather than a direct correlation between the two. Second, the data we generated are from a nonprobability sample, which limits our ability to generalize the findings. Elders who are involved with senior centers are a select group, and it is plausible that they are less suspicious than those not involved. Third, we used published BSI norms to document that this sample was identical to the nonpatient normative sample; however, these samples were not race matched.

The strengths of this study include sample size, reliability of the instruments, and the analysis of recognized variables. The sample was a large one, representing a diverse community of urban elderly African Americans. We used several standard instruments with well-established reliability and validity to measure paranoid ideation, depression, memory deficit, cognitive deficit, social support, and other key concepts. Last, this study included most of the variables that have been previously identified as potential correlates of paranoid ideation.

The occurrence of paranoid ideation in such a significant population should not be summarily dismissed as premorbid psychosis. Cultural contexts that promote feelings of alienation, powerlessness, and victimization are more likely to engender paranoid thinking, which may serve a self-protective function. It is this psychological self-protection that may be the link between paranoia and elderly persons in culturally oppressed groups (Whaley 1998). These findings suggest a need to critically examine and appropriately diagnose elderly African Americans' paranoid symptoms (Abedimpe 1981; Fabrega et al. 1988; Raskin, Crook, and Herman 1975) and to produce a greater volume of high-quality scientific data toward the provision of adequate and equal health care for this growing population.

The research reported in this article was supported by National Institutes of Health Grant G12 RRO 3026-09 from the National Center for Research Resources, NIH/NCRR/RCMI. We are grateful to Lindsay Berg and Risa Flynn for their professional editorial and helpful comments and suggestions.

References